Meshel N. S. - The Grammar of Sacrifice PDF

Page 1
T H E “ G R A M M A R ” O F S AC R I F I C E
WITH
A “GRAMMAR” OF Σ
11:50:13:06:14 Page 1
Page 2
11:50:13:06:14 Page 2
Page 3
The “Grammar” of Sacrifice

A Generativist Study of the Israelite Sacrificial
System in the Priestly Writings
with
A “Grammar” of Σ
NA PH TA L I S . M E S H E L
11:50:13:06:14 Page 3
Page 4
3
Great Clarendon Street, Oxford OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries.
© Naphtali S. Meshel 2014
The moral rights of the author have been asserted
First Edition published in 2014
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2013950786
Data available
ISBN 978–0–19–870556–7
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
11:50:13:06:14 Page 4
Page 5
for Maya, Eyal, Ellah, Hadas, and Itai
11:50:13:06:14 Page 5
Page 6
11:50:13:06:14 Page 6
Page 7
Preface
There are two aspects to this study. On the one hand, it is a work in biblical
studies, the main objective of which is to contribute to the interpretation
of texts in the Hebrew Bible pertaining to sacrificial rituals. Although
the sacrificial laws constitute a significant part of the Pentateuch, and
although sacrifice was clearly a cardinal institution in ancient Israelite
culture, many aspects of the highly technical biblical texts that pertain to
sacrifice remain obscure. This study is thus part of an ongoing scholarly effort
to understand more fully the Israelite sacrificial system within its ancient
Near Eastern setting––in all of its particulars as well as in its broader
contours.
On the other hand, it is a theoretical study in religion, more specifically in
the study of ritual, ritology. Its main objective is to demonstrate that it is
possible to formulate a “grammar” of a ritual system. In this sense, it is an
attempt to pursue the idea––first suggested in ancient times and reiterated in
modern anthropology––that ritual systems possess grammars analogous to
those of natural languages, by proceeding to compose one such grammar.
As a work in the field of biblical studies, this study is based on an analysis of
biblical sacrificial texts within their ancient Near Eastern context, using the
philological, text-historical, and literary tools of biblical criticism. Since such
philological work has been carried out in a very comprehensive manner,
particularly in late twentieth-century biblical scholarship, the present study
can rely heavily on existing scholarship in this regard. However, since a certain
number of widely accepted interpretations of biblical texts that pertain to the
Israelite sacrificial system are called into question upon reexamination, these
are subjected to new analysis and the results are presented.
Following this preliminary groundwork, I shall attempt to identify more
precisely some of the basic elements employed in the Israelite sacrificial
system––for example, to arrive at the precise denotation of a number of
technical terms for the categories of sacrificial animals.
The next task is to identify the operative categories underlying the Israelite
sacrificial system represented in the texts under examination (henceforth
referred to as Σ). The rationale underlying this venture is that if grammars
of natural languages consist of operative categories such as phonology and
syntax, and if Σ, too, has a grammar, then one should expect to find in it a
number of operative categories that are either identical to or different from
those found in the grammars of natural languages. If the operative categories
of Σ are found to be different from those of natural languages, then they must
be defined clearly and given appropriate names.
11:50:13:06:14 Page 7
Page 8
viii Preface
The present volume comprises two separate works––the bulk of the present
volume, The “Grammar” of Sacrifice, is dedicated to the identification and
definition of the operative categories of Σ. It should be perceived as an
introduction to the “Grammar” itself, A “Grammar” of Σ, which is located at
the end of the present volume.
A word is in order regarding the use of the term “Generativist” in the
subtitle. A brief glance at the present volume might suggest that it is
essentially a structuralist study of a ritual system. However, while I
gratefully acknowledge this study’s debt to Structuralism, there are various
reasons–discussed in sections 1.3.2, 1.3.3, 1.4 and 5.7––for preferring a
generativist model.
The present book is the result of research begun during 2005–2009 at the
Hebrew University of Jerusalem. It was conceived as a doctoral dissertation
under the supervision of my teacher and mentor, Professor Baruch J.
Schwartz. I thank him from the depths of my heart. I wish to thank the
members of the Ph.D. committee, Professor Israel Knohl and Professor David
Shulman for the inspiration through continued interchange. No words are
sufficient to express my gratitude to the late Professor Milgrom and to Dr. Jo
Milgrom. I am particularly grateful to Professor Roy E. Gane, whose specific
comments have been invaluable for the evolution of this project. Shortly after
I was introduced to his work on “ritual syntax,” I learned that he was in
Jordan. By the end of that week I was in Amman with him, for what would
evolve into a lasting mentorship and friendship. The specific comments that
I received from Professor Gane on many occasions have been invaluable for
the evolution of this project.
Professor Yochanan Grinshpon lit up the occasional dark hours of
intellectual despair that I experienced while working on this book.
Thanks to David and Shari Satran for their enduring friendship and
support; and to my friends and colleagues at the Hebrew University––in
particular Yakir Paz and Tzakhi Freedman––for their helpful comments.
I thank Yad Hanadiv for an ideal postdoctoral fellowship: first in Mysore
with Professors H. V. Nagaraja Rao and Gangadhara Bhatt; and later at the
University of Pennsylvania with Professor Jeffrey Tigay.
Princeton University has been an ideal setting for completion of this
project. Comments from my colleagues, particularly those in the Department
of Religion and in the Program for Judaic Studies, have improved the final
product in many ways. In particular, I am grateful to Professor John Gager
and Professor Martha Himmelfarb for the conversations we have had on
many aspects of this work. Professor Jonathan Gold read sections of the book
and offered many insightful comments. I am thankful to Professor Gary
Rendsburg from Rutgers University for his support and advice.
The Institute for Advanced Study at Princeton was a most gracious host
during 2013–14, enabling me to explore the ways in which the “grammar” of
11:50:13:06:14 Page 8
Page 9
Preface ix
Σ is applicable more broadly to Sanskrit and Semitic sacrificial ritual texts,
and to identify the limits of this applicability.
Above all, I wish to thank my students for the intellectual exchange I have
had with them over the course of the years. Anton Fleissner’s fingerprints are
discernible on every page of this book: he commented on everything from the
grand contours of the “grammar” of sacrifice to the minutiae of the grammar
of each English sentence. I thank my colleague Judah Kraut, who, with the
acumen of an expert Bible scholar and a discerning critical eye, led me to
rethink several specific details, as well as general theoretical claims. Rebecca
Khalandovsky contributed significantly to the chapter on Jugation. Liane
Marquis assisted with the proofs, and offered several insightful comments.
Special thanks are due to Jessica O’Rourke Suchoff for her continuous
engagement in this project and in particular for her invaluable contribution
to the final stages of the book’s production.
I wish to thank the Memorial Foundation for Jewish Culture for support-
ing this project financially. I wish to express my gratitude to the Office of
the Dean of the Faculty at Princeton University, to the Anonymous Fund, and
particularly to the Program in Judaic Studies at Princeton University and to
Professor Peter Schäfer for the generous support I have received, especially
during 2013–14. Thanks are due to the Firestone Library staff and to the
Princeton Theological Seminary Library––in particular to Kate Skrebutenas.
The library staff at the Institute for Advanced Study have been extraordinarily
helpful.
I would like to thank Tom Perridge, the editorial board, and the staff at
Oxford University Press for the publication of this book.
I wish to thank my parents, Yonit and Yitzchak Meshel, and my sister, Yael
Mishael. Above all, thanks go to my wife, Maya , my better half, my friend and
companion for life, and to our children, Eyal, Ellah, Hadas and Itai.
As a final remark, I would like to note that it is easy to forget, amidst the
academic parlance, the graphic illustrations, and abstract formulae, that
somewhere at the end-nodes of this “grammar” there stood priests and
laypeople, sharp knives, a consuming fire, and animals of flesh and blood,
subject to the violence of the cruelest and most systematic killer of them all.
11:50:13:06:14 Page 9
Page 10
11:50:13:06:14 Page 10
Page 11
Contents
THE “GRAMMAR” OF SACRIFICE

List of Figures and Tables xii
List of Abbreviations xiv
List of Definitions xviii
1. Introduction 1
2. Zoemics 29
3. Jugation 63
4. Hierarchics 104
5. Praxemics 130
6. Meaning 174
7. The Grammar of Sacrifice and the Sacrifice of Grammar 198
Bibliography 210
Index 227
A “GRAMMAR” OF Σ
Zoemics 1
Jugation 4
Hierarchics 9
Praxemics 9
Index 27
11:50:13:06:14 Page 11
Page 12
List of Figures and Tables
FIGURES
Figure 1. Structure of an agnicayana 15

Figure 2. Classification according to zoological class 34
Figure 3. Two-dimensional zoemic map 36
Figure 4. Three-dimensional zoemic map 36
Figure 5. Basic jugational pattern according to Numbers 15 65
Figure 6. Jugational pattern of semolina offering in Leviticus 2:1–3 66
Figure 7. Relation of jugates in Numbers 6 71
Figure 8. Relation of jugates in Exodus 29 and Leviticus 8 74
Figure 9. Relation of jugates in Leviticus 7 76
Figure 10. Jugational pattern in Leviticus 23:18–20 79
Figure 11. Ezekiel’s linear jugational model 81
Figure 12. Model A: male goat with subordinate jugation of wine 90
Figure 13. Model B: male goat without subordinate jugates 90
Figure 14. Basic jugational pattern including salt 98
Figure 15. Jugation of saltB to saltA 102
Figure 16. Never-ending jugation of salt 102
Figure 17. Hierarchic structure of wholeburnt offerings in Numbers
28:11–15 105
Figure 18. Hierarchic structure for the New Moon ceremony 105
Figure 19. Hierarchic structure in Leviticus 9:18 107
Figure 20. “Purification offering of purgation”: minimal hierarchical
structure 108
Figure 21. Rule H1 109
Figure 22. The “purification offering of purgation” in the Temple Scroll:
hierarchic structure 110
Figure 24. Hierarchic structure in Ezra 8:35 113
Figure 25. Hierarchic structure of an ordination offering according to
the Temple Scroll 114
Figure 26. Hierarchic structure of a purification offering according to
the Temple Scroll 116
Figure 27. Ibn Ezra’s interpretation of the offerings of the wealthy and
the indigent offerers 117
11:50:13:06:14 Page 12
Page 13
List of Figures and Tables xiii
Figure 28. Hierarchies above and beneath the zoemic level 118
Figure 31. Hierarchics within a single zoeme according to b Zebah. 9b 120
Figure 32. Purification zoemes in P 120
Figure 37. Two representations of Patañjali’s allegory 208
TABLES
Table 1. Zoemes attested in P 37

Table 2. Table of jugation according to Numbers 15 65
11:50:13:06:14 Page 13
Page 14
List of Abbreviations
ABBREVIATIONS USED IN BIBLIOGRAPHY

AND REFERENCES
For standard abbreviations used in biblical studies, see JBL 117 (1998), 555–79.
ADL Jonas C. Greenfield, Michael E. Stone, and Ester Eshel (eds), The Aramaic
Levi Document: Edition, Translation, Commentary. Studia in Veteris
Testamenti pseudepigrapha 19. Leiden: Brill, 2004.
ANET J.B. Pritchard (ed.), Ancient Near Eastern Texts Relating to the Old
Testament. Princeton: Princeton University Press, 1969.
Ant. Jud. Flavius Josephus, Judean Antiquities 1–4, Translation and Commentary
by Louis Feldman. Flavius Josephus: translation and commentary 3.
Leiden: Brill, 2000.
BETL Bibliotheca Ephemeridum Theologicarum Lovaniensium
BWDW Gerhard Wahrig, Hildegard Krämer, and Harald Zimmermann (eds),
Brockhaus Wahrig Deutsches Wörterbuch. Wiesbaden: Brockhaus,
1980–1984.
BHS Biblia Hebraica Stuttgartensia
BKAT Biblischer Kommentar Altes Testament
BN Biblische Notizen
CAD The Assyrian Dictionary of the Oriental Institute of the University of
Chicago. Chicago: The Oriental Institute, 1956–2010.
CGED Cassell’s German–English English–German Dictionary. New York:
Macmillan, 1978.
CTH Catalogue des Textes Hittites
DLF Emile Littré, Dictionnaire de la langue française. Paris: Gallimard, 1958.
DSL Philo, On the Special Laws (De Specialibus Legibus). Translated by F. H.
Colson. Loeb Classical Library, Philo 7. London: W. Heinemann, 1937.
DUL Gregorio del Olmo Lete and Joaquín Sanmartín (eds), A Dictionary of
the Ugaritic Language in the Alphabetic Tradition. Translated by Wilfred
G.E. Watson. Leiden: Brill, 2003.
FAT Forschungen zum Alten Testament
Gen. Apoc. Joseph A. Fitzmyer, The Genesis Apocryphon of Qumran Cave I: A
Commentary. 2nd rev. edn. Rome: Biblical Institute Press, 1971.
GKC E. Kautzsch (ed), Gesenius’ Hebrew Grammar, 2nd English edn; rev. in
accordance with the 28th German edition by A.E. Cowley; Oxford:
Clarendon, 1910).
11:50:13:06:14 Page 14
Page 15
List of Abbreviations xv
GNT Kurt Aland, et al. (eds), The Greek New Testament. 4th edn. Stuttgart:
Deutsche Bibelgesellschaft, 1994.
HIM Maimonides, Laws of Prohibitions Pertaining to the Altar (YRWSYA 'LH
XBZM).
HKM Maimonides, Laws of Temple Vessels ($DQMH YLK 'LH).
HMQ Maimonides, Laws of Sacrificial Procedures (TWNBRQH H$EM 'LH).
HPM Maimonides, Laws of Disqualified Sancta (OY$DQWMH YLWSP 'LH).
HQP Maimonides, Laws of the Paschal Offering (XSP IBRQ 'LH).
HR History of Religions
HTM Maimonides, Laws of Regular and Festival Offerings (IYPSWMW IYDYMT 'LH).
HTR Harvard Theological Review
HUCA Hebrew Union College Annual
HYH Maimonides, Laws of the Service of the Day of Atonement (TDWBE 'LH
OYRWPKH OWY).
IHQ Indian Historical Quarterly
JA Journal Asiatique
JAOS Journal of the American Oriental Society
JB Jerusalem Bible
JBL Journal of Biblical Literature
JIPh Journal of Indian Philosophy
JM Paul Joüon A Grammar of Biblical Hebrew. Subsidia Biblica 14/i–4/ii;
trans. and rev. T. Muraoka. 2 vols. Roma: Editrice Pontificio Istituto
Biblico, 1996.
JTS Journal of Theological Studies
KAI H. Donner and W. Röllig, Kanaanäische und aramäische Inschriften.
vols 1–2. 2nd edn. Wiesbaden: Harrassowitz, 1973.
KBL L. Koehler and W. Baumgartner (eds), Lexicon in Veteris Testamenti libros
Leiden: Brill, 2nd edn. 1958.
m. Mishna
M.–W. Monier Monier-Williams, An English–Sanskrit Dictionary Ottawa:
Laurier Books, 2001.
WNCD Webster’s New Collegiate Dictionary. Springfield, MA: G. & C. Merriam,
1980.
Mah. S.D. Joshi and J.A.F. Roodbergen, Patañjali’s Vyākaran.a-Mahābhās.ya
vol. 1: Paspaśāhnika. Publications of the Centre of Advanced Study in
Sanskrit 15. Pune: University of Poona, 1986.
Maim. Maimonides
MMT Elisha Qimron and John Strugnell, Miqs.at Ma‘ase ha-Torah. Discoveries
in the Judaean Desert 10. Oxford: Clarendon, 1994.
NCFD The New Cassell’s French Dictionary. New York: Funk and Wagnalls, 1962.
11:50:13:06:14 Page 15
Page 16
xvi List of Abbreviations

Neof. Alejandro Díez Macho Neophiti 1: Targum palestinense ms de la biblioteca
vaticana, Tomo III: Levítico. Textos y estudios 9. Madrid: Consejo
Superior de Investigaciones Cientificas, 1971.
NEB New English Bible
NSK Neuer Stuttgarter Kommentar
PEQ Palestine Exploration Quarterly
RSV Revised Standard Version
Sam Avraham v.e-Ratson Tsedak.ah (eds), H
. amishah h.umshe Torah : nusah.
Shomroni im hadgashah meduyek.et shel ha-shinuyim ben shte ha-
nush.aot. Tel-Aviv, 1961–1965.
SED Alexander Militarev and Leonid Kogan, Semitic Etymological Dictionary,
AOAT 278/1–2. Münster: Ugarit-Verlag, 2000–.
t Mosheh Shemuel Tsuk.ermandel (ed.), Tosefta: Al pi kitve yad Erfurt.
u-V.inah im mare mek.omot v.e-h.ilufe girsaot u-mafteh.ot. Jerusalem:
Bamberger et V. ahrman, 1937.
Tanh. Midrash Tanh.uma al h.amishah h.umshe Torah. 2 vols. Jerusalem: Eshkol,
1963.
TDOT G. Johannes Botterweck, Helmer Ringgren, and Heinz-Josef Fabry (eds),
Theological Dictionary of the Old Testament. Translated by John T. Willis
and David E. Green. 15 vols. Grand Rapids: Eerdmans, 1977–2006.
TS Elisha Qimron, The Temple Scroll: A Critical Edition with Extensive
Reconstructions. Beer Sheva: Ben Gurion University of the Negev, 1996.
TSKY The Taittirı̄ya Sam
. hitā of the Black Yajur-veda, with the commentary of
Bhat.t.abhāskaramiśra (ed). A. Mahādeva Śāstri and Panditaratnam K.
Rangāchārya (vol. 9, Government Oriental Library Series, Bibliotheca
Sanskrita 17; Mysore: Government branch press, 1898).
UF Ugaritische Forschungen
Vg Vulgata
VP Vākyapadı̄yam Part I: With the Commentaries Svopajñavr.tti by
Harivr.s.abha and Ambākartrı̄ by ‘Padmaśrı̄’ Pt. Raghunātha Śarma.
Varanasi, 1988.
VT Vetus Testamentum
WMANT Wissenschaftliche Monographien zum Alten und Neuen Testament
GENERAL SIGLA AND ABBREVIATIONS
Aram. Aramaic
BH Biblical Hebrew
D the Deuteronomic document
Gk Greek
11:50:13:06:14 Page 16
Page 17
List of Abbreviations xvii

H the Holiness Legislation
inf. abs. infinitive absolute
J the Yahwistic document
MH Mishnaic Hebrew
ms(s) manuscript(s)
MT Masoretic Text
P the Priestly document in the Pentateuch, including P and H. Where P is
used in distinction from H, this is clearly indicated in the text.
pron.suff. pronominal suffix
pron.pos.suff. pronominal possessive suffix
text partially preserved text (in English translation)
[text] reconstructed text (in English translation or in original language)
USQU Hebrew text not preserved in any of the manuscripts of TS
|| parallelism
indicates a transformation performed in the process of the construction
of a single grammatical sequence
indicates a diachronic shift. The text to the left of the arrow applies to the
more ancient system; the text to the right applies to the later system.
* (before a linguistic form or a text) indicates an unattested form or a
hypothetical text
11:50:13:06:14 Page 17
Page 18
List of Definitions
The pronunciation symbols follow the simplified system used by Merriam-

Webster and other American dictionaries.
GENERAL
: (small caps) is used in contradistinction from “grammar” (lower-

case) to designate the particular grammar of Σ offered in this study.
Note that other grammars of this system can be composed, and other
sacrificial systems may have their own “grammars.”
Σ: the ancient Israelite Priestly sacrificial system
TECHNICAL SACRIFICIAL TERMS
asham see reparation offering

atomact n sgl: an elementary unit of action (see §5.5.1)
bovine adj, also nominalized adj: (used narrowly to denote) a domestic “cow”
(Bos taurus) of any age or sex
calendric offering: a sacrificial offering required at a designated time of the
day, week, month, or year, as determined by the calendar
caprine adj, also nominalized adj: (used narrowly to denote) a domestic
“goat” (Capra aegagrus hircus) of any age or sex
cereal offering (also: grain offering): an offering the main constituent of
which is wheat or barley (note that only some of the cereal/grain
offerings are considered HX Ni )
mvM
co-jugates n pl: two or more jugates at the same level that are not jugated to
different elements (see p. 000)
coterminous offerings: calendric offerings occurring on the same day
grain offering see cereal offering
hatta’t see purification offering
hierarchics n pl but sgl or pl in constr: the study of the composition of
sacrificial types wherein one or more sacrificial types constitute
another sacrificial type
jugate \jə-gāt\ vt: to join (one sacrificial material to another)
jugate \jü-gət, jü-gāt\ n: any sacrificial material, either animal or non-
animal, when joined with at least one other sacrificial element
11:50:13:06:14 Page 18
Page 19
List of Definitions xix

jugation \jə-gā-shən\ n sgl: the joining of two or more animal or non-animal
sacrificial materials––jugational adj
materia sacra: animal and non-animal material placed on the altar or, in the
case of blood, applied to the sancta; primarily: flesh, suet and blood
millu’im see ordination offering
offerer: a person on whose behalf a sacrificial ritual is performed
ordination offering (also OYA iiWdLi
M, OYA id i , milluim, transliteration:
uLM
millûîm): a rare sacrificial type, attested usually in the context of the
ordination of priests. In P it entails the application of some of an
animal’s blood on the priests who are being ordained, tossing some of
the blood on the altar, and burning the suet on the altar.
‘olah see wholeburnt offering
ovine adj, also nominalized adj: (used narrowly to denote) a domestic
“sheep” (Ovis aries) of any age or sex
praxeme \prak-sēm\: a discrete segment of a ritual sequence (see §5.1)
praxemics \prak-se-miks\ n, pl but sgl or pl in constr: the study of ritual
from the point of view of the actions involved in its performance
purification offering (also TAU m
dxX, hatta’t, transliteration: h.at.t.āt): 1. a
sacrificial type in which the blood of an animal is applied at least to
the horns of the bronze altar (or to the wall and to the base of the altar,
if the animal is a bird), its suet is incinerated on the upper surface of
the bronze altar (if the animal is a quadruped), and its flesh is con-
sumed by a priest or incinerated outside the camp 2. a grain offering
offered in accordance with the law in Lev 5:11–13
reparation offering (also O$ moAm , asham, transliteration: āšām): a sacrificial
type in which blood of a zoeme is tossed upon the altar, its suet is
incinerated on the upper surface of the altar, and its flesh is consumed
by a priest
sacrificial complex: a rite that entails the offering of two or more jugates
shelamim see wellbeing offering
subordinate jugate n sgl: a jugate offered together with another jugate as its
adjunct or additive
type see sacrificial type
volitional offering: an offering brought on the offerer’s spontaneous
initiative
votive offering: an offering brought in fulfillment of a vow
wellbeing offering (also OYM im Lv$o, shelamim, transliteration: š elāmîm): a
sacrificial type in which the blood of an animal is tossed upon the
bronze altar, its suet is incinerated on the upper surface of the altar,
and most of its flesh is consumed by the offerer and his or her party
wholeburnt offering (also HL m oWE, HLm oE, ‘olah, transliteration ‘ōlāh or ‘ōlâ): a
sacrificial type in which the blood of an animal is tossed upon the
altar (or squeezed against the wall of the altar, if the animal is a bird)
11:50:13:06:14 Page 19
Page 20
xx List of Definitions
and its flesh and suet are incinerated on the upper surface of the
altar
zoeme \zō¯-ēm\ n: 1. a class of animals to which a set of sacrificial rules
applies 2. a member (specimen) of such a class––zoemic adj
zoemics \zō¯-e-miks, zō¯-ē-miks\ n pl but sgl or pl in constr: the study of the
classes of animals used in ritual sacrifice
O$
moA m see reparation offering
TAU m
dxX see purification offering
iiWdLM
OYA i see ordination offering
m WoE see wholeburnt offering
HL
OYM
im Lv$o see wellbeing offering
SETS USED IN THE GRAMMAR
sacrificial class = {calendric, non-calendric}

sacrificial genus = {votive, volitional}
sacrificial kind = {private, public}
sacrificial type = {wholeburnt, wellbeing, purification, reparation, ordination}
SIGLA USED FOR FORMAL ANALYSIS
Sigla for Zoemics (see p. 000 Table 1)
bovine (“cow” in colloquial English, e.g., RWo$o)

웨 female bovine (“female cow” of any age; HL Ge in BH)
mvE
웧 male bovine (“male cow” of any age, e.g., RK m–RQ
mZ mdm
BxH IM i , perhaps LG a)
eE
B mature bovine (either sex; no common English term)
B웨 mature female bovine (“cow,” possibly HRmm P)
d
B웧 mature male bovine (“bull,” RPd)
x
b immature bovine (“calf” of either sex)
b웨 immature female bovine (“female calf”)
b웧 immature male bovine (e.g., “male calf,” WoTN m$
vo½IB ie * . . . LG
eEa)
ovine or caprine (“sheep or goat, member of the flock,” “small cattle,”
H$oe )
f웧 (comprising s웧+g웧; immature male sheep or goat, HN m$ ie RK
mo½IB mZ oe )
m . . . H$
caprine (“member of the goat family,” ZE a)
웨 female caprine (“she-goat” of any age; OYZd
iEi TRxYE ivo$)
웧 male caprine (“he-goat” of any age, e.g., RK
mZm . . . OYZ
diE
im H IM i)
G mature caprine (either sex, no common English term)
11:50:13:06:14 Page 20
Page 21
List of Definitions xxi

G웨 mature female caprine (“nanny-goat”)
i TE
G웧 mature male caprine (DW d x)
g immature caprine (“kid,” either sex; d
HTmm$Nvo½TB d
x ZE a)
g웨 immature female caprine (“female kid”)
g웧 immature male caprine (“male kid”)
bird (turtledove or pigeon), equivalent to /
quadruped (bovine, ovine, or caprine)
ovine (“member of the sheep family,” e.g., B$ oe d
eK, rarely o$e K)
Bd
e
웨 female ovine (“female sheep” of any age; HB d
mv iK)
o$d
웧 male ovine (“male sheep” of any age, e.g., RKmZ m . . . OYB
i$omd
v
Kx i)
H IM
S mature ovine (no common English term; either sex)
S웨 mature female ovine (“mature ewe,” i.e., mature female sheep, LX a Rmin BH)
S웧 mature male ovine (“ram,” LYiA x)
s immature ovine (“lamb,” either sex)
s웨 immature female ovine (“female lamb,” d HTmm$
Nvo½TB d
x H$om vdK)
Bx
s웧 immature male ovine (“male lamb,” WoTN m$ i
vo½IB
e o$e BdK)
e
T mature turtledove
turtledove (any age or sex; RT o)
d
/ turtledove of any age or pigeon of any age, equivalent to
T/y mature turtledove or young pigeon
pigeon (any age or sex; HN ie )
mWoY-IB
y fledgling pigeon
sacrificial animal (quadrupeds and birds)
! (exclamation mark, before zoeme) indicates the zoeme and all of the
grammatical combinations of narrower zoemes within it. For example,
if a zoeme x comprises four blocks in the “zoemic map,” (below, p. 66)
the set !x would consist of the zoeme x itself, and all other zoemes
comprising one or more of those same four blocks.
E.g., !S = {S, S웧, S웨}
In words: The symbol !S designates a set of zoemes that includes the
zoeme “mature ovine” (S) itself, and two narrower zoemes, “mature
male ovine” (S웧) and “mature female ovine” (S웨).
Sacrificial type in superscript following a zoeme designates its sacrificial
type: S웧wholeburnt designates a mature male ovine (“ram”) offered as a
wholeburnt offering.
Sigla for Jugation
jx In words: a sacrificial material by the name of j on Level x

j designates the sacrificial material (a zoeme or non-animal sacrificial
material), whereas x indicates its jugational level––Level A, Level B,
11:50:13:06:14 Page 21
Page 22
xxii List of Definitions

Level C, etc.
Examples:
wineB indicates the jugate wine offered on hierarchic Level B, i.e.,
a B-level jugate of wine directly subordinate to another jugate on
Level A
S웧A indicates the zoeme S웧 (“ram”) offered on jugational Level A
S웧Awholeburnt indicates on A-level ram offered as a wholeburnt offering.
Sigla for Hierarchics
hx In words: a sacrificial material by the name of h on hierarchic Level x

h designates the sacrificial material, x indicates its hierarchic level,
e.g., Level –1, Level 0, Level +1.
Note: The plus sign (+) is always used where x is a positive integer
(e.g., h+1), in order to avoid confusion with simple ordinal indexation,
which is never indicated with a plus sign.
In hierarchics, sacrificial types can be located on hierarchic levels, e.g.:
wholeburnt+1 = {s웧0wholeburnt, s웨0purification}
In words: wholeburnt complex on hierarchic Level +1 comprises a male lamb
offered as a wholeburnt offering on Level 0 and a female lamb offered as a
purification offering on Level 0.
Sigla for Praxemics
a(g,o,t,l) a praxeme consisting of atomact a, agent g, object o, target t and

location l
+ indicates the combination of components in a single praxeme
(e.g., daub+blood+horns)
# indicates the combination of consecutive praxemes (e.g., daub+
blood+horns # pour+blood+base).
11:50:13:06:14 Page 22
Page 1
Introduction
.
catvā´ri śŕ.ngā tráyo asya pā´dā / dvé śı̄´rs.e saptá hástāso asya // trídhā
.
baddhó vr.s.abhó roravı̄ti / mahó devó mártyām ā´ viveśa
Four horns, three feet he has, two heads, seven hands he has. Bound in
three ways the bull bellows loudly. A mighty god has entered mortals.
R.gveda 4.58.3
1.1 PATAÑJALI
Masked in the form of a many-headed snake, according to an Indian

tradition, a master reveals a grammatical text to his students. The eager
grammarians cannot see their teacher––though perhaps they are able to
discern the sleek snake’s silhouette slithering behind the screen he has set up.1
Before turning to the minutiae of the grammar of the Sanskrit language, the
master offers a general introduction, extolling the virtues and joys of those
who study grammar: just as a wife filled with desire, beautifully dressed,
discloses herself to her husband, even so language discloses her body to the
man who is learned in speech.2 On the other hand, he reminds his audience
of the dangers that inhere in a failure to master correct speech: the demise of
the demons, the asuras, is associated with their substandard grammar;
and the destruction of Vr.tra resulted from the mere misplacement of an
accent in a compound. An unintentional but powerful speech-act, intended
to render this demon the vanquisher of the god Indra, instead made Indra
his vanquisher (that is the implication of índraśatru, thus accented).
Seemingly minute grammatical errors, one concludes, may have cosmic
ramifications.3
The sage is Patañjali, the text his Mahābhās.ya. Apparently composed in
1
See Śivadatta et al. 1934:27 and Chakravarti 1926:262–4.
2
Mah. 1.38; Joshi and Roodbergen 1986:57–8; Chatterji 1964:24.
3
Mah. 1.22–24; Joshi and Roodbergen 1986:37–42.
11:50:13:06:14 Page 1
Page 2
2 The “Grammar” of Sacrifice
the second century ,4 the text is essentially a commentary on the classical
Sanskrit grammar of Pān.ini. It is the first recorded text to suggest that there is
a formal homology between language and ritual, an idea that serves as an
organizing principle for this book. The rituals that we will examine, however,
belong to a culture that is geographically removed from Patañjali’s own: the
ancient Israelite culture, as reflected in the literature of the Hebrew Bible.
In Patañjali, the formal homology between ritual and language pertains to
a specific tension inherent in both systems, a tension between the ephemeral
and the eternal, the limited and the infinite. Patañjali relates a story that
demonstrates this duality: for thousands of years, the god Br.haspati tried to
relate to Indra the complete collection of individual words but could not
come to the end;5 since human lives are much shorter than that, claims
Patañjali, some general rules and exceptions must be composed if language is
to be taught. This conception of the relationship between the finite number
of utterances actually expressed and the potentially infinite number of
utterances that can be expressed closely resembles the distinction between
what Noam Chomsky calls performance and competence.6 But for Patañjali,
this is only half of the story: he reminds his readers of mahāsattras, imaginary
sacrificial rituals that can last up to a thousand years.7 Patañjali draws
an analogy between these lengthy sacrifices and the infinite number of
utterances, which, though grammatical, remain as-yet-unuttered: competence
in either cannot be acquired by direct exposure.8 Thus we find that rituals
must have “grammars,” in the sense of a finite inventory of building blocks
and a finite set of rules that can be used, once internalized, to generate an
unlimited array of combinations.
Ironically, Patañjali does not claim that rituals should have grammars. On
the contrary, his aim is to convince his readers that knowledge of a natural
language also requires mastery of a grammar, revealing that he takes it for
4
The date is approximate. See Staal 1972:xxiv, who dates the Mahābhās.ya to ca. 150 .
In Indian tradition, the grammarian Patañjali came to be identified with the later author of the
yogasū¯tras. See Puri 1957:iv.
5
Mah. 1.51; Joshi and Roodbergen 1986:74–5; Chatterji 1964:40–1. Note that even if “words”
alone are implied, and not full “utterances” in the more inclusive sense, the use of compounds
(samāsa) nonetheless makes the list of nouns alone infinite.
6
See, for example, Chomsky 1981a.
7
See Mah. 1.94; Joshi and Roodbergen 1986:136 and Chatterji 1964:78–9, where the
mahāsattras are discussed in detail.
8
This probably includes, but apparently is not limited to, infinitely long utterances:
aprayukte dı̄rghasattravat yady apy aprayuktās tathāpy avaśyam . dı̄rghasattraval laks.an.en-
ānuvidheyāh. (Mah. 1.94; Joshi and Roodbergen 1986:136; Chatterji 1964:79: “in the case of
unused [words], [it is] like long sacrifices. Though they are not used, they must necessarily
be taught by the rules of the Śāstras like protracted sacrifices.”). It is important to note that
these unuttered words are conceived of by Patañjali as grammatical, only not used in common
speech. They are thus not equivalent (and in a sense opposed to) the apaśabdas, which are
ungrammatical words mentioned earlier in the Mah. (1.46; Joshi and Roodbergen 1986:70;
Chatterji 1964:35).
11:50:13:06:14 Page 2
Page 3
Introduction 3
granted that rituals have such “grammars.” Thus, structure of sacrificial

ritual appears to be the prototype for language––not, as some modern
anthropologists would expect, the other way around.
It is perhaps not surprising that the analogy between linguistic and ritual
“grammars” is found in India, and that in modern times this line of thought
has been pursued primarily by Indologists. The interpretation of ritual along
linguistic lines in Patañjali’s grammatical treatise is in accordance with the
special status of grammar among the sciences in classical India, analogous to
the status of mathematics in the West since at least the Renaissance.9
Moreover, there exists a particular affinity between grammatical and ritual
literature in Sanskrit.10 This affinity is evident in an external similarity in form,
inasmuch as the two corpora share several technical terms and concepts.11
Thus, theoreticians in the grammatical and ritual schools, striving to for-
mulate rules that would encompass the systems they were describing as
economically as possible, resorted to similar methods of composition.
The similarity between the ritual and grammatical manuals consisting of
aphoristic rules (sūtras) is most striking in the use of meta-rules (paribhās.ā),
rules about the application of rules.12 These suggest a similarity not between
the unconsciously internalized underlying grammars of language and ritual,
but between the consciously formulated grammars (“grammar” in the sense
of a textbook) of each.
Let us now turn to the Vedic passage that serves as the epigram of this
chapter, and examine its treatment in the Mahābhās.ya (I.1.36):13
Four horns, three feet he has, two heads, seven hands he has. Bound in three
ways the bull bellows loudly. A mighty god has entered mortals.
9
See Staal 1989:453.
10
See Renou 1942 and Cardona 1990.
11
For example, composition in the form of sūtra; the use of anuvr.tti (the carrying over of
preceding rules); and a metalinguistic usage of nominal declensions to imply logical statuses
and relationships between components of the rules. See Renou 1942:445 and Hastings 2003,
particularly 280; the use of adhikāra (“governing-rule,” influencing a number of succeeding
rules), which was borrowed from ritual to grammar according to Renou 1942:445, and the use
of vā (“or”) in a technical sense to indicate “optional” (Renou 1942:446; see also 452–5, 458).
12
Renou 1942:444. Inspired by this similarity, in his proposed model of a grammar of ritual,
Mishra (2010:95 n. 21) demonstrates the usefulness of As.t.ādhyāyı̄ 1.3.10, yathāsam . khyam
anudeśah. samānām, for the grammar of ritual: if two lists have the same number of elements,
then the elements of the following list correspond respectively to the elements of the previous
list.
.
13
The Vedic text quoted in Patañjali’s work is indicated here in bold as follows: catvā´ri śŕ.ngā
´ ´
tráyo asya pādā / dvé śı̄rs.e saptá hástāso asya // trídhā baddhó vr.s.abhó roravı̄ti / mahó devó
.
mártyām ā´ viveśa.
.
catvāri śr.ngān.i catvāri padajātāni nāmākhyātopasarganipātāś ca / trayo asya pādās trayah.
kālā bhūtabhavis.yadvartamānāh. / dve śı̄rs.e dvau śabdātmānau nityah. kāryaś ca / sapta hastāso
asya sapta vibhaktayah. / tridhā baddhas tris.u sthānes.u baddha urasi kan.t.he śirası̄ti. See Joshi
and Roodbergen 1986:11–12 (Sanskrit section), 52–3; see also Cardona 1990:12 and notes;
in Chatterji 1964:18–19 (Sanskrit section) the penultimate line, saptá hástāso asya––sapta
vibhaktayah., is missing.
11:50:13:06:14 Page 3
Page 4

Four horns––the four classes of words, noun, verb, upasarga, and particle
Three feet––the three divisions of time, past, future, present
Two heads––the two natures of words, “permanent” and “to be produced”
Seven hands––the seven case endings
Bound in three ways––formed in three places, the chest, the throat and the head.
The passage is structured as a commentary on a verse from the R.gveda
(4.58.3, represented above in bold), which describes a deity, perhaps
Soma,14 in zoomorphic terms. Patañjali interprets these monstrous attributes
allegorically, as referring to aspects of the Sanskrit language. The result is an
awkward isomorphy between a bellowing, bound bull and the language to
which Patañjali’s grammatical treatise is devoted: its four horns are the four
parts of speech identified by traditional Sanskrit grammarians; its three feet
are (roughly) the three tenses; its two heads are two types of denotation,
or two distinct lexical corpora;15 its seven hands are the seven nominal
declensions;16 and as spoken language is bound to the three physical organs of
speech (chest, throat, and head), the bull is bound threefold.
Having created a correspondence between the parts of this bull and various
aspects of the Sanskrit language, Patañjali proceeds to expound on the moral
of his allegory. This moral, according to Patañjali, is itself allegorically
embedded in the last part of the Vedic verse: “A mighty god has entered
mortals.” The mighty god is identified with Speech/Sound (śabda),17 and the
mortals who compose Patañjali’s audience are encouraged to study grammar,
for, as Patañjali suggests, “we must study grammar in order to become like the
mighty god.”18 How precisely the study of grammar is conducive to divinity
is not explicit,19 but Bhartr.hari, a later grammarian referring to this same
14
As Geldner (1951:488) notes, this hymn is notably obscure, and there is some disagree-
ment concerning the deity or deities referred to throughout the hymn. According to Joshi and
Roodbergen, the hymn identifies the ghee (ghr.ta) used in sacrifice with Soma, which is
described as a bull; the sound of the soma trickling into a jar is taken as the bellowing of the
bull. This is one example of the sacrificial connotations of the image of the bull. For other
interpretations of this image, see Joshi and Roodbergen 1986:36, n. 163 and the bibliography
cited there.
15
Concerning the distinction between nitya and kārya, see Cardona 1990:12, and 19 n. 51;
and see Joshi and Roodbergen 1986:52 n. 165 and 53 n. 170; both list three different interpret-
ations of this dichotomy offered by Bhartr.hari.
16
But see Cardona 1990:19 n. 52; Joshi and Roodbergen 1986:53–4 nn. 172, 173.
17
See Cardona 1990:18.
18
mahān devah. śabdah.. martyā maran.adharmān.o manus.yāh.. tān āviveśa. mahatā devena
nah. sāmyam . yathā syād ity adhyeyam . vyākaran.am. See also 1.42, Joshi and Roodbergen
1986:64–5.
19
The idea is probably already embedded in Patañjali’s commentary on the word roravı̄ti
(“bellows”). Patañjali notes that this verb is equivalent to śabdam . karoti (“makes sound”). At
first sight, this appears to be merely a literal lexical interpretation of the word roravı̄ti. However,
if considered as part of the allegory, this short comment may imply that the grammar-bull
“makes” (karoti)––in the sense of “causes one to become” (as at RV 10.16.6)––the deity
“Sound” (śabda).
11:50:13:06:14 Page 4
Page 5
Introduction 5
allegory, explains, “Attainment of faultless speech is the attainment of

Brahman. He who knows the secret of its functioning enjoys the immortal
Brahman.” The study of grammar is, apparently, one way out of the
lamentable cycle of births and rebirths.20
Carrying Patañjali’s allegory to its natural conclusion, it is hoped that the
present study, like the roaring beast and the grammatical allegory wherein
it figures, will somehow be conducive to a better understanding of the
labyrinths of the minds of the creators of sacrifice.
1.2 MAIMONIDES’ GENERALIZATIONS
One example may demonstrate the applicability of Patañjali’s insights to a

corpus of texts that are geographically and linguistically distant from the
Vedic sacrificial manuals. In the twelfth century, the Jewish scholar Moses
Maimonides formulated several general rules about the sacrificial laws in
biblical and rabbinic traditions. One cluster of these abstractions, consisting
of Maimonides’ own generalizations as well as generalizations culled from
earlier rabbinic literature (e.g. b. Tem. 14a), appears towards the end of his
introduction to tractate Zebah.im, in his Commentary on the Mishna:21
SYL IA D'GT TNAP AHRK'D ODQT YTLA TWNBRQLA EYM'G TYRQTSA A'DAW
HBQN DYXY TAUX LK IA D'GT VL'DKW . . . H'GWB HBQN RWBCH TWNBRQ EYM'G YP
ALW IA'CLA EWN IM TAUX RWBCH TWNBRQ EYM'G YP SYL IA IYBY VL'DKW . . .
. . . ZEAMLA EWN IM HLWE
And if you consider all of the abovementioned sacrifices, you will find that there
is not a single female among all of the public offerings. . . Similarly you will find
that every layperson’s purification offering is a female animal. . . Similarly it will
become evident to you that there is not among all of the public offerings an ovine
[i.e., sheep] purification offering, nor a caprine [i.e., goat] wholeburnt offering. . .
Note that many of Maimonides’ generalizations are found nowhere in the
ancient sources he had at his disposal; they are Maimonides’ own abstrac-
tions, designed to accord with several examples of sacrifices within the
biblical text (as interpreted by the early rabbinic authorities) and with a
number of sacrificial combinations newly introduced in rabbinic traditions
but not found in the biblical text. Yet of the many dozens of sacrificial
combinations found in late Second Temple literature, including Qumranic
20
Bhartr.hari, VP, 1.131–2. The translation is according to Pillai 1971:30. Bhartr.hari refers to
this allegory in his commentary on Patañjali, the Mahābhās.yadı̄pikā, 4.8.1 (Bronkhorst
1987:58). Concerning the soteriological aspect of the study of grammar, see Cardona 1990:19.
See also Shulman 2005:376, who sees “nothing metaphorical” about the statement that
“Grammar is meant to turn the grammatically informed speaker into God.”
21
Kafah. 1963:19. See nn. 000–000 below for philological comments.
11:50:13:06:14 Page 5
Page 6

literature, Josephus, Philo, and a large body of apocryphal and pseudepi-
graphic literature––combinations that differ substantially from the instruc-
tions of the pentateuchal law––not a single one substantially violates these
rules formulated by a twelfth-century Spanish-North African rabbi. Most of
these texts were unavailable to Maimonides, so we can only suppose that he
has hit upon certain rigorous, underlying rules, the same rules which the
authors of the late Second Temple period had internalized as a result of their
exposure to the biblical text, and which guided them when generating their
own new sacrificial combinations.
Consider a common linguistic analog: young children hear grammatical
utterances from their parents. The children are exposed to numerous––but
finitely many––grammatical utterances. The fact that these children regularly
acquire fluency in a language on the basis of that finite number of utterances,
and in particular that they can then produce grammatical but as-yet-unheard
utterances, demonstrates that they have internalized grammatical rules from
a limited input. A linguist can compose a grammar for that language on the
basis of those utterances of one speaker, a grammar that would then prove
valid for the utterances of other speakers as well even though the linguist had
not heard them.
So too, the Jewish sources of antiquity that survive today were exposed
to (roughly) the same pentateuchal text, and some of them presumably
witnessed similar practices in the Temple. These authors describe rituals that
do not appear in the sources they would have read but that nevertheless
accord with general rules they had internalized from their exposure to a finite
set of biblical rituals. Writing in the twelfth century, Maimonides (like the
aforementioned linguist) postulated generalizations in order to characterize
the pentateuchal and rabbinic literature, but his rules also prove valid for
other rituals to which he could not have been exposed. In short, Maimonides
was describing a “grammar” of rituals that earlier practitioners and authors
had internalized.
1.3 MODERN ANTHROPOLOGY
Outside India, the notion that there exists a substantive affinity between the
structure of language and that of sacrifice was at best vaguely intuited, but
never explicitly addressed until recent times. Thus, despite the elaborate and
versatile usage of the grammatical metaphor in European traditions, there
is hardly any evidence of a conception that rituals, like natural languages,
have “grammars.” Instead, the analogy between language and ritual focused
on other points of contact between the two systems, such as their respective
modes of signification, the role of substitution in both systems, and the use of
11:50:13:06:14 Page 6
Page 7
Introduction 7
verbal utterances in the place of sacrificial deeds.22 The idea that there exists
a similarity between the rules governing the formation of linguistic and ritual
sequences emerged in the Western tradition relatively recently––as we shall
see, it has its roots in the school of the Année sociologique––and it is to this
particular analogy that we will direct our attention.
In recent years, the scholarly investigation of ritual has been pursued
within a large variety of theoretical frameworks, including linguistics,
information theory, systems theory, ethology, and cognition.23 Of these,
linguistics has been the single most important framework for the study of
ritual. In fact, several of the other frameworks employed in this undertaking
are modeled upon, or at least inspired by, linguistics.24
Within the theoretical framework of linguistics, there has been a tendency
among anthropologists and scholars in comparative religion to liken the
structures of ritual systems to the grammars of natural languages. This has
resulted in the widespread use of grammatical terminology to describe the
structures of ritual systems, and even in the postulation of an evolutional
relationship between ritual and language;25 it has also resulted in a conviction
that rituals, like languages, have grammars.26 The literature on this theoretical
issue sometimes singles out sacrificial ritual as a special case, perhaps because
some scholars believe that sacrifice holds a unique place among rituals, as if it
22
For example, in a survey of the use of the grammatical metaphor in the Middle Ages,
Alford (1982) finds that “grammar” was used in Western Europe as a metaphor for social
order, for sex, for nature, for the ways of God––but not for sacrifice (but see his brief note on
Isaac’s sacrifice, p. 737). Mazzotta finds traces of a conceived grammar of sacrifice in Dante’s
Paradiso, and traces this conception through Thomas Aquinas back to St Augustine (Mazzotta
1993:34–55, particularly p. 45). However, here the analogy does not pertain to the formal
structure of these systems, but to their modes of signification: sacrifice is viewed as a symbolic
mode of signification modeled after language. Just as in a spoken language a metaphor may
substitute for direct representation, so in sacrifice, the sacrificial victim substitutes for the
sinner. See already the view of Dionysius of Halicarnassus as interpreted in Petropoulou
2008:42. On the replacement of sacrifice with verbal sequences, see for example Stroumsa
2009:62–70. A vague reference to an analogy between one branch of grammar and
religious ritual is found in Varro, who likens the highest level of knowledge of the “origins
of the individual words” to the mysteries of the High Priest (De Lingua Latina 5.7–8, Kent
1938:8–9).
23
See Staal 1979, 1980, 1989 (linguistics); Lévi-Strauss 1963c (information theory); Gane
2004, 2005 (systems theory); Burkert 1987 (ethology); and Lawson 1976 and Lawson and
McCauley 1990 (cognition). For a sense of the proliferation of such approaches in the past few
decades, see Kreinath 2006.
24
For example, Lévi-Strauss’s use of information theory is related to his conception of
language, and Lawson’s use of cognition is modeled upon Chomsky’s idea of I-languages.
25
Staal 1989:262.
26
See for example Ferro-Luzzi 1977; Lawson and McCauley 1990; Payne 1999; Gruenwald
2003:176; Gane 2005; and the various contributions to Michaels and Mishra 2010. These are
only a sampling of a large body of literature from diverse cultural contexts. See also below,
n. 000. The idea, according to Michaels 2010, was first suggested by Franz Boas. As Michaels
notes, it is Susanne Langer who associated this idea with Boas in Philosophy in a New Key
(Langer 1948:40). However, one must admit that Langer’s reading of Boas in this regard is
rather generous (cf. Boas 1911:198–9, and see also p. 143).
11:50:13:06:14 Page 7
Page 8
were the ritual par excellence, to which all other rituals may ultimately be
reduced.27
At present, current discussion in the field of comparative religion about
the grammar of ritual greatly outweighs the descriptive literature that ought
to serve as the foundation for such discussions. In fact, despite the relative
abundance of detailed discussions of grammatical features within particular
rituals, a systematic outlay of even a single ritual system’s grammar is entirely
lacking in the literature.28 It is striking that there has even been some dis-
cussion of a “universal grammar” of sacrificial ritual––rules that may be
applicable to all sacrificial systems in diverse human societies––even though
not one complete grammar of a specific sacrificial system exists29. This
paradoxical situation is analogous to an attempt to describe a Universal
Grammar (UG) of language before composing a single grammar for a natural
language, such as Attic Greek or modern English.
This book is the result of a preliminary attempt to formulate one grammar
for one particular ritual system. The ritual system under investigation is
the ancient Israelite ritual sacrificial system (hereafter also referred to as Σ )
as represented in a set of biblical texts that are commonly referred to as
“Priestly”––essentially, several detailed sacrificial texts found in the books of
Exodus, Leviticus, and Numbers.30
Stated simply, underlying this study is the basic question, “If ritual systems
have ‘grammars’, what do these ‘grammars’ look like?”––or, more specifically,
“If Σ has a grammar, does its grammar resemble those of natural languages?
If not, what are its operative categories, and in what sense is it a ‘grammar’ at
all?”
It is important to stress that the category “ritual sacrifice,” as a subject of
scholarly inquiry, is not clearly demarcated, and does not reflect a self-
contained system of activity that is marked off, from a practical perspective,
from other activities that can take place by an altar, in a butcher shop, or in
the kitchen.31 Moreover, the English terms “ritual” and “sacrifice” received a
27
See Milbank 1995 and the bibliography cited there, to which one might add Evans-
Pritchard 1965:53; Sperber 1975a:110–11; Smith 1987:196–7; Burkert 1987:212. For a criticism
of the predominance of animal sacrifice (over vegetable sacrifice) in modern scholarship, see
McClymond 2008; for a brief historical overview see Naiden 2013:3–15.
28
See Staal 1989:447; as well as Lawson and McCauley 1990:56, quoting Roy Rappaport’s
criticism of Leach: “Leach does talk about the similarity between the grammatical rules of a
language and the grammatical rules of symbolic-cultural systems [. . .], but he nowhere suggests
what such rules might look like. This is equally true of Turner, who clearly states his intention to
develop both techniques for decoding symbols and categories for describing their internal
relationships [. . .], but fails to do so.” More recently, more concrete steps have been taken in this
direction. See for example Michaels 2012 and below, Section 1.3.4.
29
Michaels 2010:7.
30
Material from the “Priestly” texts is also found outside these books, and the narrative
sections of P are often essential for understanding the legal and cultic prescriptions; the latter
are nevertheless all concentrated in Exodus, Leviticus, and Numbers. See § 1.4.1.
31
See Margueron 1991:236; Detienne and Vernant 1989.
11:50:13:06:14 Page 8
Page 9
Introduction 9
great amount of scholarly attention in the past, and still serve as a productive
driving force in recent scholarship on ancient and modern religious
phenomena;32 however, these terms do not correspond perfectly with any
particular set of terms in the texts under examination.33 Despite these caveats
we shall see that in the case of the ancient Israelite texts it is useful to center
on a set of activities here considered the core of Σ, namely the offering of
materia sacra derived from animals upon an altar.34
Within modern anthropology, the “grammatical” analogy between
ritual and language has undergone several developments, from its earliest
intimations over a century ago to its extensive use in the 1960s, followed by its
more rigorous application in the last decades of the twentieth century and a
more skeptical approach in the last decade.35
While a complete reconstruction of the history of the concept of “ritual
grammar” will not be offered here, let us examine the most important stages
in the development of this concept in order to observe what scholarship has
achieved thus far and why it has arrived at what appears to be an impasse.
1.3.1 Hubert and Mauss
In its modern anthropological and comparative context, the association of

sacrificial ritual with language owes much to the work of Hubert and Mauss
on sacrifice, first published in 1898.36 In an attempt to reconstruct a general,
universal scheme of sacrifice, Hubert and Mauss abstract structural properties
common to sacrificial rites recorded in diverse cultural traditions, particularly
the biblical and Vedic.
Hubert and Mauss do not attempt a description, even in the abstract,
of these two systems; rather, their comparative study is geared towards
32
See, for example, Stroumsa 2009; Knust and Várhelyi 2011; Halbertal 2012. A recent
historical overview can be found in Naiden 2013.
33
See for example the use of IBRQ in Num 31:50.
34
This is not to say that the offering of animals holds theoretical prominence over the
offering of non-animal materials (see Eberhart 2011:23–5; McClymond 2008). However,
animal sacrifice lends itself particularly well to an economic “grammatical” description––its
permutations are highly restricted, and the number of praxemes it involves is strictly limited. In
later traditions, vegetable offerings are explicitly modeled upon animal sacrifice (compare m
Men 1:1 with m Zeb 1:1 and m Men 2:1 with m Zeb 2:2). A more comprehensive “grammar”
may take account of a wider array of activities, such as will become apparent in due course.
The extent to which other activities belong to Σ may be one of degree rather than absolute
association––as in the problematic cases of the red heifer, the goat dispatched to the desert, and
other ritual activity discussed in Chapter 5.
35
An outline of this idea’s history in modern anthropological studies is also found in Lawson
and McCauley 1990:32–59. See also Lawson 1976, Kreinath et al. 2006, and in particular Severi
2006 and Kreinath 2006:468.
36
The work (Hubert and Mauss 1964) was published in L’Année sociologique in 1898 as
“Essai sur la nature et la fonction du sacrifice.”
11:50:13:06:14 Page 9
Page 10
discovering certain universals of sacrifice. If a particular structure is shared by

several rituals in two historically unrelated traditions, they reason, it is
most likely due to some inherent property of sacrifice, and therefore likely
to be found in sacrificial rituals outside of these two cultures as well;37 since
sacrifice as they define it38 is an all-but-universal institution, these structures
are apparently universal.
Their work, though provocative and influential when first published, has
been criticized for its methodology39 and an abundance of factual errors.40
Furthermore, the underlying structure that Hubert and Mauss discover is
extremely rudimentary and has a very low explanatory power.41
Hubert and Mauss do not use explicit linguistic terminology in describing
the sacrificial systems that they study, but rather call the underlying structural
properties that they abstract from these sacrificial systems a “schème
abstrait.”42 It is only Evans-Pritchard, in his foreword to the 1964 English
translation of their work, who with great hesitation suggests the appellation
“grammar.”43 Though Evans-Pritchard himself attempts to describe an
abstract scheme underlying Nuer sacrificial rituals,44 he makes no explicit
attempt to identify his own scheme as a grammar, akin to that of a natural
language.45
37
Hubert and Mauss 1964:8 (1929:9). The question of a possible common origin for
(possibly pre-linguistic) Indo-European and Semitic rituals need not be resolved here, as the
data are sparse and the hypotheses highly speculative. There is evidence of certain typological
affinities between Israelite, Hurrian (non-Indo-European), and Hittite (Indo-European) rituals,
particularly from Kizzuwatna, as well as straightforward evidence of an affinity between the
Vedic ritual system and some Anatolian practices. See Weinfeld 1983, Moyer 1983, and Feder
2011.
38
Note that their definition, “Le sacrifice est un acte religieux qui, par la consécration
d’une victime, modifie l’état de la personne morale qui l’accomplit ou de certains objets
auxquels elle s’intéresse” (1929:13), is very broad.
39
For example, a professed uncritical attitude towards the biblical text (1929:9–10 n. 1)
and the implicit assumption that rites as described in the ritual manuals correspond to what
practitioners actually practiced. See Smith 1982 and Feeley-Harnik 1981:2–6.
40
See Gruenwald 2003:182 and Gane 2004:339–45. Among these errors are the misunder-
standing of the opening verses in Leviticus 4 as referring to the Day of Atonement (Hubert and
Mauss 1929:34; 1964:36), the statement that for OYM im
Lv xeZ there is no hand-leaning (Hubert
$o XB
and Mauss 1929:47; 1964: 52), and the mistranslation of HP miWNT
v as a “turned offering” (offrande
d
tournée, Hubert and Mauss 1929:35; see 1964:36). More importantly, central to their treatise
is an unproven assumption that “there is in the victim a spirit which it is the very aim of the
sacrifice to liberate” (Hubert and Mauss 1964:30, 32). However, in the Israelite Priestly tradition,
the immolation of the animal is a preliminary act that does not necessitate an officiating priest
and may be performed by laypersons. This suggests that at least within the Israelite system, the
immolation of the animal is by no means the center of the sacrificial rite. See below, n. 000.
41
This has been expressed somewhat bluntly by van Baal, and subsequently by Staal,
according to whom Hubert and Mauss prove that “every rite has a beginning, middle, and end”
(see Staal 1979:15). For a more generous view, see Tambiah 1979.
42
Hubert and Mauss 1929:17.
43 44
See Hubert and Mauss 1964:viii. Evans-Pritchard 1956:197–230, 272–86.
45
Evans-Pritchard’s own terminology is non-linguistic. Aside from the Nuer terminology
quoted in his work to describe the components of the ritual acts as termed by practitioners,
11:50:13:06:14 Page 10
Page 11
Introduction 11
1.3.2 Lévi-Strauss, Structuralism, and Neo-Structuralism
The application of linguistic concepts to the study of non-linguistic cultural

phenomena was first systematically pursued by Lévi-Strauss.46 Using struc-
tural linguistics as a model, Lévi-Strauss attempts to understand kinship
systems, myths, and rituals, all of which he views as semiological systems,47 as
a branching of binary oppositions that he assumes result from the structure
of the human brain.
One important application of this analogy concerns the modes of signifi-
cation in non-linguistic systems. Durkheim and Van Gennep have already
noted that the semantic value of an individual ritual act is never straight-
forward, but rather is always dependent on its place within a system that
includes other ritual acts with which it stands in complex logical relations.48
More radically, Lévi-Strauss argues that the “elementary units” of non-
linguistic semiotic systems (such as his “mythemes”) are devoid of semantic
value altogether, and that meaning is assigned only to combinations of
these “elementary units.” In this respect, they resemble phonemes, or psycho-
logically recognized sound units, in natural languages. This view has been
reformulated succinctly by Leach:
Notice how the same elements of ritual behavior keep on recurring, but linked
together in different combinations and different sequences. The elements are like
the letters of the alphabet; in different combinations they can be made to say
different things.49
Leach’s identification of “elementary units” with “letters” rather than with
phonemes as in Lévi-Strauss’s writings is somewhat controversial,50 but the
Evans-Pritchard refers to logical relationships pertaining to combinations of these components
with general terms such as “variations” (1956:220) and “supernumerary acts” (p. 208). These
are not conceived of as analogous to syntactic relationships within a natural language (e.g.,
p. 215: “I do not think that this reversal of phases has in itself any special significance”).
46
Lévi-Strauss 1963c. Though the bulk of Lévi-Strauss’s work concerns the study of
mythology rather than ritual, his earliest work pertains to ritual in a more general sense of the
term (see Lévi-Strauss 1958:69: “kinship systems are languages with women as words”), as does
his work on totemism (Lévi-Strauss 1963) and the finale to The Naked Man (Lévi-Strauss 1981).
A number of passages in The Savage Mind (Lévi-Strauss 1966:223–8) are dedicated to the inner
logic of sacrifice.
47
Lévi-Strauss considers these non-linguistic systems to be structurally analogous, and
apparently evolutionally related to language. However, he does not suggest that linguistic
sequences are translatable into non-linguistic sequences or vice versa (1976:198–210); see also
Kunin 1995:19 and n. 3.
48
See Staal 1989:125–51. A careful reading of Hubert and Mauss (1964:59–60) also leads
to this conclusion. This idea is found later, in the wider context of symbolism, in Turner’s
category of “positional meaning” (1967: 50–1, quoted in Sperber 1975a:63).
49
Leach 1976:88.
50
See Matthews 2001:81, but cf. 79. Concerning the indiscriminate Alexandrian usage of the
term στοιχεα to designate both graphemes and sound units, see Swiggers and Wouters
1996:150.
11:50:13:06:14 Page 11
Page 12
analogy to language, here in a form of graphic representation, is retained:

in ritual, according to the Structuralists, the smallest meaningful units are
themselves composite units, just as meaningful units in language comprise
smaller (meaningless) sound units.51
The analogy to linguistics was applied by some Structuralists more
vigorously than rigorously. In structural linguistics, clear criteria have been
developed to identify elementary units of sound and meaning. By contrast,
Lévi-Strauss’s conception of “elementary units” and their combinations
within non-linguistic semiotic systems, so central to his writings, is at times
too vague to be verifiable or conducive to further investigation. In fact, his
writings contain no clear criteria for identifying these “elementary units,” and
it is even less clear what semantic value the complex, combined structures
would have.52
Lévi-Strauss’s approach has been instrumental in countering the crude
conception of ritual acts as straightforward significations of non-ritual ideas,
values, or statements, though this conception is still found in much of the
interpretive literature.53 However, his portrayal of ritual systems as sets of
binary oppositions is far too rudimentary a description to constitute a
“grammar” of any such ritual system. At most, it is a universal structural trait
shared by ritual, myth, kinship systems, natural languages, and presumably
other systems that Lévi-Strauss views as semiological.54
One of the contributions of Neo-Structuralism, in contrast to the classical
structuralist approach, is its consideration of a number of factors that Lévi-
Strauss did not take into consideration sufficiently, such as special attention
to culture-specificity of some underlying structures, sensitivity to diachronic
change, cultural diversity within a single societal unit, and the existence of
differential levels of underlying structure.55
51
See Hockett 1960.
52
Both Penner and Sperber argue that Lévi-Strauss’s insistence on the existence of some “last
resort” meaning, located outside mythology, is unnecessary, and that it in fact contradicts his
own discoveries. See, in general, the strong criticism in Geertz 1973:3–30. For a more favorable
evaluation, see Paz 1970; Sperber 1975a:51–84; and Goody 1977:5–8.
53
In biblical studies, there is clearly a struggle within the Milgrom School concerning this
issue. For example, contrast Milgrom, Leviticus 1.856 with Leviticus 1.643–742; two examples
of a struggle evident within a single paper are Olyan 1998:621–2 and Wright 1986 (contrast
pp. 433–4 with 435–6). For an approach that is sensitive to the complexity of the problem at
hand, see Marx 2005:123 on the “connexions” made by blood in various contexts. For a discus-
sion of this problem in the interpretation of Vedic ritual, see the opening passages of Bodewitz
1976.
54
If Lévi-Strauss’s works consist of anything comparable to a grammatical rule in natural
languages, it would be the canonical formula presented in 1963b:230–1 and elaborated upon
in Lévi-Strauss 1988. This canonical formula has been extensively applied––with varying
degrees of success––by Lévi-Strauss himself and his followers (e.g., Sperber 1975a). However,
its application is disturbingly unpredictable, and it is precisely this pseudoscientific aspect of
Lévi-Strauss’s theory that has caused his writings to fall out of favor in some schools of thought.
55
See Kunin 2005:204–8; on the differentiation between levels of underlying structures see
Kunin 2004:5–28, and briefly Kunin 1995:12–13.
11:50:13:06:14 Page 12
Page 13
Introduction 13
Kunin, for example, demonstrates how these factors are relevant to a study
of biblical rituals, in particular of incest laws and dietary taboos.56 Moreover,
he stresses the importance of conscious agency, particularly with regard to
diachronic shifts in underlying structures. This last point is relevant to our
own study, since as we shall see, it is likely that some of the changes in rituals
described here were introduced consciously by literate intellectual elites.
Thus, they differ significantly from changes in the grammars of natural
languages, such as phonological shifts, which are for the most part entirely
unconscious.
While structuralist anthropology contributed considerably to the idea of
ritual as a “language” with a “grammar,” the more immediate impetus for
a rigorous, detailed analysis of ritual in “grammatical” terms began with
the work of Frits Staal, who applied a generativist model in linguistics to
the study of ritual, and reintroduced the project envisaged by Patañjali into
modern scholarly research.
1.3.3 Frits Staal and the Shift to Generativism
The analogy between language and ritual has often been characterized by
a tension between two focuses of research, resulting in a spectrum of
approaches from examining meaning to investigating formal structure.57 On
the one hand, theorists interested in the semantics of ritual, such as Turner
and Geertz, have tended to focus more on the modes of signification in ritual,
under the assumption that ritual entails a relationship between a physical
ritual signifier and a non-ritual signified. On the other hand, theorists inter-
ested in the formal structure of ritual, such as Leach and Tambiah,58 have
56
Kunin 1995, 2004, 2005.
57
See Bell 1992; Houben 2010 refers to two competing traditions in ritual studies––the one
going back to Hubert and Mauss and interested primarily in structure, the other harking back
to Tylor and Frazer and focusing primarily on meaning (Houben 2010:30 n. 8). While this
dichotomy is useful for general orientation, many important studies pay attention in both
structure and meaning; see, for example, Turner 1967; Geertz 1973; Leach 1976; and Tambiah
1979, on which see immediately below.
58
Tambiah, however, does not abandon the notion of symbolic signification. Staal
(1989:238) also notes approvingly that Tambiah establishes “that the exclusive emphasis on
meaning had become an obstacle to an adequate understanding of ritual, and that ritual acts
had to be understood in terms of rules.” However, see Tambiah 1979:119 and, more explicitly,
his criticism of Rappaport and McLuhan (p. 138): “there is one extreme semiotic school
that supposes that form can be tackled apart from the presentation of contents and the inter-
pretation of symbolism.” Lawson and McCauley (1990:51) argue that Tambiah, Rappaport, Ray,
and Finnegan are correct in noting the analogy between speech acts and ritual acts. However,
it is paradoxical that Tambiah advocates a performative analysis of ritual under the influence of
Austin’s analysis of language, since Austin analyzes utterances, not deeds (see Austin 1975,
Searle 1980). Since rituals, unlike utterances, are acts to begin with, it is not surprising that they
are intended to perform changes in the world. See also Grimes 1990:191–4.
11:50:13:06:14 Page 13
Page 14
concentrated on the relationship between this structure and syntax in natural

languages.59
The theory of Frits Staal lies at the extreme latter end of this spectrum.
According to Staal, ritual activity lacks semantic value, so its analysis should
pertain to formal structure only.60 Staal offers such an analysis of several
Indian rituals, most notably the recitation of mantras and the agnicayana
ritual that he attended and recorded in Kerala in southern India in 1975.61
As opposed to Lévi-Strauss, for whom the analogy to linguistics primarily
relates to phonology, Staal’s analog pertains mostly to syntax. The reasons
for this shift are in part historical: for Lévi-Strauss, formulating his ideas in
the late 1950s and early 1960s, Structuralism was the leading framework
of thought in linguistics, and since a structural analysis based on binary
branching was most successfully demonstrated in the field of phonology,
Lévi-Strauss was led to liken the combination of ritual acts to the com-
bination of phonemes in natural languages. For Staal, however, writing
in the late 1970s and 1980s, Chomsky’s generative linguistics was the central
framework in linguistics. Perhaps it is because Chomsky’s generative
grammar was most elaborately applied in the realm of syntax that Staal was
led to attempt a syntactic analysis of ritual. So too, whereas Lévi-Strauss views
ritual systems as quasi-linguistic systems (“metalanguages”),62 Staal explicitly
claims that ritual is neither a language nor a system of communication, but
pure syntax, devoid of semantics.63 The justification for the linguistic–ritual
analogy then lies in the fact that both are rule-governed activities64 employing
transformational rules, some of which Staal finds strikingly similar.
59
According to Lawson and McCauley, it is often theoretically impossible to distinguish
between formal structure and modes of signification, since the latter, if they are present,
are necessarily determined by the former. Lawson and McCauley are to be commended for
attempting to formulate a rigorous theory of ritual that can be tested empirically by asking
informants about what a possible ritual would look like. However, judging from the single
example that they offer (1990:175), the results are not entirely satisfactory in this regard. In fact,
upon their own admission, their thesis presented in chapter 5 is “quite commonplace” (p. 6).
60
See, for instance, Staal 1989:451 and the criticism of Penner 1985:1–2. Staal admits that
some practitioners and theologians may attach significance to ritual acts, but this does not
imply that meaning inheres in the rituals themselves. Moreover, these superordinary attached
meanings are, in Staal’s view, in constant flux, in contrast to the ritual acts themselves, which
remain relatively stable over comparatively long periods of time (p. 125). The opposite view,
however, is espoused by Levine 1974:45–52, who implies that the basic meanings of some rituals
can be relatively stable even though they undergo many formal changes over time.
61
Staal 1983.
62
Lévi-Strauss even plainly refers to ritual systems as languages, although this is probably
meant as a metaphor; see Lévi-Strauss 1958:69.
63
Staal 1989:162, 188, 260.
64
Staal 1989:451. Furthermore, many of the rituals Staal analyzes, particularly in Rules
without Meaning, are mantras. Since mantras are themselves utterances (though not linguistic
utterances, according to Staal) to which certain linguistic rules clearly apply––e.g., phonological
rules––it is reasonable to suspect that these complex utterances may be governed by a syntax as
well.
11:50:13:06:14 Page 14
Page 15
Introduction 15
In an article published in 1979, “The Meaninglessness of Ritual,” Staal

argues that ritual systems are based on transformational grammars akin to
those that Chomsky identifies in natural languages.65 Staal offers an idealized
abstraction of the agnicayana ritual that he recorded in Kerala in 1975,
which involves rules pertaining to what he calls “modification induced by
embedding.” Stated briefly, an agnicayana (“fire-heaping”) ritual consists of
darśapūrn.amāsa (“new- and full-moon”) rituals, paśubandha rituals (literally
“animal-binding,” and by extension “animal sacrifice,” though no animals
are used in this case), among several other rituals. Each darśapūrn.amāsa
ritual in turn includes the recitation of a number of sāmidhenı̄ verses (to be
recited during the kindling of the fire). Each paśubandha ritual also consists
of several rituals, two of which are themselves darśapūrn.amāsa rituals. Thus
the darśapūrn.amāsa is repeated several times within this ritual complex,
first as an element immediately embedded in the agnicayana and again as
an element that is doubly embedded (within a paśubandha that is itself
embedded in the agnicayana).
What Staal finds striking is that practitioners intuitively know to modify
the darśapūrn.amāsa whenever it is doubly embedded––and not when it is
immediately subordinate to the agnicayana. This modification pertains only
to the number of sāmidhenı̄ verses that are recited, seventeen rather than
the usual fifteen, as in Figure 1. Here, d1* indicates the modified (seventeen-
verse) sāmidhenı̄ recitation, rather than the regular (fifteen-verse) recitation
d1 typically found in darśapūrn.amāsa (D). The transformation will occur
Fig. 1. Structure of an agnicayana.

A = agnicayana
D = darśapūrn.amāsa
P = paśubandha
a1, a2, a3, a4 = surface-level constituent rites of agnicayana
d1, d1*, d2, d3 = surface-level constituent rites of darśapūrn.amāsa
p1, p2, p3, p4 = surface-level constituent rites of paśubandha
65
There is essentially nothing new in Staal’s reformulation of his theory from 1979 one year
later in “Ritual Syntax” (Staal 1980). Though Staal later addresses the same topic in greater
detail (see particularly Staal 1989), his example in “The Meaninglessness of Ritual” remains the
strongest and most compelling formulation of his argument.
11:50:13:06:14 Page 15
Page 16
invariably, even though the ritual sequence is carried out in the uninterrupted
linear form d1, d2, d3, a1, a2, p1, d1*, d2, d3, p2, d1*, d2, d3, p3, p4, d1, d2, d3,
a3, a4 (corresponding to the bottom line of Fig. 1) and there are no external
indicators of this embedding.
An obvious analogy, one which Staal may have had in mind, is found in
certain languages in which the morphology of verbs is affected by syntactic
considerations. For example, one form of indirect speech in modern German
entails verb forms that are sometimes morphologically distinct from the
indicative, as in the following example:
(a) Er kommt (“He comes” or “He is coming”)
(b) Er sagt, dass er komme (“He says that he comes/is coming”)
In this case, the morphology of the verb is affected (komme rather than
kommt) by the fact that the verb in sentence (b) appears within indirect
speech. In this particular case, one might argue that this morphological
change is due to considerations of meaning––for instance, that the modality
of komme is a result of the speaker’s insecurity as to whether the person is
really coming. However, consider the following example of the Akkadian
“subjunctive,” where no such argument is possible:
(c) awı̄lum ikrub (“the man prays/is praying”)
(d) awı̄lum ša ikrubu (“the man who prays/is praying”)
Here, the morpheme -u is introduced for purely syntactic considerations––
namely, because the verb is found in a relative clause.
As in the ritual example formulated by Staal, the modification in such
phrases results from the embedded status of the phrase and pertains only to
one aspect of the phrase, the morphology of the verb. In Staal’s view, such
similarities between linguistic and ritual structures are striking enough to
suggest an evolutionary relationship between them. Since Staal believes that
ritual is meaningless, it appears that ritual is located on a more primitive
evolutional rung than language, which has both syntactic and semantic prop-
erties. Thus, Staal believes that the capacity for syntactic arrangements
evolved in humans first in the realm of ritual (probably in the recitation of
mantras), only later to be employed in meaningful verbal utterances. If Staal
is correct, then one might say that some hypothetical humanoids engaging
in ritual but lacking language are the missing link between organisms that
lack the linguistic faculty altogether and Homo sapiens: these hypothetical
creatures are endowed with the syntactic faculty but not the semantic.66
Staal’s theory has been harshly criticized on a number of grounds, most
notably the conjectural nature of his evolutionary hypothesis, the question-
66
See Staal 1989:262, 280. As his analogy to biological evolution of fins and wings suggests,
the process is somewhat more complex.
11:50:13:06:14 Page 16
Page 17
Introduction 17
able relevance of his empirical data, and a number of inaccuracies.67

Furthermore, most of the rules governing ritual that Staal presents are dif-
ferent from those that govern the syntax of language. In fact, even the example
from the agnicayana discussed here is only roughly homologous, as we have
seen, to the examples drawn from natural languages, since in the latter only
one level of embedding is necessary for morphological changes. Staal does not
deny this dissimilarity between linguistic and non-linguistic “syntax,” and he
refers to it particularly in his analogy to music (which is also pure syntax
devoid of semantics, in his view);68 but in the context of ritual, he is somewhat
less eager to admit it. Even if these criticisms are valid, Staal’s theory neverthe-
less offers an important discovery for the study of ritual grammar. At the very
least, he has demonstrated a set of underlying generative rules restricting
the variability and dictating the form of a theoretically unlimited number of
possible ritual sequences.
1.3.4 Recent Scholarship and the Heidelberg Ritualdynamik School
Staal’s provocative work is not intended as a grammar of any ritual system;

it is comparable to a collection of sentences, each of which demonstrates
a certain grammatical rule. There is, for instance, no attempt to distinguish
different levels of ritual isomorphic with the syntactic, morphological, and
phonetic levels in a grammar of language.69 Staal points to the Āpastamba
Śrautasūtra,70 a semi-formulaic explanation of ritual restrictions, as an
ancient South Indian precedent for ritual grammars, but it is difficult to guess
on this basis what he thinks the structure of a modern “ritual grammar”
would be.
Recent scholarship since Staal, in particular within the Heidelberg Ritual-
dynamik circle,71 has been explicitly calling for the composition of a grammar
in the past few years. As Michaels notes, previous scholarly literature on the
subject has failed to supply such a grammar, being “more programmatic than
proto-grammatical.”72
67
See in particular Penner 1985. Note, too, Staal’s loose use of the term “syntax” with
reference to the formation of the word “mama” (Staal 1989:275).
68
Staal 1989:220.
69
At one point, Staal does hint at a multi-layered view of the analogy to language––namely,
of discovering several dimensions analogous to syntax, morphology, and phonology in a non-
linguistic system––but this is suggested in the context of music, not ritual (Staal 1989:331).
70
See Narasimhachar 1944 (critical edition) the German translation of Caland (Caland
1921), and Thite 2004.
71
See, for example, the works inspired by this circle in Kreinath et al. 2006 and the various
contributions to Michaels and Mishra 2010.
72
Michaels (2007:241) alliterates: “mehr programmatisch als protogrammatisch.”
11:50:13:06:14 Page 17
Page 18
Michaels states that if there is any future for the study of ritual as a
scientific discipline then real attempts to compose grammars of ritual systems
are a sine qua non. However, he also believes that this is impossible:
In keeping with the possible generativity of rituals and thus their potential
grammaticality, it is necessary, although probably not possible, to compile a
finite quantity of ritual elements––the ritual inventory––in such a way that,
together with an also finite number of formal rules, it produces an infinite
number of rituals and thus allows new rituals as well as deviations within the
ritual involved.73
Michaels’s assertion sums up a century-long attempt to convert the vague
intuition that ritual systems have grammars into a branch of scientific
investigation––an attempt he believes is doomed to failure, on grounds that it
is “necessary, although probably not possible” to compose such a grammar.
The grammar Michaels imagines is of the form [E, R], where E represents a
finite list of ritual elements (a ritual inventory) and R a set of formal rules
that determine which strings can be generated.
Michaels is right in that such a formulation is necessary, but unduly
pessimistic in asserting that it is “probably not possible.” Aside from the
problem of discretizing a continuously performed ritual into psychologically
real units––a process which he admits is not insurmountable74 ––Michaels
does not explain why composing a ritual grammar is impossible.
As we shall see, the grammar of Σ is somewhat more complex than a
simple [E, R] grammar. Like the grammars of natural languages, it consists of
distinct sets of rules for multiple operative categories that interact to generate
numerous sacrificial sequences.
1.4 METHODOLOGICAL CONSIDERATIONS
In this study we will see that the analogy between language and sacrifice, and
the notion that the two systems are isomorphic, has been a useful tool in
advancing our understanding of the phenomena of sacrifice and language,
and may further our understanding of these systems. In this sense, the present
work cautiously follows a tradition in the humanities and social sciences that
views linguistics as model for the study of other types of human activity.75
At the same time, investigation in the field of ritual grammar has reached
something of an impasse because this isomorphism has been applied
too rigorously, as scholars have been led to believe that some of the basic
categories of natural languages are to be found in non-linguistic systems
73 74
Michaels and Mishra 2010:12 (see Michaels 2007:244). Michaels 2007:244.
75
See Staal 1989:453 and Lawson and McCauley 1990:45.
11:50:13:06:14 Page 18
Page 19
Introduction 19
like ritual. On the contrary, it will be argued that ritual and language are
isomorphic only inasmuch as both are based on a finite set of generative rules,
in part unconsciously internalized, that are amenable to concise, formulaic
notation, such that illicit combinations are immediately discernible. The basic
operative categories of the grammar of Σ, as we shall see, are distinct from
those of linguistic systems.
Of the two dominant approaches to linguistics in the previous century, it
is particularly Generativism, rather than Structuralism, that has served as a
model for the present study (although the classification of sacrificial animals
is inspired by both schools). The reason for this is that this study attempts to
account for a phenomenon similar to that which Generativist linguists such
as Chomsky have tried to explain in natural languages: in certain systems,
the exposure to a finite set of sequences enables a practitioner to create an
unlimited number of new sequences, none of which he or she has been
exposed to in the past, and yet all of which comply with a set of rules that
have never been consciously formulated or taught.
1.4.1 The Siglum “P”
Throughout our discussion of biblical passages, I will use a number of sigla

that require some clarification. In biblical scholarship it is customary to use
the siglum “P” (from German Priesterschrift, “Priestly document”) to refer
to a body of literature within the Pentateuch that is believed to have been
composed by a priestly elite active in ancient Israel. The authors of “P” shared
a common idiosyncratic narrative pertaining to the history and pre-history
of Israel, certain unique technical terms, and a common set of traditions
concerning legal matters, in particular sacrificial ritual.
Similarly, the siglum “H” (for Heiligkeitsgesetz, “Holiness Code”) is
commonly used to denote a body of literature which shares many of the
characteristics of P but is distinct from other pentateuchal Priestly material in
its stylistic, thematic, and legal peculiarities––in particular, a specific interest
in the continued sanctification of all Israelites.
While there exists a relatively wide consensus among biblical scholars
concerning the existence of a body of “Priestly” literature,76 which includes
76
Nihan 2007:20 summarizes, “Still today, the distinction between ‘Priestly’ and ‘non-
Priestly’. . . remain[s] one of the few unquestioned results of pentateuchal criticism.” Con-
sidering the major shifts in scholarly investigation into the text-history of the Pentateuch that
have taken place over the course of the past half-century, this consensus is quite telling.
Pentateuchal scholarship in the past half-century may be characterized as attempts to fill the
vacuum left by the decrease in popularity of the classical, Graf-Wellhausenian documentary
hypothesis. See the various contributions to Dozeman et al. 2011, in particular Schwartz 2011;
Schmid 2011; and Gertz 2011:170; in contrast to these attempts see Schwartz 2011 and Baden
2009. For a summary, see Carr 2014; for a discursive introduction to a formative historical
period, see Rogerson 1985.
11:50:13:06:14 Page 19
Page 20
(but is not limited to) much of the sacrificial ritual material in Exodus,
Leviticus, and Numbers, many aspects concerning the Priestly texts are
currently debated. For example, the scope of P,77 its nature as an independent
source or a redactional layer,78 the extent of H and its distinctness from P,79
the relative dating of P and H, and questions of absolute dating80 ––all these
remain the subject of an ongoing debate.
Moreover, most scholars writing on P agree that the Priestly material is
itself the product of different authors writing in diverse historical settings,
though there is no scholarly consensus concerning the precise identification
of different strata within P. In some cases, it is possible to separate several
strata within P on the basis of lexical and stylistic diversity, and especially
disputes on specific legal matters,81 and in a number of cases even to
arrive at a relative dating of these strata with a reasonable degree of
77
See for example, the summaries in Ska 2006:147–51 and Nihan 2007:20–5, and the
bibliography cited in Boorer 2012:45–6.
78
For a brief summary, see Nihan 2007:20; Kratz 2011:36–38; Schmid 2011:18–20 esp.
nn. 4–7, who claims that most contemporary scholars agree that this Priestly material
constitutes an independent source, not a redactional layer, and notes a few exceptions. For a
somewhat more extensive list, see Dozeman 2011:282 and Römer 2008:7.
79
On the relationship between P and H, see Knohl 1995; Cholewiński 1976; Milgrom,
Leviticus 1.13–42, 2.1319–1443; and the bibliography cited in Nihan 2007:4–11.
80
On the absolute dating of P––particularly whether the Priestly texts are for the most part
pre-exilic, exilic, or post-exilic––see Blenkinsopp 1996; Haran 1981, 2008; Hurvitz 1974, 2000;
Rendsburg 1980; Zevit 1982; Barkay et al. 2004; Dozeman 2011:280 n. 76; Ska 2006:159–161; as
well as the general bibliography on P cited above.
81
One argument in favor of the use of historical developments to explain legal contra-
dictions within P is found in Milgrom, Leviticus 1.411–12, 473–81, with reference to the
following contradictions: (a) Lev 6:19 implies that the flesh of the eaten purification offering is
eaten by the officiating priest; 6:22 implies that it is distributed among several priests; (b) Lev
7:9 implies that the processed cereal offerings are handed over to the officiating priest;
2:10 implies that these, too, are distributed among several priests; (c) Lev 7:33 implies that the
right thigh does not undergo elevation (HPWNT) but is handed over directly to the officiating
priest, in contrast to the breast, which does not undergo elevation and is divided between several
priests; 9:21, 10:15, on the other hand, imply no such distinction between the two portions of
flesh and state explicitly that both must undergo elevation. In all of these cases, it is possible that
an earlier stratum of P, reflecting the practice in pre-centralization Judea, assigned the prebends
(flesh of purification offering; cooked grain offering; right thigh of wellbeing offering) to a
single priest, and only a later stratum of P required that they be split among a cadre of priests.
Milgrom may be correct concerning the last two (cooked grain offering and right thigh), but
concerning the flesh of the purification offering, one need not presume a contradiction
between the two texts (as Milgrom himself admits). Other contradictions may be understood as
a result of diachronic developments. For example, Leviticus 3 implies that there is a distinction
between flock animals and bovines in terms of the number of priests involved in their offering:
a single priest offers the suet of a flock animal offered as a wellbeing offering, and several
priests manipulate its blood (3:8, 11, 13, 16); but several priests are involved in the offering of
both the suet and the blood of a bovine offered as a wellbeing offering (3:3). In contrast,
7:33 implies that a single priest offers both the suet and the blood of all zoemes. This, again,
would suggest that 7:33 is the earlier text, confirming Milgrom’s hypothesis concerning the right
thigh.
11:50:13:06:14 Page 20
Page 21
Introduction 21
certainty.82 Some scholars have gone further, attributing passages here con-
sidered part of “P” (e.g., Numbers 15 and 28–9) to post-Priestly redactional
strata which, however––upon their own admission––display linguistic,
stylistic, and thematic affinities to the rest of the Priestly corpus.83
Due to the relative unity of the Priestly literature in general, in particular
with regard to its use of technical terminology and its agreement on major
legal issues, which will be apparent throughout this study, I use the siglum “P”
rather broadly, to include passages attributed to H, which is thus considered
here a subset of “Priestly” literature.84 Reference to text-historical issues will
be found throughout this book, primarily in the footnotes. Our focus, how-
ever, will be on properties common to the detailed sacrificial ritual texts in
the Pentateuch, describing unwritten rules and logical schemes underlying
virtually all of P (and on a rare occasion also the non-Priestly source, “D”).85
This relative unity, as we shall see, ultimately stems from the relative stability
of the grammar of Σ and its rituals’ resilience to change.
As a final point it should be mentioned that among some scholars, focus
has turned to larger literary units that have previously been divided between
Priestly and non-Priestly sources, and to the formation of the discrete books
of the Pentateuch.86 While this book considers the category P (cutting across
the books of the Pentateuch) a valid text-historical category, I would note that
a holistic reading of the various books of the Pentateuch, or even of the
Pentateuch as a whole, lends itself quite well to the “grammatical” analysis
advocated here. In fact, as we shall see, the grammatical analysis proves to be
relevant for several texts composed by the early readers of the Pentateuch in
the first few centuries  and , who were unaware of the Pentateuch’s
complex textual history.
82
See Meshel 2010. On the identification of various strata within P, see in general Rendtorff
1993; Nihan 2007; and Koch 1959. Elliger 1966 is particularly attentive to the text-history of
smaller units within P.
83
See, for example, Nihan 2007: 148–50 on Lev 10; Achenbach 2003:604–6 on Numbers
28–9. See already Noth 1987:121–34 concerning much of Numbers; and Noth 1968:114, 219
for the claim that Numbers 15 is one of the very latest sections included in the Pentateuch, and
that Num 28–9 is even later. On the view that the entirety of Num 27–36 comprises various
interpolations following a process of Fortschreibung, see Nihan 2008 (in particular nn. 3 and
142).
84
For a convenient scheme see Milgrom, Leviticus (3 vols; see in particular the summary
scheme in 1.63). For the present discussion, it is not necessary to accept Milgrom’s absolute
dating of P.
85
While the source “D” is not exactly coextensive with the book of Deuteronomy, many
scholars chose to refer simply to “Deuteronomy” when considering “D” (see Stackert 2007:1–12
and bibliography cited there).
86
Achenbach 2003; the bibliography cited in Nihan 2007:1–19, 84; the various contributions
to Römer 2008, in particular Römer 2008b; Zenger and Frevel 2008. For a similar argument
from a different perspective, see Douglas 1999; Rendsburg 2008.
11:50:13:06:14 Page 21
Page 22
1.4.2 The Generativity of the Grammar
The discussion of the composite nature of P raises a fundamental question

concerning our present “grammar” and its resilience to change over time. In
the following chapters, we describe Σ as a single system, essentially extending
from the Pentateuch to Maimonides, which emanates from a single grammar
that can be abstracted from P. New combinations and sequences unattested in
the pentateuchal law are thus described as generated from a single grammar
and complying with the rules of this grammar. According to this mode of
description, new sacrificial sequences are analogous to new utterances that
comply with a single, fixed grammar of a natural language––a significant
difference being that in natural languages, new utterances are generated on an
hourly basis, whereas new sequences in ritual systems are generated at a much
slower pace. If a grammatical theory of ritual is to be upheld, one must adhere
to this mode of description as far as possible.
There is, however, an alternative mode of description that cannot be dis-
missed out of hand, one which takes into account the possibility that a ritual
grammar may itself undergo changes over time: as in natural languages,
sacrificial sequences that were ungrammatical in the seventh century 
may have become grammatical by the first century ; conversely, acceptable
ritual sequences may have become ungrammatical. This mode of description
is plausible prima facie, since we will refer repeatedly to significant diachronic
shifts that are reflected in Israelite sacrificial ritual texts and in Second Temple
ritual texts.
Since the shifts in question take place over the course of many centuries, it
would stand to reason that several grammars are necessary to describe
the systems at various stages, much like the grammars of Old, Middle,
and Modern English differ significantly. If sacrificial laws changed at a pace
similar to the English language, one should expect to find at least three or four
different grammars for Σ.
However, it is clear that the more one resorts to positing diachronic shifts,
the more grammars must be composed, and the more grammars one com-
poses, the weaker the explanatory power of each grammar. If adhered to
ad absurdum, positing diachronic shifts would necessitate the composition
of an endless number of grammars for every textual stratum in P and the
post-biblical texts discussed here. This would naturally render the com-
position of grammatical rules devoid of theoretical value, and would be
able to explain very little. Recall that standard grammars of many languages
(in particular ancient languages, such as Biblical Hebrew, Akkadian, and
Sanskrit) reasonably account for several hundred years––and more than one
dialect.
An even greater sense of unity than is found in a natural language
11:50:13:06:14 Page 22
Page 23
Introduction 23
characterizes the sacrificial procedures under examination. Recall that the

initial catalyst for composing the present  was the fact that
Maimonides’ generalizations hold true for texts he had never read, composed
more than a thousand years before his time. Therefore, wherever possible,
we describe new ritual sequences as grammatical sequences emanating from
a single grammar rather than indications of diachronic developments that
warrant multiple grammars.
However, diachronic developments certainly took place; we will in fact
encounter them time and again throughout the texts examined in this
book. How then can one distinguish between the two types of innovations,
namely new sequences generated by a single grammar versus diachronic
developments within a grammar?
An analogy to chess may be useful here: throughout this grammar of Σ we
attempt to describe every new sequence as a move in a game of chess, where the
rules of the game remain unaltered. However, important shifts in the rules of
chess occurred throughout the history of the game. For example, the invention
of castling and en passant rules not only allowed new games to be played but
also profoundly affected chess strategy and the average length of most games.
Similarly, in sacrifice one could describe certain innovations as new com-
binations of moves that accord with a fixed, unchanging set of rules. However,
major shifts in the sacrificial system, such as the inclusion of new sacrificial
materials, might be considered diachronic shifts in the grammar itself.87
Two examples should suffice to demonstrate why seemingly ungram-
matical sequences are important for identifying diachronic changes. As we
shall see (in Chapter 0), the inventory of sacrificial animals in P appears to
distinguish individual specimens according to bifurcations based on bio-
logical features. An individual animal is either avian or quadruped, male or
female, mature or immature, and so forth. However, in a small number of
post-biblical texts, the binary nature of the age groups breaks down, and more
than two age groups can exist within a single species. The Book of Jubilees
appears to refer to a tripartite age division of goats, and according to the
opinion of some rabbinic authorities, a distinction is made within the
sacrificial system between male bovines that are immature, mature, and old.88
As we shall see, such tripartite divisions would have been ungrammatical in P,
and their attestation in post-biblical texts suggests that the grammar itself has
undergone a shift.
87
The studies of Rendtorff (1967) and Marx (1998, 2005) highlight these changes. Marx,
who examines Israelite rituals in general (including pre-P, P and post-P material) thus speaks of
“systèmes” (in the plural)––in contrast to the view espoused here, which considers the entirety
of the sacrificial system influenced by P as a “system.” See for example Marx 1994:155–65.
It is often assumed that the religious systems of “simple societies” are more open than those of
“complex societies,” hence allowing for more rapid diachronic change (e.g., Goody 1977:43).
88
See p. 000.
11:50:13:06:14 Page 23
Page 24
The second example might be viewed as an exception within the grammar

rather than an indication of a large-scale shift in its underlying structure. The
pseudepigraphic Testament of Job (15:4–7)89 mentions the offering of pigeons
whose flesh is, in all appearances, shared among a large number of
non-priestly guests, in the manner of wellbeing offerings. If this is in fact the
intent of the passage, it is not only wrong from the point of view of P (recall
that in P birds are not offered as wellbeing offerings but only as wholeburnt
or purification offerings), but also entirely unparalleled in a multitude of
apocryphal, pseudepigraphic, Jewish-Hellenistic, Qumranic, and rabbinic
texts.90 Whatever the reason for this author’s “ungrammaticality”––if one
views the grammar of Σ as prescriptive––given the attestation of this com-
bination in the Testament of Job, we must either admit that in this text the
grammar has changed, or at least that Σ allows for exceptions.
1.4.3 Historicity of the “Grammar”
It may prove beneficial to clarify our assumptions concerning the historicity

of sacrificial rituals described in P and the relation between ritual activity and
ritual texts. Stated simply, one might ask whether the present “grammar”
pertains to sacrificial rituals that actually took place at a certain historical
time or merely to rituals that were conceived of by intellectual elites but never
took place outside of their texts.
On the one hand, there is relatively abundant comparative evidence––
particularly from northern Syria and from Anatolia, but also from
Mesopotamia––suggesting that many of the ritual details portrayed in P are
not inherently unlikely, even if an early date is ascribed to P.91 The evidence––
which pertains both to the general contours of the ritual systems in question
and to many of their particulars––obviously cannot demonstrate that the
rituals described in P were in fact practiced, but it does lend some historical
credibility to P’s textual representation of sacrificial rituals.92
89
Kraft 1974:36–7. The text was composed before the fourth century ; see Brock 1967.
90
One may resort to harmonistic solutions, such as the claim that the protagonist, a pre-
Mosaic non-Israelite, was not conceived as bound by the laws of Leviticus. Alternately, one
might argue that the passage allows for an alternative reading, according to which only the flesh
of the quadrupeds was fed to the poor. However, it is more likely that the author, while well
versed in the Greek version of Job, had not read the legal sections of Leviticus very carefully.
91
See for example Moyer 1983; Weinfeld 1983; Wright 1986, 1987; Beckman 1983 (in
particular Texts J.–K.); Schwemer 1995; Feder 2011 (Hittite); Klingbeil 1998; Fleming 2000
(Emar); Scurlock 2002, 2006 (Mesopotamian); Pardee 2000, 2002 (Ugaritic).
92
Archaeozoology offers a potentially fruitful direction of research on this issue. In par-
ticular, the recent examination of osteonic remains in southern Levantine contexts suggest that
P’s ritual system is not entirely detached from reality. For example, of about 300,000 animal
bones discovered by the sacred precincts at Mt Ebal (alongside several votive inscriptions),
bones of ovines, caprines, bovines, and pigeons have been identified by Magen 2008:160–2
11:50:13:06:14 Page 24
Page 25
Introduction 25
On the other hand, a comparison of P with non-Priestly texts in the

Hebrew Bible suggests that P’s system was not followed to the letter, even
in the Persian period. Moreover, even in the case of Chronicles and Ezra–
Nehemiah, which resemble P in ritual details more than any other biblical
account, it is impossible to demonstrate reliably that the similarity evidences
anything more than literary dependency, direct or indirect.93
So too it has been argued that several law codes from the ancient Near
East fulfilled an ideological programmatic purpose rather than serving as a
blueprint for actual practice. In some cases it appears that P’s legislation is
likewise theoretical rather than practical in its thrust––for example in the
ritual taxonomy of Leviticus 11.94
Finally, when considering the corpus of texts that reflect Σ, it is immedi-
ately apparent that several of the most important (and most extensive)
sources of Σ are patently imaginary: the rituals described in Ezekiel 40–8, the
Temple Scroll, Jubilees, the Aramaic Document of Levi (ADL), and the Genesis
Apocryphon (to mention but a few)95 are all embedded in settings that are
for the most part imaginary––and in the case of Ezekiel, in a temple that is
explicitly visionary.96 Early rabbinic sources, all available from a period in
which the Israelite sacrificial system was no longer practiced (post-70 ),
and even the works of Philo and Josephus, which contain testimonies that
are ostensibly more reliable from a historical point of view, cannot be relied
upon uncritically for their depiction of ritual minutiae even in the late
Second Temple period.97 This is not to say these texts do not contain evidence
of actual practice; the current evidence suggests that at least some of the
particulars of these Jewish Hellenistic ritual texts were not merely imagined,
(cf. Lev. 1). A recent study (Davis 2008) demonstrates that in certain Iron Age sacred precincts,
the right femur of large and small cattle is found to be four times more common than its left
counterpart. This suggests unequal ritual status of the left and right cuts––specifically, the
predominance of the latter––in the ritual distribution of meat (cf. Lev 7:33).
93
See Marx 1994:155–65; 2005: 211–12; Rendtorff 1967.
94
Meshel 2008; see Greenberg 1968; Liss 2006; and in brief Klingbeil 2007:52–4. On P as a
rhetorical text rather than a manual or handbook for priests, see Watts 2007; and Watts
2008:94–5. More generally, see the caveats expressed in Kunin 1995:49–50.
95
On Ezekiel 40–8, see for example Zimmerli 2.327–553; and, from a different angle,
J.Z. Smith 1987a:47–73. For Jubilees, see Vanderkam 1989; for ADL, Greenfield, Stone, and Eshel
2004. On the Genesis Apocryphon, see Fitzmyer 1966 and Machiela 2009.
96
The burden of proof lies on those who would claim that Ezekiel’s ritual system reflects
actual practice in the Jerusalem Temple, projected to the eschaton. See, for example, Haran 2008,
2009a.
97
For Josephus see the notes in Feldman 2000:291–306; for Philo, DSL (Colson 1937),
Daniel-Nataf 2000; see the discussion of both in Petropoulou 2008:131–2, cf. 188–9. For a
balanced view of the use of rabbinic and late Second Temple textual evidence (Aristeas,
Hecataeus of Abdeira, Sirach, Philo, Josephus, and other sources) for the reconstruction of
actual practice in Jerusalem, see Schwartz 2008. Note, however, the statement on p. 10*,
regarding an early Second Temple text––1 Chron 29––which he claims “undoubtedly” reflects
common practice. See recently Cohn 2013:133 nn. 6–7.
11:50:13:06:14 Page 25
Page 26
particularly where two or more post-biblical sources independently

corroborate a ritual detail not found in P.98 However, the abundance of
“fantastic” ritual texts raises certain suspicions concerning the historicity
of Σ in general.
Concerning the ritual texts of P, matters are even more complex, and more
hotly debated.99 Admittedly, it stands to reason that the authors of P super-
imposed, via literary retrojection, the rituals practiced in their own circles
upon their narrative of the Tabernacle in the Wilderness.100 This suggests
that one may glean information about sacrificial procedures in the
southern Levant in the eighth to sixth centuries or in Yehud in the Persian Era
(depending on how one dates P)101 on the basis of the rituals described in the
narrative. However, this assumption ignores a few serious hurdles that should
impede the mental leap from ritual text to ritual practice.
Even if it were certain that the authors of P attempted to superimpose a
contemporary ritual system onto a tabernacle that existed in illo tempore,
clearly not all details could be retrojected with equal success. Many questions
thus remain open due to significant differences between P’s (the authors’)
milieu and P’s (the narrative’s) setting. For example, it is unclear whether the
single tabernacle in the wilderness should be construed as a corollary of local
altars (“high places”), of several major temples, or of a single sanctuary in a
central location.102 The authors of P do not usually refer to the manner in
which they thought that their narrative constructs would apply in their own
day.103 Thus, there is no reason to assume a priori that any specific detail
reflects the reality known in P’s day rather than, for example, the manner in
which the authors imagined that events must have transpired.104
Together, these considerations all suggest that a degree of skepticism is
justifiable, at least as a starting point, with regard to the historicity of P’s
98
An oft-cited example is found in Dion 1987; Wevers 1995:221–2.
99
See for example Grabbe 1997; 2007, esp. 23–25. With particular reference to some of the
details in P, see Greenberg 1968; see also Gadegaard 1978, and the retort of Zevit 2001:288 n. 59.
100
For an explicit formulation of this presupposition, see Wellhausen 1973:34.
101
Above, n. 78.
102
See Brichto 1976; Milgrom 2000. Similarly, the distinction between common Israelites
and chieftains, which has ritual implications (Lev 4), has no obvious correlate in any historical
period outside P’s own narrative. And one cannot tell for certain whether the complex rites of
ordination described in Exod 29 and in Lev 8–9 were meant to refer to a unique historical event
(at the first establishment of the priestly lineage) or to a recurrent event. If the latter case is true,
then it is unclear who would be expected to play the role played by Moses, a non-priest, in Exod
29 and in Lev 8.
103
For a rare exception, see Meshel 2013a:286.
104
Ironically, the narrator in the Priestly literature only rarely implies that the rituals related
in P actually took place: for the most part,  commands that certain offerings be made in
accordance with a set of rules, but their actual performance is not described. Judging by the
rhetoric employed by several prophets concerning many divine decrees, and insinuated in the
Priestly literature as well (Lev 26), one would be led to believe that at least some of the Priestly
laws were not carried out.
11:50:13:06:14 Page 26
Page 27
Introduction 27
ritual system. We will proceed with the tentative conclusion that the ritual
details in P most likely reflect an idealized form of actual practice.
It should be stressed that from a purely formal point of view, the distinction
between real and imaginary ritual sequences is not cardinal for the gram-
matical investigation of ritual. Generative grammars of natural languages
must account for sentences that are actually uttered as well as sentences
that are merely imagined but never uttered (such as arbitrarily long or prag-
matically unlikely sentences). In fact, sentences of the second group often
prove to be particularly elucidating for the linguist.105
The centrality of such “unreal” sequences is already indicated clearly in
Patañjali’s Mahābhās.ya. The story of Br.haspati and Indra referred to in
Section 1.1 bears this point out. A grammar of the Sanskrit language was not
invented for the description of actual utterances: thanks to the longevity
of both teacher and disciple, Indra would have been able to internalize all
utterances spoken thus far by simple repetition; according to Patañjali, it is
only for the sake of unuttered linguistic sequences––such as infinitely long
utterances––that a grammar must be formulated.106
Patañjali’s analogy to the mahāsattras, sacrificial rituals that can last
hundreds of years, explicitly reveals that a similar situation holds true for
rituals. The sacrifices that practitioners would have been exposed to,
though manifold, are presumably not so numerous that they could not be
internalized without a “grammar”; it is primarily for the sake of imaginary
sacrificial rituals that are never carried out in practice, and which cannot be
internalized by direct exposure, that a “grammar” must be composed.107
1.5 CATEGORIES OF THE GRAMMAR OF Σ
Having established the desirability, but current lack, of even a single systematic
grammar of ancient ritual, we will now begin to establish a grammar for Σ,
the priestly sacrificial system of ancient Israel. As mentioned above, the
operative categories of this grammar need not––and in fact do not––resemble
the morphology or syntax of natural languages, as such categories are foreign
to the Israelite ritual system and need not be meaningfully sought within it.
Instead, we introduce the following four internally meaningful categories:
1. “zoemics”––the study of the categories of animals acceptable as
sacrificial offerings;
105
See below, § 3.9.1 (p. 000), and Pinker 1994:209–10.
106
See n. 7.
107
On the importance of imaginary rituals (such as sacrifices carried out by deities) for the
theoretical study of ritual, see Patton 2009.
11:50:13:06:14 Page 27
Page 28

2. “jugation”––the study of the acceptable combinations of animal and
non-animal sacrificial material;
3. “hierarchics”––the study of the acceptable higher-order structures made
up of one or more zoemes; and
4. “praxemics”––the study of the sequential ordering of ritual activities.
Accordingly, Chapter 2, , reveals that sacrificial animals within Σ
are subject to a system of bifurcation that leads to the creation of twenty
building blocks, and that these in turn can combine to create new classes of
animals fit for particular sacrificial occasions. Chapter 3, , investi-
gates the specific rules governing animal and non-animal sacrificial com-
binations in Σ, for example in situations in which the offering of a zoeme
entails subordinate a grain offering and a libation. Chapter 4, ,
demonstrates that terms designating sacrificial types such as HL m WoE, TAU
m
dxX,
and O$ m (wholeburnt, purification, and reparation offerings, respectively)
moA
are not mutually exclusive as is usually believed; rather, each term has
two distinct denotations, so that a sacrificial animal, for example, can be at
the same time an HL m oWE in one sense and a TAU m
dx X in another. Chapter 5,
, abstracts a finite number of discrete ritual acts from P and dis-
tinguishes between combinations of ritual acts that are grammatical (that is,
licit) and combinations that are ungrammatical and hence unlikely to occur,
although they are physically possible.
After a chapter discussing each of these operative categories, we will
turn to examine the place of meaning in the grammar of sacrifice. Finally, A
“G” of Σ integrates the findings of this study with previous findings
on Israelite sacrifice, all presented in a concise, partially formulaic manner.
11:50:13:06:14 Page 28
Page 29
Zoemics
2.1 INTRODUCTION
To begin, let us consider the operative category of zoemics (from Greek ζον,
“animal”), the classification of animals within a sacrificial system. We begin
with zoemics because of its centrality in Σ, and because the other operative
categories, as will become apparent in due course, are best explained in
terms of zoemics. The term “zoeme” is coined by analogy to “phoneme,” a
psychologically recognized discrete unit of sound in a natural language. By
way of a first definition, a zoeme is a group of individual specimens of
animals that share certain relevant biological traits and to which a common
set of sacrificial rules apply.
For example, the categories designated by the terms LYiA x (“ram”),
RP
xd (“bull”), and HN i
e (“pigeon”), as used in the legal Priestly literature, are
mWoY-IB
zoemes, inasmuch as each of these animal sacrifices is required of a Nazirite,
a high-priest, and (optionally) a parturient, respectively, under specified con-
ditions (Num 6:14; Lev 4:3; Lev 12:6). Likewise, the categories designated
by the terms B$ K (“sheep”) and ZE
oe d
e a (“goat”) are zoemes inasmuch as
different sets of laws apply to these animals if offered as wellbeing sacrifices
(Lev 3:7–11, 12–16).
Like a phoneme, a zoeme is an abstract category consisting of elements
that share some features that distinguish them as a category, though they may
differ from one another in certain irrelevant aspects.1 Also as with phonemes,
the term “zoeme” may additionally, for ease of reference, be used to refer to a
member of that abstract category.
A linguistic example may help to clarify my meaning. Native English
speakers are typically able to distinguish between the sounds represented by
“l” and “r,” easily telling “glow” from “grow,” or “plowed” from “proud.”
Such pairs of words, which differ only in one element (here, an “l” versus
an “r”), are known as minimal pairs and are central to distinguishing
1
See Chomsky and Halle 1991:64–5.
11:50:13:06:14 Page 29
Page 30
phonemes in languages: because they can find examples where an “l” versus
an “r” makes the difference between words, the two sounds must represent
two distinct phonemes to them. However, some non-native English speakers
from Japan and Korea, who do not have such “l”/“r” minimal pairs in their
native languages, have difficulty pronouncing two distinct sounds, or even
telling the above minimal pairs apart. These difficulties, as well as the absence
of minimal pairs in their native languages, suggest that they consider the
sounds a single phoneme.
As another example––this time where English is the language that fails
to perceive the difference––consider “aspirated” versus “unaspirated” con-
sonants, a distinction made in languages like Mandarin Chinese. A common
way to experience the difference in aspiration is to hold a sheet of paper
vertically by its top corners in front of your mouth and repeat the words
“spring” and “pit.” The paper should fly forward much more during “pit”
than during “spring” due to the aspiration, or puff of air, inherent in the “p”
of “pit.” The different movement of the paper indicates that a physically
different sound is being produced, but to native English speakers, the sounds
are “the same ‘p’.” Because native English speakers, unlike native Mandarin
speakers, think of them as identical, within English the sounds constitute a
single phoneme /p/.
As with phonemes, different sacrificial systems can make different zoemic
divisions. One can imagine, for instance, a system in which some rituals
require the sacrifice of specifically white rams while others require black rams.
In such a system, there would be a zoemic distinction between /white ram/
and /black ram/. However, all rules in Σ pertaining to rams only pertain to
rams in general, with no distinction concerning specimen color. In a system
like Σ, white and black rams are of course physically no less different than in
other sacrificial systems, but here they are members of one and the same
zoeme, /ram/.
Some zoemes in Σ comprise members that, from a physical perspective,
closely resemble the members of other zoemes, though each zoeme remains
typologically distinct. For example, a lamb in the first year of its life (which, as
we shall see, is usually designated in P by the term WoTN
m$ ie o$e
vo-IB Bd
eK)2 externally
resembles a ram (LYiA x ) more than it resembles a bull (RP d) because lambs
x
and rams belong to the same zoological species (they are domestic sheep,
Ovis aries). However, in the grammar of ritual, this external similarity is as
accidental as the phonetic similarity between English “lamb” and “ram.” In
the Israelite system, there is a zoemic distinction between “immature male
ovine” (WoTNm$ ie o$e
vo-IB Bd
eK) and “mature male ovine” (LYiA x ), with the result that,
2
See the discussion in §2.4.5, p. 000 below.
11:50:13:06:14 Page 30
Page 31
Zoemics 31
where the law requires a ram (as at Lev 5:15), a lamb is as unacceptable as
a bull.3
The structural analogy between phonemes and zoemes is limited. As we
shall see, it is most economical to define the zoemic structure in such a
manner that one zoeme can be a subcategory of another zoeme. For example,
HT
d mm$
Nvo-TBd
x H$ om v
Bd
xK (“immature female ovine,” or “female lamb”) and B$ oe d
e
K
(“ovine”) are two distinct zoemes,4 even though outside the zoemic scheme,
“immature female ovine” is obviously a subcategory of “ovine.” In phonemic
systems this is impossible––it would be as if /s/ and “sibilant” (a category
comprising several sounds and including /s/) were two distinct phonemes in
English.
Despite the elementary nature of the zoemic classification system and its
centrality for any descriptive account of ancient Israelite sacrifice, the identifi-
cation of certain zoemes has long invited a host of problems. Many cases of
mistranslation and misidentification of the categories of sacrificial animals
have persisted from antiquity to the modern period. A case in point is the
term RWo$o, which has been misconstrued as “bull,” though the term actually
denotes a “bovine” of any age and either sex (“cow” as used in colloquial
English in the sense of “member of the species Bos taurus”).5 This and other
similar cases are discussed below, in Section 2.4.
Some of this lack of clarity is the result of lexical variation, which is
notoriously confusing when pertaining to animal species, sexes, and age
groups, even if one neglects any changes in the language that may have taken
place over time. For example, in some languages, a single noun in one context
may designate a species as a whole (e.g., “cow” meaning a domestic bovine of
any age or sex) and in another context an adult of one of the sexes (a female
“cow” as distinct from a male “bull”). As demonstrated in Sections 2.4.3 and
2.4.10, this feature is found in BH in general and in the specific vocabulary of
terms for zoemes in particular.
Semantic shifts over time offer another source of confusion. An example
comparing BH with the Hebrew of the Temple Scroll and Mishnaic Hebrew
oe in BH denotes “member of the
is particularly telling in this regard: while H$
3
Note that in one case, Tannaitic tradition reflects the insertion of a “buffer zone” between
these two zoemes. According to this tradition, the zoeme o$e BdK (identical with WoTN
e m$ ie o$e
vo-IB Bd
eK)
includes specimens up to the age of one year, whereas the zoeme LYiA x includes specimens that
are over thirteen months old but under two years of age. A specimen between twelve and
thirteen months is considered acceptable neither as a o$e K nor as an LYiA
Bd
e x . See m. Parah 1:3. This
“buffer zone” is analogous to a buffer zone between phonemes––a sound that is physically
located between phonemes but is not part of a language––essential for the creation of phonemic
distinction in natural languages (see Leach 1964). This SG xL
vxP (vocalization following MS
d
Kaufmann) can therefore be said to have been placed “between zoemes” in order to clarify the
zoemic distinction between o$eBdK and LYiA
e x.
4
These terms will be discussed at greater length in §2.4.7, p. 000 below.
5
See Péter-Contesse 1985 (note the correction on p. 161 in that volume), 1992.
11:50:13:06:14 Page 31
Page 32
flock” (hence an ovine or a caprine), the Temple Scroll contains the formula
ZEW H$W RW$ (52:5,13), which clearly implies that H$ oe here refers to ovines
alone. This usage is found also in MH.6 Further differences between BH and
MH pertaining to the precise denotation of terms for animal species can be
found: (1) the term LXa Rm can be used in MH to include male sheep, although
in BH it refers only to females;7 (2) YDiG
v , which probably denotes only goat
d
kids in BH, can include sexually mature he-goats in MH;8 and o$e K can
Bd
e
be a subcategory of LYiA x in MH, while the opposite hierarchical relation
sometimes applies in BH.9
This confusing situation is aggravated by extreme diversity in the common
and technical usage of terms for cattle in modern languages, such as the
English “ox,” “cow,” and “heifer,” and so forth, which vary widely regionally
and diachronically. Modern English “ox,” for instance, can denote a sexually
undistinguished “domestic bovine quadruped,” or specifically a “male
castrated” one.10 Therefore, the first task of the scholar who wishes to study
systems of animal sacrifice should be to identify, as far as possible, each of
the different zoemes in the sacrificial system under consideration and to
6
This understanding also appears to underlie the question AWH OYYALK H$ WYBA (“if
its father is an ovine [sic], is that kil’ayim?”) in b. Bekh. 12a––otherwise the pericope is
unintelligible.
7
See Gen 31:38; 32:15; SoS 6:6. In all of these cases, the context reveals that a mature female
is implied. The same may be the case in Isa 53:7, though this cannot be proven from context.
By contrast, MH LX a Rm denotes either an adult female or the species as a whole (compare BH
as well as MH ZE a, which may designate either the adult female or the species as a whole). See
m. Bekh. 2:5 LXR IYMK HDLY$ ZEW (“and a goat that birthed [a male birthling] in the form of a
lamb [sic]”), where it is clearly a male lamb that the author of this mishna has in mind, since the
laws of the first-born are applicable here. See also OYLXR L$ RKZ (“a male of the ovines [sic]”)
in 5:3. But LX a Rm in MH also means a specifically mature female sheep, as in m. Shab. 5:2, 4
(distinct from OYRKZ, see Aram. IYRKD, equivalent to BH OYL a ). Similarly, concerning donkeys,
i YA
see m. Bekh. 1:3: OYRKZ YN$ HDLYW HRKYB AL$ RWMX (“a donkey that, having never given birth
before, births two males”).
8
See t. Bekh. 1:13: YDGH IMW LXRH IM OYALKH (“the kil’ayim [born] from a [female] ovine
and a male caprine”). Compare BH WMA BLXB YDG (“a kid in its mother’s milk,” Exod 23:19;
34:26; Deut 14:21); see also Jub. 14:6 and particularly 1 Sam 10:3.
9
t. Parah 1:5 clearly implies that H$ om v
Bd
xK is a subcategory of LYiA x . Compare Gen 32:15
et passim, where LYiA x must denote mature male specimens, and see p. 000 for o$e Bd
eK in the
general sense of ovine.
10
OED, s.v. “ox.” Note that etymologies and comparative semantics are not necessarily
helpful in sorting out the meanings: in fact, in the “ox” example, the Sanskrit cognate refers
specifically to an uncastrated male. Note that many Bible translations and dictionaries use terms
that are ambiguous in modern languages, such as English “sheep,” French “chèvre,” or German
“Lamm” and “Ziege.” Thus, NCFD (153a) renders chèvre as either “goat” or “she-goat”––in
either case, it is a feminine noun––although DLF (2.259) renders chèvre only as “la femelle du
bouc. . .” Conversely, CGED (743b) renders Ziege––always a feminine noun–– as either goat or
she-goat, though BWDW (6.833b) has “Angehörige einer Gattung der Horntiere. . .” As noted
above, English “cow,” may denote either “the mature female of cattle” or “a domestic bovine
animal regardless of sex or age” (see also WNCD 268b). As a result, even some of the most
perceptive studies falter on elementary identifications, which may lead to entire passages and
their import being misunderstood.
11:50:13:06:14 Page 32
Page 33
Zoemics 33
designate these with a clear set of abstract terms or signs.11 Accordingly, I will
use the following signs: five Roman letters––B, S, G, T, and Y––to designate
cows, sheep, goats, turtledoves, and pigeons, respectively; capital letters
to designate mature specimens and lowercase for immature (hollow capitals
for specimens of any age); and the symbols 웧 and 웨 to designate male and
female specimens, respectively.
2.2 THE ZOEMIC CLASSIFICATION SYSTEM
2.2.1 Criteria for Categorization
As a preliminary to explaining the nature and function of zoemes in Σ, I must

explain three criteria or biological variables that, despite several differences
in the zoemic classifications of biblical and post-biblical sources, suffice to
categorize all sacrificial animals into elementary “blocks.” These criteria are
(1) zoological class (i.e., taxon), (2) sex, and (3) age.12
(1) Zoological taxon

Contrary to what may be inferred from previous attempts at schematic
description,13 the zoological classification in P consistently follows a multi-
level binary pattern. A sacrificial animal is either a quadruped (HM maHvB) or a
d
bird (FWoE, more precisely “winged creature”). If it is a bird, it is either a
turtledove (RT o )14 or a pigeon (HN
d
ie ); if it is a quadruped, it is either large
mWoY½IB
cattle (RQmdB) or small cattle (IAC
m o ). If it is large cattle, it is of necessity bovine,
but if it is small cattle, it is an animal “of the flock,” and hence either an ovine
(B$ K, rarely o$e
oe d
e BdK) or a caprine (ZE
e a). Note that this bifurcation, in its entirety,
is explicit in the language of P (Lev 1:2–3, 10, 14).
11
See Goody 1977:122: “it is the existence of a notation far removed from speech that makes
possible [. . .] the kind of highly abstract, decontextualised and arbitrary procedures that are
typically represented by the formula.”
12
The cross-classification of animals according to these criteria is roughly analogous to the
classification of phonemes according to distinctive feature analysis. See n. 000.
13
For example, Marx 1994:47. Maimonides’ opening words in HMQ 1:1 are TWNBRQH LK
OYRWTH IMW OYZEH IMW OY$BKH IMW RQBH IM :DBLB IYNYM H$MXM IYAB HYX $PN YNYM L$
HNWYH YNB IMW (“all animal sacrifices come from five species only: bovines, ovines, goats, turtle-
doves and pigeons”). This presentation of “five species” blurs the binary nature of the zoemic
tree. Coincidentally, Vedic traditions also often enumerate five species of animals fit for sacrifice,
though in some texts the number is vastly augmented. For a discussion, see B.K. Smith
1994:249–55.
14
For a different opinion, see Staubli 2008.
11:50:13:06:14 Page 33
Page 34
This criterion is based on a taxonomy, a system of categorization that does

not allow for cross-classification.15 Therefore, classification according to
zoological class can be illustrated reliably in the form of a branching tree, as
Figure 2 shows.
Fig. 2. Classification according to zoological class.
(2) Sex
Among quadrupeds, a distinction is made throughout P between sexes.
For example, P distinguishes between OYZ d
iEi RYE $ (“male goat,” Lev 4:23
i ov
et passim) and OYZ d
iEi TRxYEov (“female goat,” Lev 4:28; 5:6) and between
i$
WoTN
m$ ie o$e
vo½IB Bd
eK (“immature male ovine,” Num 6:12, 14, et passim) and H$ om v
Bx
dK
16
dHTmm$
Nvo½TB x (“immature female ovine,” Lev 14:10; Num 6:14). This dis-
d
tinction does not apply to birds: contrary to a hyper-literal understanding
of the word IB a , the designation YN
d aBd in the phrase HN
v mWoY½YN
aBd does not denote
v
“male,” but merely “members of the category.”17
15
That is, the categorization of each element is exclusive and absolute: a single element
cannot belong to two distinct categories simultaneously unless one is a subset of the other, nor
can it belong to a particular category to a greater or lesser extent. Empirically, most known soci-
eties––modern as well as primitive––prefer taxonomy to other cross-classification schemes for
animal categorization. See Sperber 1975b.
16
Note that the reading is H$ om v
Bd
xK, not H$
om viK, according to the Masoretes.
Bd
17
See Rendtorff 1985:74. This is explicit in b. Kid. 24b, but it is also true in P. See also BRaEo½YN
aB
d
v
(crows, probably of all ages and surely of either sex, Ps 147:9); $oYA i -YN
aB
d ,OD
v mAm ½YN
aB d (often humans
v
of any age or sex, Ps 90:3; Lam 3:33); and R$ eoN
e-YN
aBd (vultures of any age or sex, Prov 30:17); see
v
also OYMiA ie (Ps 29:6), IAC
a Rv-IB o -YNd (Ps 114:4, 6). Even RQ
aB
v md
m ie can designate bovines in general
B-IB
(see Num 15:8, where such an animal may be offered as a XB xeZ or as an HLm WoE; it is therefore
synonymous with the term RWo$o in v. 11).
11:50:13:06:14 Page 34
Page 35
Zoemics 35
(3) Age
Similarly among quadrupeds, P clearly distinguishes between immature and
mature specimens. This distinction is also binary, since at most two age
groups are considered in each case. For the sake of clarity, I refer to these
two age groups simply as “mature” and “immature.” It is likely that this
distinction corresponds to the biological distinction of sexual maturity, the
ability to reproduce. However, since there may be more than one criterion
for maturity and, consequently, several stages of maturation, this is a clear
case of discretization of a natural spectrum. There is no internal evidence
within P that the “mature”/“immature” distinction also pertains to birds.18
2.2.2 Schematic Structure of the Zoemic Classification System
Together, these three criteria place any sacrificial animal in one of twenty
elementary categories, or “blocks” (20 = 5 taxa · 2 sexes · 2 age groups).
Figures 3 and 4 represent these blocks on two- and three-dimensional
“zoemic maps,” as abstracted from P. Both diagrams represent the same data,
but each serves a distinct illustrative purpose.
The two-dimensional diagram (see , Zoemics rule 2) is some-
what simpler to conceptualize; more importantly, the diagram (Fig. 3) clearly
demonstrates the essentially binary nature of the zoological, sex, and age
distinctions in P’s zoemic system, and it provides (in the shaded squares) an
accessible listing of all twenty possible combinations of species, sex, and age
that a single animal specimen may represent. These practical advantages will
render it more useful in most future discussions.
However, Figure 3 has its limitations. Most significantly, by graphically
privileging zoological species as its dominant structural feature, the diagram
might suggest that zoological species is in some ways a more essential factor
18
Contra Milgrom, Leviticus 1.168. One might argue that P assumes that such a distinction
is made between the ages of turtledoves and pigeons, according to an ancient custom (see Gen
15:9, in a non-sacrificial context, where the species of the LZ mWodG is not mentioned), but the law
nowhere implies that this is the case, and it certainly does not require it. In proving that HN mWoY½YN
aB
d
v
denotes fledglings in BH, Milgrom only adduces that YN aBd would otherwise be superfluous, since
v
in BH HN mWoY also appears without a modifier (Gen 8:8; Isa 60:8). However, R$ eoN
e-YN
aBd (“vultures,”
v
Prov. 30:17) is similarly found in BH alongside OYRi$ movN (2 Sam 1:23; Isa 40:31), and yet R$eoNe-YN
aB
d
v
are clearly mature (according to context: fledgling vultures do not go about pecking at people’s
eyes, since they wait in the nest for their parents to feed them). Similarly, the term RQ mm
B-YN d is
aB
v
found in P alongside RQ md
mB (Num 7:15 etc., 87–8; 28:11; see previous note, and generally in BH
as at Exod 21:37). Moreover, the phrase RQ mdm
Bx e is used in P in contrast to OYZ
H-IB
d d
iEi ,OYBi$omd
vK
(which are IAC do x i , Lev 1:2 10, but never in P IAC
H½IM o ½YNd )––a situation analogous to the use
aB
v
of HN mWoY-YN
aBd in contrast to OYRiWoT
v d ––and yet the text clearly does not imply that sheep must be
mature, bulls immature. Therefore, there is no reason to suppose that such a distinction should
be made in P between turtledoves and pigeons.
11:50:13:06:14 Page 35
Page 36
Fig. 3. Two-dimensional zoemic map.
Fig. 4. Three-dimensional zoemic map.
than either sex or age. In truth, however, all three criteria are equally
important within the zoemic system. The three-dimensional diagram (Fig. 4)
is therefore a fundamentally more accurate representation of the interplay
among the criteria, since each is afforded equal importance within the
graphic depiction.
11:50:13:06:14 Page 36
Page 37
Zoemics 37
A zoeme comprises any combination of one or more of these blocks;

twenty blocks therefore yield 220 possible zoemes.19 We have thus defined
“zoeme” in two ways, (1) as a group of animals to which a set of sacrificial
laws may pertain and (2) as a combination of one or more blocks. To see the
equivalence of these definitions, consider the set comprising the members of
s웨, b웧, and Y웧. In theory, a specific rule could apply to these animals and
these alone: for instance, a specific amount of wine may be offered when
one of these animals is sacrificed. In reality, however, only a relative handful of
the 220 possible zoemes are actually attested entities. This severe limitation
suggests that certain rules of zoemic formation are active within P. For a full
list of these rules, see the section  in the  booklet.
Table 1 includes all of the relevant zoemes attested in P, explaining
their composition according to these two definitions. Philological notes
Table 1. Zoemes attested in P
BH term(s) Zoeme English name(s) Constituent Example

block(s) passage or rule
LYiA
x S웧 “ram” S웧 Lev 5:15
iWBYRiQ
vx
Y R$
eoAq HMmaHvd
B “quadruped” b웧, b웨, B웧, Lev 1:2; 27:9
HWHYLx IB
d
m RvQ
m HN
d
meM
dMi B웨, s웧, s웨,
S웧, S웨, g웧,
g웨, G웧, G웨
HN ie
mWoY-IB “pigeon”a y웧, y웨, Y웧, Lev 5:7; 12:6
Y웨
WoTN
m$ ie o$e
vo-IB Bd
eK ,$
o e
Bd
eK s웧 “male lamb” s웧 Num 28:3–4
HT
d mm$
Nvo -TB
d
x H$
om v
Bd
xK s웨 “female lamb” s웨 Lev 14:10
o$e
Bd
eK rarely, B$
oe d
e
K “ovine,”b “sheep” s웧, s웨, S웧, S웨 Lev 3:7–11
. . . o$e
Bd
eK, HB
d
m$ov d
iK 웨 “ewe” s웨, S웨 Lev 4:32; 5:6
HBmaQvN (differs from
웨 inasmuch
as broad tail is
offered)
i . . . IAC
IM do xH IMi 웧 (no common s웧, S웧, g웧, Lev 1:10–11 (as
IM
i WoA OYB i$omdv
KxH term) G웧 wholeburnt
RK
mZm . . . OYZd
iEimH offerings,
slaughtered on
northern flank
of altar)
IAC
do x
H IM
i HB ma QvN 웨 (no common s웨, S웨, g웨, Lev 5:6
TRxYE
i$ov WoA HB
d
m$ov d
iK) term) G웨
(OYZd
iE i
LG
eE
a 웧 “male bovine” b웧, B웧 Lev 9:2
19
To be precise, 220 − 1 = 1,048,575 possible zoemes, since an empty set could not possibly be
a zoeme.
11:50:13:06:14 Page 37
Page 38

Table 1.––cont.

LG
eE a b웧 “male calf” b웧 Lev 9:3
c ie
WoTN
m$vo*-IB
FWoE “avian,” “bird,” t웧, t웨, T웧, Lev 1:14
“fowl” T웨, y웧, y웨,
Y웧, Y웨
ZE
a “caprine,” “goat” g웧, g웨, G웧, Lev 3:12 (note
G웨 distinction
from )
HT
d mm$
Nvo½TB
d
x ZE
a g “kid” g웧, g웨 Num 15:27
i
DWTdE
x G웧 “mature billy G웧 Num 7:17, 29,
goat” etc.
RP
x
d B웧 “bull” B웧 Lev 4:3
(coll.) IAC
o ,H$
oe “animal of the s웧, s웨, S웧, Lev 3:6–16
flock,” “small S웨, g웧, g웨, (note
cattle” G웧, G웨 distinction
from bovines
in number of
priests, vv. 11,
16 || v. 5)
HN
m$ ie RK
mo-IB mZ
m . . . H$
oe f웧 (no common s웧, g웧 Exod 12:5d
term)
e
(coll.) RQ
md
mB ,RWo$o “bovine” b웧, b웨, B웧, Num 15:8–10
B웨
RYE
i$om ,OYZ
d
iEi RYE
i$ov 웧 “he-goat” g웧, G웧 Lev 4:23
OYZ
d
iEi TRxYE
i$
ov 웨 “she-goat” g웨, G웨 Lev 4:28
RWoT
d “turtledove” t웧, t웨, T웧, Lev 5:7
T웨
(none) “sacrificial (all twenty Lev 1 and 3
animal” (no blocks) (immolation
common term) required)
(none) 웧 (no common b웧, B웧, s웧, Lev 1:3, 10
term) S웧, g웧, G웧 (acceptable as
wholeburnt
offering,
slaughtered)
(none) s웧+s웨+S웨 (no common s웧, s웨, S웨 (all Num 15:4–5,
term) of except for 11 (specific
S웧) amounts of
grain offerings
and libation),
combined with
Lev 3:9
11:50:13:06:14 Page 38
Page 39
Zoemics 39

(none) s웧+s웨+S웨+ (no common s웧, s웨, S웨, Num 15:4–5,

term) g웧, g웨, G웧, 11 (specific
G웨 (all of amounts of
except for S웧) grain offerings
and libation)
(none) 웧+ (no common b웧, B웧, s웧, acceptable as a
term) S웧, g웧, G웧, wholeburnt
t웧, t웨, T웧, offering
T웨, y웧, y웨, (Lev 1:3, 10, 14)
Y웧, Y웨 (all of
except 웨)
a
Note that one could not determine that is a zoeme according to Lev 1:14 or 12:6, since no specific set
of rules applies exclusively to (every rule that applies to it applies to as well). However, technically may
be defined as a zoeme since the following rule applies to it: if a pigeon is offered within a ritual that
requires two birds, the second must also be a pigeon (and similarly for ).
b
I follow the English usage of these terms and not Linnaean terminology. Thus, “bovine” denotes “cows”
(genus Bos, species Bos Taurus), “ovine” denotes “sheep,” and “caprine” denotes “goats” (not the subfamily
caprinae). According to Linnaean terminology, family Bovidae includes the subfamily Bovinae as well as
caprinae. Caprinae, in turn, includes genus ovis and genus capra.
c
The term only appears in the phrase HN m$
mo-YN aB
d o$e
v BeW
Km LGa. See 2.4.5.
eE
d
On the questionable sacrificial status of this ritual, see Eberhart 2002:274–8 and 2011:20, 30. See,
however, the references to XBZ and IBRQ in Exod 12:27 and in Num 9:7 respectively, which suggest an early
interpretation of the paschal ritual in sacrificial terms.
e
For example, OYiNx$
vo RQmd
mB (e.g., Num 7:17, 23, etc.) and TWo$oRvXo iWYH
m RQ
mdm
BxH (Job 1:14; epicene), as well as
when RQ md
mB appears in the stock phrase RQmd
m
B-IBie , which usually follows RPd (e.g., Num 28:11, Lev 1:5).
x
concerning problematic and controversial cases listed here will be discussed

in Section 2.4.20
2.2.3 Formal Properties of the Zoemic System
Now that the majority of zoemes have been identified (Table 1), it is possible
to note two important properties of the zoemic system: its independence
from the languages that describe it, and its binary nature.
2.2.3.1 Zoemes and Names

Even at first glance, the uniquely technical terminology employed in the
zoemic system stands apart from the terms used commonly in all other
20
A vast number of additional zoemes can be constructed artificially, by means of uniting or
intersecting the sets listed in Table 1. Such zoemes, though technically “attested in P,” are hardly
relevant to the discussion and are not presented here.
11:50:13:06:14 Page 39
Page 40
biblical texts. The terms OYZ diEi RYE i$ov and B$ K, for example, are hardly found
oe d
e
outside Priestly literature (but see Gen 30:32, 33, etc.; and 37:31), and HB d
m ov
$diK
may very well be a neologism of P. On the other hand, in poetic and narrative
texts pertaining to domestic cattle in ancient Israel, other terms (such as
v , RK
AYRiM x , and I$
d mom
B-YN d ) occur beside P’s zoemes (as at Deut 32:14; Isa 1:11;
aB
v
Amos 5:22). Therefore, it is evident that the priests who composed these
sacrificial texts were not simply following common usage, but rather creating
a parlance of their own.21
As Table 1 demonstrates, only some of the zoemes actually correspond to a
single noun in BH. By contrast, other zoemes are not precisely named. Some
comprise two distinct groups of animals that do not share a common name
outside the sacrificial texts, and the authors of these texts did not deem it
necessary to coin new terms for them. For example, the author of Exod 12:5
(in connection with the paschal ritual) knows of no economical way to
designate the zoeme f웧––immature male from the flock, including both s웧
and g웧––and hence refers to this category in a roundabout way, using several
words, H$ oe , RK m, and HN
mZ m$ ie , and adding for the sake of clarity, OY$
mo-IB oi m
Bd
vKx
H IM i
OYZ
d
iEiH
m IM i
i W. On the other hand, there are categories of animals which, though
offerable, and sometimes even denoted by a particular name, are not attested
as zoemes, inasmuch as no particular laws pertain to these groups. For
example, the category G웨 (mature female caprine, “nanny-goat”), though
clearly consisting of animals fit for sacrifice in P, is “too narrow” to constitute
an attested zoeme, since even when a nanny-goat is offered, for example as a
OYMimLv
$o XB xeZ in accordance with the law in Lev 3:12 (where ZE a = ), it is only
recognized as part of a larger category, (caprines of any age or sex).
i TE
Similarly, it is possible that DW d x (English “mature billy-goat,” i.e., mature
22
male caprine, G웧), though included in Table 1, is too narrow a category
to constitute a zoeme in P, even though it is mentioned along with several
animals donated to the sanctuary in Num 7:17, 23, and so forth. The zoeme
donated by the chieftains is in fact a , to be offered as a wellbeing offering
(see for example the law pertaining to a caprine wellbeing offering in Lev
3:12–16). On this occasion, the caprine happens to be a mature billy-goat
(G웧), probably the most expensive kind of , but zoemically it was thought
of no more specifically than as a general caprine. Thus, only if one assumes
that the lists in Numbers 7 are prescriptive in some sense is it justified to
i
consider DW T
dE x a zoeme in P.
“Wide” and “narrow”––that is, general and specific––zoemes may appear
side by side in a single list, even if one is a subcategory of the other.
For example, the chieftains offer both a RP xd (B웧) and OYiN x$vo RQmd
mB (2,
21
For a comparable phenomenon in Ugaritic ritual texts, see Levine 1963.
22 i TE
Thus commonly in BH, though in a number of cases, OYDiW d x denotes F웧 (see Gen
31:10, 12).
11:50:13:06:14 Page 40
Page 41
Zoemics 41
Num 7:15–17, 21–23, etc.).23 This does not imply that the RQ B are zoologically
md
m
distinct from the RP d ––for instance, that RQ
x mdB implies female specimens, as
m
LXX presumes. The RQ md
mB in this narrative may very well have happened to be
d. In fact, v. 87 suggests that the offered RQ
bulls, just like the RP
x md
mB were male,
contrary to LXX.24 However, from a zoemic point of view, the same kind of
animal that was offered as a B웧 (mature male bovine, “bull”) in v. 15 would
be considered a (bovine of any age or sex) if offered as a wellbeing offering,
as in v. 17.25
2.2.3.2 A Note on the Binary Pattern

The binary pattern described above is by no means a universal characteristic
of sacrificial systems. The sacrificial systems practiced in diverse cultures
display a variety of structures, not all of whose zoemes are easily reducible
to a binary pattern. For example, a list of the daily sacrifices to the gods of
Uruk contains such zoemes as (1) a first-class, fat, clean ram fed with barley
for two years; (2) a fat, milk-fed kalū-ram; (3) a large bull; (4) a milk-fed
bullock; (5) a fat ram (not barley-fed); (6) a lamb; (7) a grain-fed duck; (8) a
(relatively cheaper) duck; (9) a crane?; (10) a wild boar; (11) a marratu bird;
(12) an ostrich egg; and (13) a duck egg.26 Although theoretically this system
23
See p. 000. According to the story in Num 7, every day a chieftain from a different tribe
presents an identical gift before the altar (x
Xda
BvM
Zd
ixH YN
aPviL, 7:10). The gift of each chieftain consists
of three utensils of precious metal, an unknown quantity of semolina mixed with oil, a ladleful
of incense, and twenty-three zoemes (see 13, 19, etc.). Since the gift consists of metalware as
well as sacrificial material (compare the carts and draught animals in vv. 3–8), and since it
is described as being presented before rather than on the altar, it appears that these materials
were set aside for future use rather than sacrificed on the spot. Thus, the occasion described
in Numbers 7 is a protracted “IKEA-week” of sorts, in which each chieftain furnishes the
tabernacle with metal and zoemic materials. The fact that some of the zoemes are designated
as wholeburnt offerings whereas others are designated as wellbeing or purification offerings
suggests that each zoeme was reserved for a specific type of offering. As one might expect,
the 웧 are reserved for purification offerings, whereas the B웧, S웧, and s웧 are reserved for
wholeburnt offerings, in accordance with the custom of using these zoemes for public offerings
of these types (e.g., Num 28–9).
24
LXX consistently translates OYiN x$
vo RQmdB in this chapter as δαµάλει δο (“two [mature
m
female] cows”), whereas RP xd is always translated µσχο (“young bull”). The incentive for the
(mis)translation of RQ B is an attempt not only to reconcile the lists in Num 7 with the law in
md
m
Leviticus 3, which indicates that female specimens are acceptable as wellbeing offerings, but to
make a point of the law in Lev 3. See n. 000. Theoretically, one might argue that LXX is not a
mistranslation, if the word OYRiP d has two distinct denotations in the summary, vv. 87–8: in the
m
formula OYRP R$E OYN$ HLWEL RQBH LK, the term RP xd denotes B웧; whereas in the formula
OYRP HEBRAW OYR$E it denotes . This is highly unlikely, though it is sometimes the case that
the masculine may designate male or female animals indiscriminately, even where there exists
a lexical differentiation between the two sexes (e.g., RWMX or OYRWMX, even though the term
IWTA exists).
25
The same logical relationship exists between LG a and RWo$o in Lev 9, on which see below,
eE
and probably between RYE $ (웧) and DW i TE
i om d x (G웧, a subcategory of RYE i$ om ) in the same list in
Num 7.
26
ANET 343–5.
11:50:13:06:14 Page 41
Page 42
could also be reduced to a binary pattern, the task would require considerable
exertion, and the results would probably appear forced.
The Israelite zoemic system appears extremely unvaried when compared to
others in the ancient Near East, as the ritual texts consider only five animal
species fit for sacrifice: three species of domestic livestock (cows, sheep, and
goats) and two species of birds (pigeons and turtledoves). By contrast,
lists compiled from a survey of ancient Near Eastern ritual texts––though
admittedly, the findings are not easily comparable to those appearing in
Israelite and Jewish traditions––are significantly longer, exceeding a dozen
animal species in a single list. Though domestic cattle––bovines, ovines,
and caprines––are standard sacrifices throughout the ancient Near East,
sacrificial inventories from Mesopotamia, Anatolia, and the Levant display a
wider variety of animal species than the Israelite system.27
Another outstanding characteristic of the Israelite priestly texts is that only
three criteria (taxon, sex, age) are used to differentiate zoemes,28 thus greatly
limiting the number of possible zoemes that can be generated within the
system (see the , Zoemics rule 3). By contrast, some Mesopotamian
lists include such distinctive features as the diet of the animal, its color, and the
length of its hair,29 and in other cultures, distinctive features defining zoemes
may include color and color-patterns.30 This tendency towards limitation
is apparently only partly an innovation of the Israelites. It appears to be a more
general tendency, reflected in ritual texts in Northwest Semitic languages,
particularly in Ugaritic inventories and Punic tariffs.31 Thus, the Israelites
27
For a survey of Mesopotamian material, see Scurlock 2002 (particularly pp. 389–92, where
the lists include bandicoot rats, gazelles, deer, and mountain rams) and Scurlock 2006. In Ugarit,
though the standard offerings are domestic cattle and particular birds, other animals are men-
tioned on rare occasions. See the extensive treatment in Pardee 2000, 2002. A Hittite birth ritual
(KBo XVII 61, see Beckman 1983:42–3) mentions the burning of a wolf’s head (“burning” is a
standard term used in sacrificial contexts): [na-aš-ta SA]G.DU UR.BAR.RA-ma ar-ha wa-ar-
nu-mi (Obv. l. 19, “and a wolf’s head I burn up”).
28
The requirement that an animal be unblemished is presented as an additional requirement
in several cases (though never with birds, and cf. Lev 22:21–4), and is not part of a systematic
classification of sacrificial animals into categories. There appears to be little or no dependency
between the various distinctive features in zoemics; cf. Clements 1985 on the interdependency
of some distinctive features in phonology.
29
Note that among cattle, the category “milk-fed” (as distinct from “barley-fed”) reflects an
age group that is a subcategory of “immature,” referring to very young specimens. See Scurlock
2002.
30
In other cultures, distinctive features defining zoemes may include color and color-
patterns. In the Vedic context see for example Taittirı̄ya Sam . hitā V.5.22 (TSKY 86–87, “a black-
necked one for Agni, a ewe for Sarasvatı̄ . . . ”); and in the context of the Greek Sacred Law (on
this term, see Parker 2004) see Lupu 2005:328–9 et passim, particularly with regard to the
distinction between black and white specimens.
31
See Pardee 2000 (particularly the convenient summary found in Appendix 1 D, 2.1024–
1051), 2002; and Appendix A in Lupu 2005. For a survey of archaeozoological findings from the
region, see Clason and Buitenhuis 1988. Concerning the findings of deer bones in a north
Israelite sanctuary, see Milgrom, Leviticus 1.723. For the Emariot texts, see Arnaud 1985–7.
11:50:13:06:14 Page 42
Page 43
Zoemics 43
seem to have adopted a general tendency towards restriction at the zoemic

level, namely by reducing to an extreme minimum both the inventory of
species fit for sacrifice and the distinctive features.32
2.3 DIACHRONIC ZOEMICS
2.3.1 Attested, Potential, and Ungrammatical Zoemes
In order to demonstrate how the zoemic system operates from a diachronic

point of view, let us define the following terms on the basis of the zoemic map
above (Fig. 3), abstracted from the Priestly pentateuchal texts:
Attested zoeme: a zoeme that appears in an ancient text. For example B웧 is
an attested zoeme in P since the rule “required as a high priest’s purifica-
tion offering” (Lev 4:3) applies to this set of animals. As Table 1 reflects, the
number of zoemes attested in P is on the order of 20 to 30.33
Potential (grammatical) zoeme: a category that is not attested in a text (e.g.,
in P) but which (1) can be designated by a combination of one or more
blocks in the zoemic map and (2) does not overstep the constraints
placed on possible combinations of blocks, described in the . All
potential zoemes are grammatical. For example, T (“adult turtledove”) is
not a zoeme in P, but its formation can be arrived at by combining the two
blocks T웧 and T웨, a combination that does not contradict any of the rules
in the . It so happens that this zoeme is attested in rabbinic
literature, so that one might say that T is a potential zoeme in P, but later
becomes “actualized” as an attested zoeme in post-biblical sources.
However, many potential zoemes (such as S웨) are attested in none of the
ancient texts discovered so far.
Ungrammatical zoeme: a combination of blocks that is not licit according to
the rules explicated in the Zoemics section of the . For example,
it is expected that no special set of rules would apply exclusively to mature
female bovines (B웨) and immature male turtledoves (t웧). The rules of
the  are designed to agree with this, with the result that the
combination {B웨, t웧} is ungrammatical and therefore unlikely to appear
in any text.
32
The reasons for this restriction are unclear, though they may be related to an analogous
restriction in the Israelite dietary laws. See Milgrom, Leviticus 1.718–736, particularly 723. All
of the species offered according to the Priestly texts––perhaps with the exception of the turtle-
dove––are domestic species. See p. 000.
33
The reasons that the number is somewhat variable are, first, that some are contested, as
discussed below, and, second, that one could create artificially many more, as demonstrated
above, n. 000.
11:50:13:06:14 Page 43
Page 44
Non-zoeme: a category of animals that cannot be described in terms of

the blocks of the zoemic map (for instance, “squirrel,” “red heifer,”34 or
“suckling black lamb”). Naturally, some non-zoemes are more alien to the
system than others: a squirrel is unofferable, whereas a suckling black lamb
may certainly be offered––for instance, as a daily tāmîd offering. However,
if it were, it would be offered as an s웧, for its color and extremely young
age would be irrelevant from a zoemic point of view.35
2.3.2 The Operation of Diachronic Zoemics
Many sacrificial texts dependent on the pentateuchal system––that is,

promulgated by authors versed in the text of the Pentateuch and bound by
its laws––nevertheless contain evidence of radical changes in the details of
sacrificial rituals. If the process of shifts and alterations in a ritual sacrificial
system is viewed as a result of conscious intervention, then we may conclude
that in the centuries following the acceptance of the pentateuchal sacrificial
texts as legally binding, writers such as Josephus, Philo, the authors of the
Temple Scroll, and rabbinic authorities still had the freedom to alter many of
the elements of sacrifice.36
As a rule, these authors introduced new elements, eliminated others, and
sanctioned ancient practices unmentioned in pentateuchal law, creating
sacrificial systems that differ, often radically, from what the pentateuchal
law prescribes. For example, the identity of the slaughterer (priest, Levite, or
layperson), the location of the priest in relation to the altar (for example, on
the altar or beside it), and the manner of applying the blood of quadruped
wholeburnt offerings to the altar (poured on top, applied to its walls, or
dashed on two of its corners)––all were subject to changes throughout the
second half of the first millennium . We will examine many such
examples in the chapters to come, particularly in Section 5.5 of .
The inventory of sacrificial materials was also subject to certain changes. The
inventory of non-zoemic elements (such as wine and grain), which does not
fit neatly into a binary tree like the zoemic system,37 was expanded to include
sacrifices unattested in P. This expansion was achieved either by the inclusion
34
This animal category is not a sacrifice in Num 19 or anywhere else in P, pace Marx 2005:59
n. 11.
35
Leviticus 22:27 contains a general law that a quadruped should not be offered before
the eighth day of its life. This rule is formulated as an “amendment,” so to speak, to the system,
and it does not substantially alter the system as a whole.
36
See for example Marx 1994:155–65 concerning late Second Temple textual traditions
pertaining to vegetable offerings (in particular, see 157 concerning TS).
37
For an attempt to identify a tripartite structure in P’s inventory of materia sacra, see Marx
1994:47.
11:50:13:06:14 Page 44
Page 45
Zoemics 45
of materials unmentioned in P (such as water)38 or by a process of reanalysis,

through which materials already mentioned in P as subordinate offerings
(such as wood [OYC E, not considered an offering in P, though placed on the
ia
altar], frankincense [HN v ], wine [IYiYx], oil [IM
mWoBL ee$o], and even salt [XL
xMe ])
acquired an independent status, becoming offerings in their own right.39
The zoemic system, by comparison, remained relatively stable. As noted
above, the pentateuchal ritual texts themselves reflect a shift in the zoemic
inventory vis-à-vis other ancient Near Eastern cultures, since the range of
animal species considered fit for sacrifice in ancient Israel was restricted
exclusively to domestic cattle (bovines, ovines, and caprines) and two species
of birds, at least one of which (pigeons) was domestic.40 Following this initial
restriction, however, the range of species considered fit for sacrifice did not
undergo significant changes. Ezra and Nehemiah, Chronicles, the Elephantine
documents,41 and extra-biblical sources from the first centuries  and ,
including Qumranic and rabbinic literature, all reflect the same limited group
of animals acceptable for sacrifice. As for the other criteria pertaining to
animals––namely, sex and age––it appears that a similar zoemic inventory
is shared, even in its details, by P, H, Ezekiel, Ezra and Nehemiah, and
Chronicles.42 It is noteworthy that whereas these sources (including diverse
strata of P and H) differ significantly in the numbers of animals sacrificed on
special occasions, and in the relationships between animal and non-animal
materials constituting each type of sacrifice, there is almost complete accord
among them as far as the zoemic classification system is concerned.43
The main difference between these sources in this respect is the scope of
the material: the code of Ezekiel, for example, makes no mention of female
zoemes, though it probably acknowledges their place within the system.44 The
reason for this omission is that Ezekiel 40–8 is concerned primarily with
public calendric wholeburnt and purification sacrifices, and such offerings
are always male (as in P).
However, even in this relatively stable system, several changes did occur,
particularly towards the end of the first millennium , but also in rabbinic
38
m. Sukkah 4:1. Non-Priestly ancient Israelite traditions mention the libation of water
(1 Sam 7:6; 2 Sam 23:16).
39 40
On these see nn. 000–000. See p. 000.
41
“The evidence from the Elephantine documents is admittedly meager, reflected in a brief
sacrificial list including “sheep, ox, goat” (but no birds) in Cowley 33:10. See Porten and Yardeni
1986:78, A4.10:10.”
42
This cannot be said of other non-P texts such as 1 Sam 7:9; Ps 66:15. See for example
Rendtorff 1967:115–18.
43
For a different opinion concerning RWoT d , see Staubli 2008.
44
See Ezek 45:15: OYM imLv
$oiLv
W HL viW . . . OYiT
m WoEL x AMmx
d H-IM
i IAC
do x
H-IM
i TX
xAx -H$ W (“and one of
oe v
every two hundred members of the flock. . . for wholeburnt and wellbeing offerings”). If Ezekiel
agrees with P that female specimens may be offered as wellbeing offerings, then TX xAx H$oe in this
verse probably includes female specimens. Note that TX xA oe is grammatically feminine here
x H$
(though not everywhere in BH), but this in itself does not imply that it refers to female specimens.
11:50:13:06:14 Page 45
Page 46
literature as late as the Middle Ages.45 The changes that occurred within the
zoemic system were usually not introductions of new elements, but rather
shifts of the old elements in two interrelated ways: (1) relatively “wide”
zoemes were narrowed down by excluding some of their constituent blocks,
and (2) the binary pattern underlying P’s system was extended to create new
zoemes unattested in P. We will examine some of these shifts immediately.
Note that the changes in the zoemic system that are of interest to us are
not simply changes in nomenclature––thus, the fact that in the Temple Scroll
H$ oe denotes (52:5, 13) whereas in P H$ oe = is a lexical shift, not a zoemic
shift. To see that this is so, note that the change does not affect the sacrificial
system promulgated by the author of the Temple Scroll when not quoting
biblical passages. Similarly, the fact that P refers to s웧 almost invariably
as WoTN m$ ie o$e
vo-IB Bd
eK whereas Ezekiel, Chronicles, Ezra and Nehemiah, and
Qumranic and Mishnaic sources refer to it simply as o$e BdK is a semantic shift
e
(o$e K = 웧 o$e
Bd
e Bd
eK = s웧), not a zoemic shift, as the same zoeme persisted
under a different name. Zoemic shifts, as we shall presently see, are instances
in which a change occurs in the scope of a category of animals to which a
particular sacrificial law pertains.
2.3.3 Zoemic Shifts
The priestly legislation requires that a quadruped offered as a voluntary HL

m WoE
(wholeburnt offering) be an unblemished male (Lev 1:3, 10), with no restric-
tion on the age of the animal. However, ancient tradition narrowed the HL m WoE
category to include young specimens only, at least for some of the species.
It has been argued that this development is evidenced already in LXX,
whose terminology in Leviticus 1 suggests animals of a relatively young age.46
The choice of words in LXX may not be decisive proof of such a development,47
45
Aside from the many sporadic discussions in the Sifra and in later Talmudic literature,
numerous discussions of the zoemic scheme appear as well (1) in the first chapter of Tractate
Parah (Mishna and Tosefta); (2) in a systematic manner, at a later period, in Maimonides (twice:
first in his Commentary, Kafah. 1963; and later in HMQ); and (3) in the work of the
seventeenth-century scholar R. Yom Tov Lipmann Halevi Heller (the introduction, dependent
on Maimonides, is entitled $DWQB RMWX). These and other rabbinic schemes depend on
rules of inference different from those used in historical-philological study. The application
of these rabbinic rules may lead to conclusions that are highly unlikely from a historical-
philological point of view. Thus, while these post-biblical texts are important evidence of
late developments within Σ, they cannot be relied upon uncritically for the interpretation
of terminology in P.
46
This is the opinion of Altshuler in Feldman 2000:292 n. 593.
47
For RQ
mdm
Bx ie , LXX has µσχο; for OYB
H-IB i$omd
v
Kx i , απ . . . τν αρνν; and κα τν ρφων
H IM
for OYZ
d
iEim iiW. The phrase τν αρνν implies , not s; µσχο does in fact denote a young bull,
H IM
but one must keep in mind that LXX had to make some choice here–– RQ md
m
Bx ie cannot be
H-IB
concisely translated in a better way––and in any case the translators did not choose µοσχάριον,
“young calf.” Concerning ριφο, Altshuler is apparently correct.
11:50:13:06:14 Page 46
Page 47
Zoemics 47
but a passage in Josephus definitely is. Independently of LXX,48 Josephus

states that a layperson’s HLm oWE must be either s웧, g웧 (each less than a year
old),49 or 웧 (any age).
Targumic evidence also suggests that, at least in some cases, laypersons were
expected to offer young quadrupeds, rather than quadrupeds of any age, as
voluntary wholeburnt offerings. Though the rendering of OYB i$
omd
v
Kx i (Lev
H IM
1:10) as AYRMA IM and OYZ d
iEi as AYZE YNB in the Targumim is not conclusive
proof of this shift,50 the translation HYZE YNBD HYDG IM in Targum Neofiti
probably is.51 Taken together with Josephus, these texts seem to reflect a halachic
development, which is non-pentateuchal and may reflect common practice.
Similarly, Josephus (Ant. Jud. 3.231) assumes that a layperson’s purification
offering (the OYZ d
iEi TRxYEov or HB
i$ maQvN . . . o$e
Bd
eK of Lev 4:28, 32; the ½WoA HB
d
m$ov d
iK
OYZ
d
iEi TRxYE ov of Lev 5:6; and the d
i$ HT mm$Nvo½TB d a of Num 15) is necessarily
x ZE
under one year of age––that is, immature. This view is corroborated by
rabbinic halacha and may similarly reflect common practice.52
One can hardly overstate the fact that reductionist materialistic explanations
do not suffice to explain ritual systems.53 However, the shifts
x
(age unspecified immature) reflected in the four examples mentioned
(웧 s웧, 웧 g웧, 웨 g웨, and 웨 s웨) may have originated
from pragmatic considerations: the sacrifice of young animals is less of an
economic burden, a particularly significant factor for laypersons (as opposed
to the Temple treasury or high-ranking personnel).
These examples, culled from diverse sources, suggest that this “narrowing”
48
See Ant. Jud. 3.9.11–12. Note that Josephus’ terminology is different from that of LXX
(Josephus has βο and αρνον (“little lamb”) where LXX has µσχο and τν αρνν, respec-
tively). Gallant argues that Josephus saw a contradiction between the general terms in verse 2
(βον, προβάτων) and the terms in vv. 5, 10, and suggested that though all ages are acceptable,
young specimens are preferable (see Feldman 2000:292 n. 593). I see no evidence of this in
Josephus; furthermore, the shift from young bulls (LXX) to mature bulls (Josephus) would
remain unexplained. It is possible that µσχο in LXX is based on a misunderstanding
of RQ md
m ie as referring to young bulls only. However, there is no BH catalyst for LXX’s
B-IB
interpretation of OYZ d i as young goats, and, as noted above, τν αρνν (v. 10) does not imply
iE
young sheep.
49
The Greek is unambiguous: βον κα αρνον κα ριφον: τα!τα µ"ν π#τεια, το$ δ" βο!
φεται θειν κα προ&κοντα. The Latin, however, which translates π#τεια as anniculos,
allows also a different reading, “yearling” (i.e., having attained the age of one year already).
50
Contra Feldman 2000:293 n. 606. The term RMA, often connoting s웧, is a standard
translation of B$ K in Targumic Aramaic, even where the specimens in question are mature,
oe d
e
e.g., Gen 30:32, 35. So too, AYZE½YNB does not necessarily imply young goats (g웧); see n. 000.
51
See, for example Neof. pp. 5, 161.
52
See Sifra, Weiss 22a, HTN$½TB HAB WZ FA HTN$½TB HAB HRZ HDWBE TAUX HM
(“just as a purification offering for idolatry is in the first year of its life, so should this one be in
the first year of its life”). The statement is made concerning the HBmaQvN . . . o$e
Bd
eK of Lev 4:32, but it
pertains also to the HB d
m$ iK mentioned in Lev 4:28 and the OYZ
ov d d
iEi TRxYE i$ov and HB d
m$ iK mentioned
ov d
in Lev 5:6, as context implies.
53
See Lévi-Strauss 1966 for a critique of Radcliffe-Brown 1952 and Malinowski 1948. A
somewhat more balanced view is found in Kunin 2004:36.
11:50:13:06:14 Page 47
Page 48
is not a chance process. There is method to the madness of sacrifice, at least as

far as zoemic classification is concerned.
2.3.4 The Binary Pattern Extended
The evidence suggests that the process of differentiation between only two age
groups persisted, with a few important exceptions (see Jubilees 32:6, and the
opinions cited in t. Parah 1:1),54 well into an era when sacrificial rituals were
no longer performed in Israel.
First, the distinction of age was carried over from quadrupeds to birds.
The text of P does not imply that turtledoves must be mature or that pigeons
must be young;55 it merely lists OYRiTo YT
a
d$vo and HNmWoY½YN
aBv YN
a$ vo (and in one case
RTo -WoA HN ie ) as interchangeable zoemes (though not intermixable: a
mWoY-IB
combination of one pigeon and one turtledove is nowhere implied). How-
ever, several ancient Jewish texts suggest that the transformations T and
y took place. The transformation y is attested in LXX (e.g., Lev
12:6)56 and in Jubilees (6:3),57 and rabbinic tradition reflects the shift T
as well as y.58
54
See above, p. 23. It appears that Jubilees 32:6 refers to zoemes of three different age groups
among goats. According to R. Eleazar, there is a zoemic distinction between RP d and RQ
x mdm
B½IB ie RP
x
d
––the former being two years old (i.e., in the first three years of its life), the latter three to five
years old (i.e., in the fourth to sixth years of its life). This dichotomy is clearly not implied in P
(compare Num 29:13, 17 with 14, 18, 32 and with 17, 20, 23, 26, 29, 32), and in any case it is
probably purely theoretical. Since R. Eleazar does not deny the existence of the zoemes LG eE a and
HNm$ ie LG
mo ½IB eEa that are younger than RQ md
m ie RP
B½IB d, there appear to be more than two age groups for
x
bovines. It is noteworthy that, according to R. Eleazar, the qualification RQ md
m ie “lowers” the age
B½IB
of the RP d; this corresponds with the case of HN
x mWoY½YN
aBd in rabbinic literature, which was rendered
v
as referring to young doves (corresponding, as we noted, with late Second Temple reality). On
the other hand, where R. Simeon distinguishes between LG a, which should be under a year old,
eE
and RQ md m ie LG
B½IB eEa, which should be in the second year of its life, it appears that RQmdm ie raises the
B½IB
age of the LG eEa. This is almost unavoidable, since the textual basis for the dichotomy is
unambiguous: the LG a (Lev 9:3) is specified as being less than a year old (HN
eE m$mo½YN
aB
d $oB
v eeWKm LGeE W).
av
There are several ways to solve this “problem,” but it is unnecessary to do so here. In any case,
Neusner’s translation of this halacha (Neusner 6.172) in the Tosefta cannot be accepted.
55
See n. 000. Note, too, that the author of Gen. Apoc. (10:15) does not hesitate to refer to the
turtledoves offered by Noah (upon disembarking from the ark) as ANYNP$ YNB (not simply
IYNYNP$ or AYNYNP$), although the turtledoves are clearly not young.
56
Jubilees 6:3 distinguishes between ’gwala rəgb (young pigeon, the equivalent of HN ie ) and
mWoY½IB
mant.ēta (the equivalent of RT o , not necessarily young). I thank Dr Michael Segal for assisting me
d
with the Ge‘ez. See also Tisserant 1921:80–1.
57
In all instances except for Lev 1:14, the law in LXX distinguishes between τρυγνα
(turtledoves) and νεοσσο$ περιστερν (fledgling pigeons), or––in Lev 12, where only one bird
is required––between νεοσσν περιστερα̃ and τρυγνα.
58
For instance, several sources (Sifra, Weiss 83b; m. Zeb. 7:5; 14:2; m. Me‘il. 3:4 speak of
INMZ RBE$ HNWY YNBW INMZ EYGH AL$ OYRWT, turtledoves too young and pigeons too old (to
be offered); similarly, m. Zeb. 3:5; t. Zeb. 3:9; m. Me‘il. 3:5 all speak of eggs found in turtledoves
(but not in pigeons). The reason for this is not that turtledoves are female and pigeons male, but
that the latter are too young to have eggs.
11:50:13:06:14 Page 48
Page 49
Zoemics 49
Clearly, there is an exegetical aspect at play here: the term YN d may be

aB
v
understood, with some exertion, to denote “young.” However, it is clear
that the process here was not merely exegetical. If pure linguistic exegesis had
been employed here, only the transformation y would have taken place,
not T. The fact that the latter transformation took place as well suggests
that an internal, implicit logic was assumed: if there is a zoemic distinction
between mature and immature quadrupeds, then there ought to be a zoemic
distinction between mature and immature birds, with an exemplar of each
age bracket; and if it is suitable to limit the pigeons to fledglings, then turtle-
doves, which are termed OYRiWoT d and not RWoTd ½YN
aBd in P, must be mature.
v
These two zoemic shifts probably reflect common practice and are not
merely theoretical. First, y is attested in two early independent sources,
LXX and Jubilees, as well as in rabbinic tradition. Second, the Mishnaic laws
presume––rather than assert––that in the sacrificial system turtledoves are
mature but pigeons immature.59 Therefore, in contrast to the following two
examples, the two aforementioned shifts probably occurred in reality, not
merely in the texts.
As already noted (p. 000), Josephus’ explicit insistence that a quadruped
(at the very least, an ovine or a caprine) offered as a layperson’s wholeburnt
sacrifice must be no more than a year old. In an adjacent passage (Ant. Jud.
3.9.22–3), when describing wellbeing sacrifices (more precisely, thanksgiving
offerings), Josephus notes that the animals offered as wellbeing offerings are
more than a year old (τν πετεων πρεσβτερα). An unbiased reading
of the passage suggests that this condition is requisite, not permissive60 ––in
other words, that wellbeing offerings must be older than one year. Since
this piece of evidence is, as far as I can see, uncorroborated,61 it likely does
not reflect actual practice. However, if it is an artificial construction of a
59
The sacrifice of pairs of birds was apparently quite common in the Herodian temple, as
may be inferred from several sources, e.g., m. Ker 1,7, Luke 2:24 and possibly from a stone vessel
depicting a pair of birds and the word IBRQ, now in the Israel Museum (see Mazar 1969:168–70
and plate 45 no. 5). Note that the etchings are of two birds of seemingly different age groups––
one has a tail, the other does not. If the birds represent a pair (mishnaic ,(IQ then this does not
dovetail with rabbinic evidence according to which both must be of the same age group (mature
turtledoves, immature pigeons).
60
See Feldman 2000:293 n. 606, who notes that Josephus’ specification was interpreted as
permissive by Thackeray. However, it appears that Thackeray’s interpretation is motivated by his
knowledge of Mishnaic and biblical law, where this requirement is lacking. Gallant correctly
notes that since this requirement immediately follows the requirement that the animal be
unblemished, it is clearly requisite and not permissive. It is unclear whether Josephus was
consciously attempting to introduce a new law that he knew deviated from common practice.
It may be a slip of the pen, but if so it is a telling slip.
61
See m. Ed. 7:6; m. Tem. 3:1, OYML$ YXBZ HRP WNL HTYH$ DYEM YNA SYYPP YBR RMA
GXB OYML$ HDLW WNLKAW XSPB HWNLKAW (“Said Rabbi Pappias, I testify that we had a cow as a
wellbeing offering; we ate it on the paschal celebration, and we ate its offspring as a wellbeing
offering on the Festival”). It is impossible to determine the precise age of the second-generation
bovine.
11:50:13:06:14 Page 49
Page 50
theoretician attempting to systematize the Jewish sacrificial practices as he

knew them, then it reveals a similar logical process described just above in
relation to /, running as follows: if voluntary wholeburnt sacrifices of
laypersons must be immature, then voluntary wellbeing sacrifices of lay-
persons must be mature.
This tendency to extend the binary pattern persisted long after the sacri-
ficial system in Israel had ceased to be practiced in reality. Thus, later rabbinic
literature, replete with details of sacrifices that probably never actually took
place, continued to be pervaded by a mental “game” of sacrifice even when
the physical “game” was no longer played. For example, some medieval rab-
binic authorities extended the binary pattern to he-goats, distinguishing
between RYE om , which is in the second year of its life, and OYZ
i$ d
iEi RYE
i$ov , which
should be less than a year old.62
2.4 SOME UNCERTAINTIES AND

WIDESPREAD MISTRANSLATIONS
The schematic reconstruction offered here differs significantly from previous

descriptions of the Israelite sacrificial system in its details and general struc-
ture. The abstract sigla representing zoemes, rather than the Hebrew terms
denoting them, serve as the organizing principle of the inventory. This form
of presentation is preferable from a theoretical point of view, since many of
the terms underwent semantic shifts in biblical as well as post-biblical
Hebrew, whereas logical divisions like mature–immature and male–female
remained relatively fixed.
In modern scholarship, there is a tendency to assume that the age and sex
of the zoeme are always specified, even when there is no real indication of
these details in the text. Such presumptions have led to much confusion and a
number of erroneous determinations. Other scholars have tended to give up
prematurely in some cases, claiming, for example, that the ages of the animals
62
R. Joseph Caro in his commentary (HN$M FSK) on Maimonides, HMQ 1,14, suggests
that a zoemic distinction exists between OYZ d
iEi RYE ov and RYE
i$ i$om ––the former denoting an
immature male caprine, the latter a mature male caprine. His motivation was to reconcile
several rabbinic sources that indicate that RYE i$om denotes an immature male goat, including
b. Yoma 65b, 66a (following Sifra), with other sources that indicate that RYE om denotes a
i$
mature male goat, e.g., m. Men. 13:7, b. Men. 91a. This harmonistic approach led him to
highly unlikely and artificial conclusions, including a textual conjecture regarding
Maimonides’ Code: Maimonides had written OYT$ IB OYZE RYE$ (i.e., OYZE RYE$ implies
[an animal] in its second year), whereas R. Joseph Caro suggested inserting three words to create
OYT$ IB –– RYE$ ;HN$ IB –– OYZE RYE$ (that is, OYZE RYE$ implies [an animal] in its first year,
whereas RYE$ implies [an animal] in its second year).
11:50:13:06:14 Page 50
Page 51
Zoemics 51
cannot be determined, even when sufficient evidence exists for reliable

determinations.63
Some zoemes can be identified with a relatively high degree of certainty.64
For example, it is clear that LYiA
x is a mature male sheep (S웧). This denotation
is consistently found in BH outside the technical priestly language65 and
appears to have been taken unaltered into the priestly sacrificial ter-
minology.66 Early post-exilic sources (where Aram. IYRiK vdi replaces OYL
D a ),67
i YA
the Versions, and ancient Jewish sources are almost unanimous on this matter
as well.68
Other zoemes are more difficult to identify, and it is possible that their
designation differs from text to text. There are therefore several uncertainties
and some widespread mistranslations, which will be noted. Some of the
mistranslations have been noted in the past, and thus do not require our full
attention, though a sample of them will be treated briefly, either because
some aspect concerning them has escaped scholarly notice or because they
continue to be a source of confusion; other mistranslations have remained
entirely unnoticed and will be discussed in detail.
oe as
2.4.1 H$
The term H$
oe , despite an ancient tradition of translation, does not connote
the immature sheep or goat (f) in P, or even in BH in general.69 Rather,
63
See Knierim 1992:24 n. 21. Levine (Numbers, 2.377) confesses that there is much
uncertainty in identifying the age groups of sacrificial animals. For a balanced view, see Marx
2005:52–3 n. 1.
64
LXX is relatively consistent and can serve as a good guide in most cases, though it has some
extreme peculiarities, to be discussed. One should keep in mind that the Greek translators
sometimes translate H$ oe () not as πρβατον (LXX’s standard term for “member of the flock”)
but as a sheep or a goat (e.g., α*µνο, χµαρο), according to context (e.g., Lev 12:8, where the
H$oe the author has in mind is clearly the o$e BdK mentioned above). Note the double translation of
e
H$ oe in Deut 14:4 (OYZ d
iEi -H$
oa v
W OYB
i$ K-H$
om v oa ), first as α*µνο, then as χµαρο.
65
See Gen 32:15 et passim and Péter-Contesse 1992.
66 ie $oB
See Lev 23:18; Num 28:11, et passim, where LYiA x is contrasted with WoTNvo-IB
m$ edK.
e
67
Ezra 6:9, 17; 7:17, where IYRiK
vdi are contrasted with IYRMA (lambs).
D
68
See for example Ezek 46:4, where LYiA x is contrasted with o$e BdK; m. Parah 1:3. In LXX, κρι
e
(ram) is the consistent equivalent of LYiA x (Num 28:11 et passim). For a rare exception to the
usage of the term LYiA x in MH see n. 000, but it is evident that this semantic shift in the Hebrew
language did not lead to a misinterpretation of the BH term LYiA x.
69
For instance, NEB on Gen 22:7 etc. “young beast”; RSV, JB “lamb.” Some have opined that
H$ oe alone can designate a sheep in BH––(rather than only ), e.g., Dorival 1994:127; or simply
a young flock animal (f), e.g., Jastrow 1943:1526. The mistranslation “sheep” may be due in
part to a diachronic lexical development (narrowing or limitation) that occurred within
Hebrew, since H$ oe eventually came to denote sheep () as distinct from caprines (e.g., TS
52:5,13, ZEW H$W RW$; see p. 000). Admittedly, there is one clear case in which the term IAC o is
used in BH to denote ovines alone as distinct from caprines (1 Sam 25:2)––but H$ oe is never
used in this way. The misunderstanding of H$ as “f” and “s”––particularly in Gen 22:7––may
11:50:13:06:14 Page 51
Page 52
oe refers to any ovine or caprine (), regardless of age or sex.70 Despite

H$
assertions to the contrary in some of the lexica and commentaries, the only
real potential source of confusion in Priestly literature is Num 15:11.71
2.4.2 RWo$o as
The term RWo$o in BH, though grammatically masculine, does not denote
B웧,72 as in later Hebrew sources.73 Rather, it denotes ––it is a generic term
for a single domesticated bovine of any age or sex (as at Lev 4:10 referring
back to 3:1; 7:23, et passim).74 In fact, in one passage in H (Lev 22:28), the
animals to which it refers are almost certainly female.75
have been reinforced by the perception of a H$ oe as the prototype young Isaac (who is pre-
sumably young; however, he is substituted by a H$ oe , , which happens to be an LYiA
x , S웧), and
the perception of Jesus as agnus dei, analogous to the paschal offering (explicitly in 1 Cor 5:7
and implicitly John 19:36), which is in fact immature (more precisely, f웧).
70
See Waschke 2004 and the bibliography cited there.
71
Even this passage is not an exception to the rule that H$ oe = , since, as we shall presently
see, mature goats (male and female) and mature ewes (female) are also indicated by H$ oe in
this verse. Cf. JB, NEB “lamb or kid”; RSV “each of the male lambs or the kids.” This point
seems to have escaped the notice of Milgrom, who writes (1990:120) in his commentary on
JPS’s (correct) translation of H$oe as “any sheep or goat,” “literally, ‘with any kid, either sheep or
goat,’ ” and “Hebrew śeh is the young of either sheep or goats.” In fact, however, the term H$ oe
here, as everywhere, denotes . In this verse, P juxtaposes H$ oe with LYiA
x , and since LYiA
x indicates
mature animals, it might appear that H$ oe denotes only immature specimens. However, upon
closer reading, it is clear that this is not the distinction that is implied. Vv. 5–6 distinguish
between the laws of the LYiA x (S웧) and the o$e Bd
eK (s웧) but say nothing of the other members of
the sheep family, and nothing at all about goats. Subsequently, v. 11 refers to these in general,
stating that whereas one set of laws applies to rams (S웧), another set of laws (the same set listed
in vv. 4–5 in connection to s웧) applies to all other members of the flock that are not rams
( with the exception of S웧). Since in P, and in BH in general, there is no simple way
to designate “––S웧” (i.e., all members of the flock minus S웧), H$ oe is used, meaning “any
other H$ oe .” Levine in his translation (2.386) translates correctly (inserting “other”), but his
commentary ad loc. (p. 392) is strangely confusing. It is possible that another source of con-
fusion in this verse is the use of OY$oi m
Bd
vK in the sense of “ovines” () rather than “male lambs”
(s웧), which is rare in P, but not unique (see §2.4.3).
72
This error is found, for example, in Elliger 1966:70; Knierim 1992:24 n. 21.
73
In MH, it is clear that alongside the use of RWo$o in accordance with its biblical sense, , it
is sometimes assumed that RWo$o designates only male bovines (웧), e.g., m. Kid. 1:6. Note,
moreover, the ensuing confusion that led to such a sentence as RW$ YWRQ WMWY IB RW$ (“a
newborn RWo$o is [still] called a RWo$o,” BK 65b), where the first RWo$o is apparently colloquial, the
second technical, following biblical usage.
74
See Péter-Contesse 1985 for a discussion of all of the occurrences of RWo$o in BH. Péter-
Contesse conclusively demonstrates that RWo$o in BH denotes a bovine of any age or sex,
though the word itself is masculine, just as H$ oe (grammatically either masculine or feminine,
regardless of the sex of the animal in question) designates any single domesticated ovine or
caprine. As is well known, the collective form of RWo$o is RQ mdB, and that of H$
m oe is IACo (e.g.,
RQB H$MX, IAC HEBRA, never TWY$/OYY$, and only once OYRW$, in Hos 12:12).
75
See the commentaries, as well as b Hul. 79b. See also TS 52:5–7 and 4Q396 (4QMMTc)
1:2 which, though fragmentary, almost certainly interprets this verse as DLWH TAW OAH T[A]
DXA OWYB (“the mother and its offspring in a single day”).
11:50:13:06:14 Page 52
Page 53
Zoemics 53
Leviticus 9:4 deserves special attention in this context. In this chapter,

which describes the events that took place following the ordination of Aaron
and his sons, Moses commands the Israelites to take a RWo$o and an LYiA x for a
wellbeing offering. Since RWo$o consistently denotes in BH, it appears that
the two zoemes that were required were S웧 (mature male ovine) and
(a domestic bovine of any age or sex).76 The appearance of one “wide”
zoeme and one “narrow” zoeme is asymmetrical.77 However, the appearance
of a wide zoeme alongside several narrower zoemes (vv. 2–4) is not unique
in P. As noted (p. 000), a comparable list is found in Num 7:15–17, 21–23,
33–35, and so forth, where OYiNx$
vo RQ
mdB (2) are listed alongside H$
m mdoi
MqX OL
i YA
a
i TE
(5S웧), H$ mdoM
iqX OYDiW d x (5G웧), H$mdoM
iqX HN
m$mo-YN
aBd OY$
v oi m
Bd
vK (5s웧), and even
RQmd
m
B½IBie DXmeA-RP
xd (B웧). In Leviticus 9, the text does not reveal whether, at
this particular historical point in the narrative, the RWo$o, which could have
been mature or immature and male or female, happened to come in the form
of a young female calf or a mature bull––within the zoemic scheme, it served
as a regardless of the particular specimen selected. This requirement in
9:4 is in accordance with the law elsewhere in P that female specimens are
acceptable as wellbeing offerings. Therefore, here as everywhere else, one
should translate RWo$o as “bovine” or “cow” (generic), contrary to many
translations (NEB, “bull”; JB, RSV, JPS more correctly “ox”––a term that may
denote male as well as female specimens).78
2.4.3 o$e
Bd
eK vs. B$
oe d
e
K
Whereas in BH in general o$e BdK and B$

e K (formed by phonetic metathesis)
oe d
e
are apparently interchangeable, in the technical language of P the two were
molded into distinct technical terms: o$e K is almost always used in P to
Bd
e
denote young male ovines (s웧), whereas B$ K is used exclusively to denote
oe d
e
79
ovines of any age or sex (). This procedure of coining technical terms out
of general nouns is exemplified elsewhere in P,80 but it is possible that here it
was carried out only imperfectly: in Lev 4:32 (MT) P speaks of a female o$e K.
Bd
e
Since there is no such thing as a female s웧, o$e BdK in this case must denote ,
e
76
This is true also of v. 18, which purports to be strictly descriptive rather than prescriptive.
77
The designation of as a required zoeme in conjunction with the ram is somewhat
surprising: one might have expected Moses to specify two “wide” zoemes (e.g., and ) or two
“narrow” zoemes (e.g., B웧 and S웧). The combination LYiA xm d ––i.e., B웧 and S웧––is found
W RP
m
several times in biblical texts (e.g., in Num 23:2).
78
Conversely, it is possible that chapter 9 in Leviticus, which differs from other tôrôt in P
linguistically, legally, and in its historical and narrative presuppositions, uses the term RWo$o in a
unique manner, meaning “bull,” as the term is used in later Hebrew sources. See n. 000.
79
Concerning B$ K, the data are unequivocal, as Milgrom has demonstrated (contra
oe d
e
Rendtorff 1967:116; Seebass, Numeri 2.133).
80
See Milgrom 1992 and Meshel 2008 concerning the term JQ eo, and below, nn. 000 and
e$
000.
11:50:13:06:14 Page 53
Page 54
and a female o$e

Bd
eK is then 웨 (any age). In other words, P slips into common
BH on this occasion,81 using o$e
Bd
eK to denote .82 The plural OY$ oi m
Bd
vK implies
on two other occasions, Exod 12:5 and Num 15:11.83
2.4.4 WoTN
m$ ie o$e
vo½IB Bd
eK and o$e
Bd
eK
In P, the term o$e K is consistently qualified by WoTN

Bd
e ie (Lev 12:6; 23:12, 18,
vo-IB
m$
et passim). If the qualifier does not immediately follow, this is usually
because WoTN vo-IB
m$ ie appears the first time the animal is referred to and is
implied in the following verses (as at Exod 29:38–9, OYiN x$
vo HN
m$mo-YN
aBd OY$
v oi m
Bdv
K
RQedB
oxB H$ oe q
ETx
d DXm e
A m
H o$ e
Bd
eKxH-TAe . . .). Notwithstanding the few instances
where OY$ oi m
Bd
vK/$o e K denotes (Exod 12:5; Lev 4:32; Num 15:11), this noun
Bd
e
is found without the qualifier WoTN vo-IB
m$ ie in only two contexts in P––in the
offering of a person purified from scale disease (see Lev 14:10 [MT], 12, 13,
21, 24, 25: OM i YM
iTv OY$
d oi m
BvK-YN
a$vo, followed by o$e BdK several times),84 and in
e
Num 15:5. This last verse is particularly telling, since a comparison of this
verse with the laws in Num 28:5–7, 13–14, and so forth reveals that the
o$e Bd
eK referred to in 15:5 is none other than the zoeme referred to as
HNm$mo-YN
aBd OY$
v oi m
BdvK in Numbers 28–9 (see §3.7.1). It is therefore probable
that there is no zoemic distinction between WoTN m$ ie o$e
vo-IB Bd
eK and o$e BdK ––
e
81
For a similar slip, see Meshel 2008:227.
82
Contra Milgrom, Leviticus 1.252, who believes that P was consistent and suggests that
perhaps the o$e Bd
eK at 4:32 [MT] should be read B$ K (compare Sam HB$K). Rendtorff
oe d
e
(1967:116 n. 2) opines that the graphemes o$e K and B$
Bd
e K are interchangeable. One may
oe d
e
account for all of the data in P by stating that o$e Bd
eK may mean either or s웧, but B$ K is
oe d
e
only (this would account for the plural forms referred to above and to the attestation of
o$eBdK in the sense of , if it is not a result of graphic metathesis as Milgrom believes). It
e
appears that Maimonides became aware of the zoemic distinction between o$e BdK and B$
e oe dK at
e
some point between the composition of his introduction to tractate Zebah.im and the com-
position of HMQ 1. In the Commentary he writes (Kafah. 1963: AY) YP HLLA LAQ AM 'TYX
ONGLA IM OAE IBA WHP OY$BK WA B$K WA HB$K WA $BK IYBARQLA (“whenever God says
in the context of sacrifices $BK or HB$K or B$K or OY$BK, it is a sheep one year old [i.e.,
in the first year of its life]”); but in the Code he changed this to HRWTB RMAN$ OWQM LK
HN$ YNB WLA YRT OY$BK WA HB$K WA $BK (“every place in the Torah where it is said $BK
or HB$K or OY$BK ––these are in the first year of their lives”), having now erased B$ K from
oe d
e
the list.
83
Note the masoretic comment on Num 15:11, according to which the sebirin is OYB$KB.
84
Of these six occurrences, one would expect the qualifier HN m$mo-YN d to appear, as often in P,
aB
v
only the first time the zoeme is introduced––i.e., in v. 10––and perhaps in v. 21. Note that in
v. 10 this qualifier is reflected in LXX (νιαυσου) and in the Samaritan (HN$ YNB). It is
noteworthy that in Ezekiel, Ezra, Nehemiah, and the Chronicles, all of which are very close in
their sacrificial terminology to the priestly law, o$eBd
eK is used without the qualifier WoTN m$vo-IBie in
passages that are parallel to where the pentateuchal law has WoTN m$ ie o$e
vo½IB BdK (e.g., when contrasted
e
with LYiA x , Ezek 46:4–11, OYMT LYAW OMYMT OY$BK H$$); but see 46:13; Ezra 8:35 (alongside
OYRiP
md, OYLi YAa and TAUX YRYPC); 1 Chr 29:21 (alongside OYRiP d and OYL
m a ); and also 2 Chr
i YA
29:21, 32. Compare also Ezra 6:9, 17; 7:17, where IYRMA (“lambs”) appears to be the equivalent
of OY$oi m
BdK, as it is mentioned alongside IYRWT (“bulls”) and IYRKD (“rams”).
v
11:50:13:06:14 Page 54
Page 55
Zoemics 55
that is, the qualification WoTN

m$ ie in this case is not restrictive but descrip-
vo-IB
85
tive. Therefore, despite the potential lexical ambiguity, it is almost certain
that o$eBdK (like WoTN
e m$ ie o$e
vo-IB K), where it does not designate , always
Bd
e
designates s웧 rather than s86 or 웧.87
2.4.5 WoTN
m$ ie as Equivalent to HN
vo-IB m$ ie in P
mo-IB
The qualifiers WoTN vo-IB

m$ ie and d
HT Nvo-TB
mm$ x in BH most probably do not denote
d
“a yearling” (an animal in its second year) but rather an animal that is in the
first year of its life. In any case, Num 7:87 (cf. 15, 21, etc.) conclusively proves
that the phrases WoTN m$ ie and HN
vo-IB m$ ie are equivalent.88
mo-IB
85
See also Ezek 46:4–6, 11, 13. Furthermore, since male sheep reach sexual maturity at the
age of approximately one year, there is no biological category of male sheep that are too young
to be termed LYiA x but are over a year old. Similar usage of descriptive qualification in the zoemic
system is found in the partitive genitive form OYZ diEi which may or may not follow RYE i$
om
(compare TAU m
dxXvL OYZd
iEi -RYE
iv$o of days 1–5 of the main autumn festival in Num 29:12–34 with
TAU m
dxX RYE i$ov of days 6–7); similarly, RQ md
m ie may or may not follow RP
B-IB d, with no apparent
x
change of meaning (compare days 1–2 of the same festival with days 3–7).
86
Contrast OY$ oi m K-YN
Bv a$vo with d
HT Nvo-TB
mm$ dx TX xAx H$om v K (Lev 14:10), where one of the OY$
Bd
x o m
Bdv
K
is offered as an HL m oWE and therefore must be male.
87 ie o$e
This cannot be proven from the cases where a WoTN vo-IB
m$ BdK is required for a wholeburnt
e
offering, since implying s웧 is not the same as denoting s웧, but according to Milgrom’s dis-
tinction (Leviticus 1.252) between o$e K (s웧) and B$
Bd
e K (), WoTN
oe d
e m$ ie o$e
vo-IB K probably does not
Bd
e
denote “s,” since one would expect “s” to be termed WoTN m$ i
vo-IB
e B$ K. Presumably, then, LXX on
oe d
e
Num 7:17, 23 etc., where o$e Bd
eK is translated consistently as s웨 (αµνά), is simply wrong. It
appears that the translators were overanxious to reconcile the lists in Numbers 7 with the law of
Leviticus 3, according to which female specimens are acceptable as wellbeing offerings; in fact,
they did not only attempt to reconcile these lists with the law in Leviticus 3, but tried to make a
point of the law in Leviticus 3 by translating OY$ oi m
Bd
vK as female lambs. The very same process is
traceable in their treatment of RQ md
mB in these verses, on which see n. 28.
88 ie denotes “yearling” in
There is some disagreement among scholars as to whether WoTN vo-IB
m$
the sense of a one-year-old animal, i.e., an animal in the second year of its life (Milgrom,
Leviticus 1.757), or “in its first year” (i.e., in the first year of its life, J.–M. 129j). The evidence in
favor of the first interpretation is the fact that the term IB in phrases such as ½IB DEW $DX½IBM
OYN$ $MX (Lev 27:6) clearly serves to designate a minimal age, not “within a month” or
“within five years.” On the other hand, IB unambiguously denotes “within” in the BH phrase
HLYL-IB (Jonah 4:10, as a stock phrase, “overnight”; for the vocalization IB i compare Josh 1:1;
d
Prov 30:1) and in rabbinic phrases such as WMWY-IB RW$ (see n. 000), which clearly denotes “in
the first day of its life.” While it is tempting to suggest that HN$½IB etc. (without a pronominal
suffix) denotes “yearling” whereas WTN$½IB denotes “within the first year of its life,” the two
phrases are evidently stylistic variants in Num 7:87 and 7:15, 21, etc. (thus Milgrom, Leviticus
1.757). Milgrom deduces that both therefore denote “yearling,” but the opposite conclusion is
more likely: the two phrases are in fact equivalent, but both denote “in the first year of its life.”
Since there is no unambiguous attestation of the phrase WTN$½IB (or WMWY-IB etc.) in BH or in
later Hebrew in the sense of a minimum age, while the phrase IB (without a pronominal suffix)
is ambiguous––attested already in BH both as a minimal age and in the sense of “within”––it
appears that in P, WTN$½IB denotes “within the first year of its life,” and, considering Num 7:15,
21, etc., the phrase HN$½YNB denotes the same age-group in this context. For this understanding
of the biblical law in Josephus and in rabbinic literature, see n. 000 and m. Parah 1:3,
respectively.
11:50:13:06:14 Page 55
Page 56
2.4.6 HB
d
m o$d
iK
Both HB
d
m$ iK and HB
ov d maQvN . . . o$e
Bd
eK (Lev 4:32; 5:6) denote a female sheep of any
age (웨).89 The use of two terms for a single category may be the result of a
diachronic development in BH, but this is difficult to prove.90
2.4.7 H$
om v
Bd
xK
H$
om vdK is found only twice in P, where it is qualified by d
Bx HT Nvo-TB
mm$ x (Lev 14:10;
d
Num 6:14). It denotes an immature female sheep, that is, a female lamb (s웨).
Note that in P there is a functional distinction between s웨 and 웨 offered as
a hatta’t: the former is offered where only purification is required (of a person
purified from scale disease, Lev 14, and of a Nazirite who has become impure,
Num 6); the latter is offered when both purification and forgiveness are
involved (Lev 4:32; 5:6).
89
Concerning HB d
m$ iK, the formulation of Lev 5:6 suggests that no specific age group
ov d
among 웨 is implied: if the formula IAC H-IM
do x i HB maQvN stands in apposition with -WoA HB d
m$ iK-
ov d
OYZ
diE i TRxYEi$ ov , as seems to be the case (see 1:10: OYZ d
iEiHm -IM
i oWA OYB i om
$d v
KxH-IMi . . . IAC H-IM
do x i ), then
HBd
m$ iK must mean 웨 and OYZ
ov d d
iE i TRxYE i$ov , 웨. See Rendtorff 1985:196. This tallies with the fact
that 웨 is known to be a zoeme elsewhere in P’s system (4:32, where o$e K must mean , since
Bd
e
it cannot mean s웧). However, one cannot be certain that H$ om v
Bd
xK alone may designate immature
specimens as o$e BdK does. Unlike o$e
e BdK, which appears sometimes without WoTN
e m$ ie but still
vo-IB
clearly implies immature specimens, H$ om v
Bd
xK appears only twice in P and is modified by
HT
d mm$Nvo½TB x . A less likely alternative is that HB
d dx$ iK is the precise mirror image of o$e
ov d Bd
eK (s웧),
i.e., s웨, and the relation between IAC do xH IMi HB maQvN and OYZ diEi TRxYE i$ov -WoA HB
d
m$ iK is not one of
ov d
simple apposition. This conclusion would accord with the rabbinic view (e.g., Maimonides
HMQ 1:14). However, in this case one might expect a different formulation, such as
OYZE TRYE$ AYBHW OYZEH IM OAW . . . HB$K AYBHW OYB$KH IM OA . . . IACH IM HBQN.
Note that P has a different word for s웨, namely d HT Nov -TB
mm$ d
x H$ om v
Bd
xK, and that even the Sifra does
not simply assume that HB QvN . . . o$e
ma Bd
eK is one year old, but takes care to deduce this by analogy
to HTN$½TB ZE in Num 15: HTN$ TB HAB WZ FA HTN$ TB HAB HRZ HDWBE TAUX HM
(Weiss 22a).
90
There are reasons to suspect that Lev 4:32–5 is an interpolation, not belonging to the
original corpus of P. (1) The possibility of offering an 웨 appears to be an afterthought.
Elsewhere in P, where two possible types of animals are offerable (not as substitutions as
in WDY GY$T AL OA, “if his means do not suffice,” but as two equal alternatives), the text
first enumerates the general rule and then enumerates the two possibilities (see Exod 12:5 H$
OYZEH IMW OY$BKH IM . . . HN$½IB RKZ OYMT; Lev 5:6 TRYE$ WA HB$K . . . IACH IM HBQN
OYZE; and even Lev 1:2 IACH IMW RQBH IM . . . HMHBH IM, 1:10 WA OYB$KH IM . . . IACH IM
OYZEH IM). Lev 4:32–5, by contrast, does not follow this pattern, and the possibility of offering
an ovine comes as a surprise to the reader who does not expect this alternative possibility from
reading v. 28 (and having also read that the chieftain may offer only a male caprine). (2) The use
of o$e BdeK (sgl.) to denote in this passage (alongside P’s standard B$ oe d
eK in v. 35) is unparalleled
in P. (3) Elsewhere in P, ovines normally precede caprines within a single list (see particularly
5:6), whereas here the order is reversed.
11:50:13:06:14 Page 56
Page 57
Zoemics 57
2.4.8 RYE
i$om
Surveying the ancient traditions regarding the precise age of the RYE om is
i$
confusing, and the innovative solutions of medieval rabbinic sources reflect
the exasperation of the commentators.91 Though it is certain that in BH
RYE
i$ om (= OYZd
iEi RYEi$ov ) designates male caprines,92 it is difficult to determine
whether it denotes 웧, G웧, or g웧, or perhaps a fourth category, borrowed
directly from its technical usage in animal husbandry, whereby it may have
designated, for example, a male goat “one winter old” (as LXX understood,
χµαρο).93
In Num 7:16–17, 21–2, etc., RYE i$om appears in a single list along with
i TE i TE
OYDiW d x. Since OYD iW d x are mature male goats (G웧), it is almost certain
that OYZd
iEi RYEiv$o does not also denote G웧.94 The most likely interpretation of
the biblical term––an interpretation that may also settle the controversy
without rejecting any of the ancient traditions––is that the term RYE i$om within
the ritual texts does not denote an age group at all. The phrase OYZ d
iEi RYEi ov
$
simply denotes 웧, and HRmYE i$ov denotes 웨. The fact that OYZ d
iEi TRxYEi$ov ,
together with HB d
m$ iK (웨, above, §2.4.6), is in apposition95 with IM
ov d i HBma QvN
IAC
do xH (Lev 5:6) lends some support to this interpretation, and the fact that
i TE
OYZ
d
iEi RYE i$ov appears alongside OYD iW d x in a single list (Num 7) does not
disprove it (see §2.4.2, concerning RP d and RQ
x md
mB in a single list).
2.4.9 RP
d, LG
x a, and WoTN
eE m$ ie * LG
vo-IB eE
a
The term LG a attested in a sacrificial context in P only three times, in

eE
Leviticus 9 (v. 2: RQ
md
m ie LG
B-IB a; v. 3: HN
eE m$mo-YN
aBd $oB
v eeW
Km LGa; v. 8: LG
eE a). The
eE
language of this chapter appears to be uncharacteristic in many ways,
91
See for example n. 000.
92
The fact that in one case, RK m modifies RYE
mZ i$ om (Lev 4:23), but never RP
d or LYiA
x x , which are
clearly masculine, does not necessitate that a RYEi$om may also be feminine (and that hence RYE i$om
= or G). The appellation RK m in Lev 4:23 probably serves only a literary purpose, contrasting
mZ
OYMT RKZ OYZE RYE$ of the chieftain with HBQN HMYMT OYZE TRYE$ of laypersons. Note that
P employs a different term for , namely ZE a (e.g., Lev 3:12).
93 i TE
That is, too old to be termed YDiGv (unattested in P), but too young to be termed DW
d dx
(attested only in one context in P, in Num 7, but common in non-Priestly sacrificial ter-
minology, e.g., Isa 1:11). Cf. LXX on Ezekiel, which employs ριφον α+γν, “kid” (e.g., 45:23).
In Dan 8:21 the term RYE $ appears to be a gloss on RYPC; from other lists it appears that RYPC,
i om
both in Aramaic (Ezra 6:17) and in Hebrew (2 Chr 29:21), is identical with the RYE om .
i$
94
Contra Marx 2005:60.
95
See n. 000.
11:50:13:06:14 Page 57
Page 58
including the use of several terms that are found nowhere else in P.96 There-
fore, the precise denotation of LGeEa in P is especially difficult to ascertain.
It has been suggested that LGeEa in this verse is synonymous with RP d, that is,
x
a mature male bovine (B웧);97 modern translators commonly render this term
“young bull.”98
A more likely possibility, in light of the use of the term HL Ge in BH to
mvE
designate female bovines of any age,99 is that LG eE a is P’s coinage for male cows
of any age (웧). This last interpretation, which I propose with some caution,
would accord with previous suggestions according to which LG a can be
eE
sexually mature, and it would not contradict the fact that LG eEa can be modi-
fied by HN m$ ie (Lev 9:3; Mic 6:6). If this interpretation is correct, then the
mo-IB
three bovine zoemes referred to in Leviticus 9 are 웧 (RQ B-IB
md
m ie LGa, v. 2
eE
and LG a, v. 8), b웧 (HN
eE i
mo-IB
m$ e LG a, v. 3), and (RWo$o, vv. 4, 18)––all of which
eE
are distinct from the zoeme B웧 (RP xd) that is absent from this chapter but
common elsewhere in P.
If, however, the terms RP d, LG
x a, and WoTN
eE m$ ie LG
vo-IB a represent three distinct
eE
age groups, then the systematic binary structure here falters, and there is an
irregularity at this point within the system.
2.4.10 ZE
a and d
HT Nvo-TB
mm$ d
x ZE
a
Finally, there is clearly a zoemic distinction between ZE

a () and d
HT Nvo-TB
mm$ d
x ZE
a
100
(g). The law in Num 15:27–8 requires an inadvertent sinner to offer an
HT
d Nvo-TB
mm$ d
x ZEa as a purification offering. Despite scholarly disagreement
96
On the peculiarity of Lev 9, see Koch 1959:69–70 (not all of the evidence adduced by Koch
is equally convincing). Moreover, the term HSKMH (v. 19) in the sense of TA HSKMH BLXH
BRQH is a hapax legomenon (in contrast to seven occurrences of BRQH TA HSKMH BLXH). The
verb XBZ is found nowhere else in P (v. 4; it is found in H, 17:5,7; 19:5; 22:29, see Milgrom,
Leviticus 1.573); this chapter uniquely employs DBKH IM TRTYH (9:10) alongside DBKH TRTY
(v. 19). Elsewhere it is termed DBKH TRTY or DBKH LE TRTYH, never DBKH IM TRTYH
(10 attestations). Verse 14 is also unique in that P nowhere else has $YGH precede hand-leaning
(but see 2 Chr 29:23). The formula HNMM WPK ALMYW (9:17) is also unique, as P normally uses
JMQ. Finally, the zoemes LG a and WoTN
eE m$ ie LG
vo*-IB a (vv. 2–3) are mentioned nowhere else in P, and
eE
the zoeme RWo$o (4, 18) is not required anywhere else in P.
97
This was already the view of Ibn Ezra on Lev 9:1 (see Shachter 1986:34). Ibn-Ezra notes
that this would render the animals of the eighth day (Lev 9) somewhat similar to those of the
seven days of ordination (HLWEL LYA OG TAUX RP, referring to Lev 8:1, 18).
98
For instance, Marx 2005:59.
99
As at Isa 7:21, where it is clear that the HL Ge produces milk (immature calves do not
mvE
produce milk). See also Gen 15:9, which refers to T$L$M HLGE. If T$L$M denotes “three
years old” or “in its third year,” then HL Ge must include mature specimens. Note, moreover,
mvE
that the constraint dHdm
B DB
d
xuE-ALo R$ A in Deut 21:3 implies that there could be an HL
oe q Ge that had
mvE
been used for ploughing. This, too, suggests that HL Ge is not limited to immature bovines.
mvE
100
Note that whereas there is a distinction between ZE a and d HT Nvo-TB
mm$ d a, the terms o$e
x ZE Bd
e
K
and WoTNm$ ie o$e
vo-IB Bd
eK often denote the same zoeme (both may denote s웧; see above). This is one
example of asymmetry in the zoemic system of classification.
11:50:13:06:14 Page 58
Page 59
Zoemics 59
concerning the relation between this law and the law in Lev 4:27–35,101
modern scholars and ancient commentators and translators unanimously
agree that the animal in question is a young female goat.102 This interpret-
ation, though possible, is unlikely.
As in other languages, the gender of nouns denoting animals in BH does
not necessarily correspond to the sex of the specimen under discussion.103 For
example, the noun RT od , probably “turtledove,” is grammatically feminine:
it is invariably qualified by a feminine adjective (OYRiT o YT
a
dv$o, technically a
construct chain), though the specimens referred to may be male or female.104
Conversely, the phrase HN mWoY-YN
aBd is always masculine, though the specimens in
v
question may be of either sex. To the best of my knowledge, no one has ever
doubted these two cases. It is possible that the fact that males and females
of these species are difficult to distinguish externally contributed to the lack
of ritual differentiation between male and female pigeons and turtledoves
in P, and subsequently to this (justified) consensus.
By contrast, the case of dHT Nvo-TB
mm$ d a ––though it may be explained with the
x ZE
same grammatical feature just introduced––has invariably been taken to
denote a female goat. This conclusion, however, is unlikely. The noun ZE a in
BH is used in two distinct senses:105 (a) as a nomen generis (genus capra, i.e.,
, as in e.g. Lev 7:23; 17:3; 22:27);106 and (b) as a narrower term designating
the mature female of this species (G웨, only in non-Priestly texts like Gen
31:38; 32:15).107 Denotation (a), by far the more common, is well established
and can be considered uncontested.108 Denotation (b) is based on a small
number of texts in which this sense is also uncontested.109 The data are found
101
See for example Kellerman 1973; Knohl 1990.
102
For example, LXX α,γα µαν νιαυσαν (acc.f.); Vg capram anniculam; since the Aramaic
AZYE and its modifiers are in any case feminine (as in BH), one cannot tell what the sex of the
animal is from the phrase HT$ TB AZYE (e.g., in Targum Onqelos). Rabbinic sources clearly
imply that it is female, on which see below. See also NEB, RSV, JPS; and Levine Numbers, 1.397,
who even writes, “Why a female animal was specifically required is not clear.”
103
See J.–M. §134c.
104
Note that its plural form happens to be –îm, as in a number of other feminine nouns
(e.g., OYN
iB
mqA, IB
eAe ).
105
A third sense, “goat-hair” (only in plural, OYZ
d
iEi ), is not discussed here.
106
Some errors are still found in the literature regarding this; see Dorival 1994:127.
107
Such duality is found elsewhere in BH animal nomenclature, though it is often the term
designating the mature males that is also used to designate the species as a whole (e.g., RWoMXq in
Gen 32:6 et passim), and not the term for the adult female, as here.
108
See Zobel 1999:578.
109
Primarily Gen 32:15, where the pair OY$YT || OYZE is juxtaposed with OYL a || OYL
i YA iaXRv
(adult female and male caprines, followed by adult female and male ovines); but also Gen 31:38
(||LX
a Rm, where the issue is miscarriage, hence an adult female is necessarily implied); in Gen
30:35 it is clear that the animals referred to are mature females due to the parallel with mature
males (OYZE || OY$YT). It should be remembered that although Jacob’s flock in theory may have
consisted of young goats as well, in the context of the plot the author has the mature animals in
mind here, particularly those which will soon (v. 38) be in heat and then conceive and bear
offspring. Two other texts that may be further evidence of sense (a) are questionable: OYZE BLX
11:50:13:06:14 Page 59
Page 60
in Péter-Contesse’s study of the subject and need not be presented here in

full.110
The question to be addressed here is which of these two senses is implied at
Num 15:27. If sense (b) (a mature female goat) were intended, then the zoeme
would consist of non-existent animals, since the qualification d HT Nvo-TB
mm$ d
x
(i.e., in the first year of its life)111 virtually excludes the possibility that the
animal could be a mature female goat.112 It appears, then, that the author has
sense (a) in mind, and ZE a refers to goats in general. Some commentators,
admitting that sense (a) is implied, nevertheless assume that the feminine
adjective d HT Nvo-TB
mm$ x indicates that the goat implied here must be a she-
d
goat.113 However, this assumption is probably mistaken on grammatical
grounds: it is likely that the noun ZEa designating the species is modified with a
feminine adjective simply because the noun is grammatically feminine in
BH, as in other Semitic languages, even when the species is implied.114
As discussed above, P’s OYRiWoT YT a
dv$o serves as an excellent parallel, as the
feminine YT a
d$vo modifies the (grammatically) feminine OYRiWoT d , even though
the specimens referred to may be either sex.115 Hence, while d HT Nvo-TB
mm$ d
x ZEa
in Prov 27:27 may refer to the nomen generis, although only the females give milk. The term
T$L$M ZE (Gen 15:9), which apparently refers to a mature specimen (T$L$M, three years
old), could be considered evidence of sense (b), but see below.
110
See Péter-Contesse 1992, who nevertheless misidentified the zoeme d HTmm$
Nvo½TBd a in
x ZE
Num 15.
111
See §2.4.5.
112
I write “virtually excludes” because goats may mature before the first year is over, after
approximately eight months. Archaeozoologists assume that goats in the first millennium 
in Palestine matured at roughly the same age as goats in modern times (if the latter are
not treated with hormones to speed maturation). I thank Dr Rivka Rabinovich of the
Hebrew University Archaeozoology Laboratory for this piece of data. See also m. Bek. 3:1, b. Bek.
19b–20a. However, it is highly unlikely that the law would require that a person bring such an
animal: this would mean that if a person sinned at the beginning of the summer, when most
kids are born, that person would have to wait eight months before offering a TAU m X, since last
dx
year’s kids would be too old already and this year’s kids still immature. (Note that the form
HTN$ TB ZE in m. Bek. 3:1 does not necessarily indicate a shift in the usage of ZE a in MH.)
113
See, for instance, Péter-Contesse 1992:72.
114
See for example CAD 180b–3a, and in particular 182a, denotation 2.a1; Sokoloff 852b
with examples; see also Militarev and Kogan, SED II.53–6, though their analysis of the Biblical
Hebrew data is incorrect as argued here. In some other Semitic languages, however, the cognates
of ZE are masculine, e.g., del Olmo Lete and Sanmartín, DUL 196 s.v. ‘z. The noun ZE a in sense
(b) is always grammatically feminine in BH. Despite ample attestation of ZE a in sense (a) in BH,
it is unfortunately never found in a grammatical context that would reveal its grammatical
gender. Theoretically at least, in terms of its gender, ZE
a in denotation (a) could be either epicene
(and feminine, like TBNRA) or common (and thus either feminine or masculine depending on
the sex of the animal implied, like LMG). See G.–K. §122b–e.
115
The turtledoves referred to in P (as well as the pigeons, termed HN mWoY-YN
aBd ) can be either
v
male or female. This is not only logical (the two sexes are hardly distinguishable) but also in
accord with rabbinic tradition and, more importantly, with the precise language of P, who
throughout Lev 1 distinguishes between male and female quadrupeds but makes no such
distinction among birds.
11:50:13:06:14 Page 60
Page 61
Zoemics 61
could denote a female kid, it most probably denotes “a goat in the first year
(of its life)” regardless of sex; that is, any kid in the first year of its life, male or
female, would qualify.116
Having finished examining the philologically contested cases, we are now
able to make some final observations regarding meaning at the zoemic level.
2.5 ZOEMICS AND MEANING
While reserving a general discussion of the problem of meaning in ritual for

Chapter 6, I would now like to consider where, if at all, meaning can be
sought within zoemic systems.
It is possible that the choice of the zoemic inventory as a whole reflects
some profound truth pertaining to the primordial meaning of sacrifice. For
example, there may be some truth in the theory that the choice of cattle to the
exclusion of feral or wild animals and the centrality of cultivated crops and
domestic fire are key to the primary meaning of sacrifice––a stylized “game”
celebrating humanity’s successive domestication of fire, animals, and finally
plants over the course of prehistoric time.117
Applying this theory to Σ may underscore the exceptional character of the
d , which––according to the independent evidence of Philo and rabbinic
RWoT
sources118 ––was hunted for sacrificial purposes in the late Second Temple
period, and therefore clearly is an example of a non-domestic sacrificial
116
One might counter by suggesting that the noun ZE a has a third denotation, (c) “female
goat” (of any age)––but this counterargument must be rejected on the following grounds.
(1) This denotation is nowhere attested in BH. (2) The Priestly texts employ a different term for
“female goat” (namely OYZ diE
i TRxYEi$ov ). This term is used throughout P in counterdistinction to
OYZ
d
iEi RYEi$ov , a male goat (of any age). In fact, the term OYZ diE
i RYEi$ov is found in this very passage
(in v. 24). It stands to reason that if the authors had wished to designate “female kid,” they
would have used d HT Nvo-TB
mm$ d
x OYZ
d
iEi TRxYE $. Therefore, even if this third denotation existed, the
i ov
term ZEa in the formula d HT Nvo-TB
mm$ d
x ZE a should still be construed according to denotation (a).
Alternately, one might counter the argument presented here by suggesting that the noun ZE a
in denotation (a) is common rather than epicene, so that denotation (c) is merely a particular
case of (a), and that the modifying adjective d HT Nvo-TB
mm$ x reveals that the author had a female
d
specimen in mind. Once again, this counterargument would be invalidated by consideration
(2) above.
117
See J.Z. Smith 1987; concerning sacrificial rites as a reenactment of the “domestication” of
fire, compare Heesterman 1993:120.
118
See Philo DSL 1.162, and the rabbinic story about King Agrippa and the poor man (Vayik.
Rab. 3:5): YNAW OWY LKB DC YNA HEBRA LWDG IHK YNWDA . . . OYRWT YT$ WDYBW DXA YNE AB
OYN$M SNRPTMW OYN$ BYRQM (“A poor man came, holding two turtledoves. . . ‘Melord High
Priest, I hunt four every day; I offer two and subsist by means of [the other] two’ ”). Concerning
the domestication of the species indicated by the BH RT o , see also Milgrom, Leviticus 1.168. This
d
argument is no less valid if one accepts Staubli’s conjecture that RT o in P is a bird of the genus
d
Gallus (Staubli 2008).
11:50:13:06:14 Page 61
Page 62
animal.119 This type of explanation, however, does not call for a grammatical
model.
The same is true of other “essentialist” or reductionist theories of sacrifice,
such as the sociological approach of René Girard, the evolutionalist approach
of Walter Burkert, and the cultural-materialist approach of Marvin Harris.120
Some or all of these theories may be true and insightful when applied to the
general scheme of sacrifice: they may explain (respectively, and somewhat
simplistically) why members of a society pick on other, weaker members,
including their own cattle; why they kill these animals without eating them;
and why they spare their donkeys. However, these theories do not purport to
account for the minute details of sacrifice, those very minutiae with which the
authors of the Priestly texts are seemingly so obsessed: they cannot explain
the self-perpetuating binary pattern underlying the Israelite system or the
internalization of an unstated rule concerning the sex of an animal offered as
a sin-offering. More importantly, these theories do not address a question
that lies at the very heart of every ritual system––“(what) do these details
signify?”
The question that is raised here, and discussed at length in the chapter
on , is not whether sacrificial systems mean anything. Rather, what
is at stake here is whether rituals have a meaning corresponding to and
dependent upon a delicate yet rigorously rule-governed interplay between
their constituent elements––the zoemes, defined in this chapter––and other,
higher- and lower-level elements defined in the following chapters.
119
Applying this theory to Σ may also offer new insights into the myth of the divine origin
of the “perpetual fire” on the bronze altar. Contrast Lev 1:7; 6:5 with 9:24.
120
Girard 1977, 1987; Burkert 1987; Harris 1974.
11:50:13:06:14 Page 62
Page 63
Jugation
3.1 INTRODUCTION
In the sacrificial system reflected in the Pentateuch, a zoeme is rarely to be

offered in isolation. In the first place, many sacrificial rituals entail offering
more than one zoeme, creating higher-order ritual complexes that will be
discussed in  (Chapter 4). Additionally, offering a zoeme often
entails accompanying non-animal materials, such as cereal gifts and libations;
such supplementary offerings are the subject of jugation (from Latin jugare,
“to join”). The tiered structure of ritual has been explored in depth in
previous discussions of the “grammar” of ritual;1 nevertheless, a systematic
description of the rules governing such structures is lacking in the literature.
Moreover, the systematic distinction introduced here between two types of
tiering––discussed separately in the chapters  and ––is
novel, and is essential for a proper understanding of Σ.
As mentioned, many sacrifices, especially animal sacrifices, often entail
ancillary offerings, which we shall call subordinate jugates. In order to discuss
the relationships between primary and supplemental sacrificial materials,
we use the following convention: a primary sacrifice receives the subscript
letter “A,” each jugate that is immediately ancillary to it receives the letter “B,”
and so on. We will also use the term “co-jugates” to refer to two or more
jugates at the same level that are not jugated to different elements. The
definition is a subtle but useful one: two or more B-level offerings jugated to
the same A-level offering are co-jugates, for example, but so are two or more
A-level offerings that are offered together under specified conditions.
Judging from P and from post-biblical material, the most common type of
jugation in Σ is the adjoining of a HX Ni (often translated as “cereal offering,”
mvM
“grain offering,” or “meal offering”)2 and a vVe N (“libation”) to a zoeme
Se
1
For discussions of the tiered structures of particular rituals within the framework of
a grammatical approach, see Staal 1979, 1980; Lawson and McCauley 1990; and Gane 2005. For
a different perspective, see Klingbeil 1998.
2
For the etymology of the term and its various denotations, see Snaith 1957; Levine 1974:7
n. 9, 17 n. 38; Eberhart 2002:78–9; Anderson 1987:30; and Marx 1994:1–28.
11:50:13:06:14 Page 63
Page 64
offered on the altar. This combination appears to be a legacy from Levantine

sacrificial practice and is couched in human realia: like the offerers them-
selves, the gods are assumed to prefer a starch and an alcoholic beverage along
with their meat.3 Similar condimentary practices may explain, to some extent,
the requirement to add salt (XL xMe ) to portions of certain offerings.4 As we
shall see, the elementary jugational structures in P are governed by a small set
of specific constraints, some of which are quite simple and intuitive, others
more nuanced.
In considering the nature of such constraints, an analogy to fast-food
menus suggests itself. In addition to orders of a lone burger, a lone soda,
and lone fries, one may choose from combinations of food that together
constitute a meal. A burger with fries (a starch prepared in oil and often
salted) and a soda is common, and a burger with fries or a burger with a soda
are also not unheard of––hence the typical adage, “Do you want fries with
that?” However, a combination of only fries and a soda is not licit: that is,
while it is not illegal to order fries and a drink separately, they do not con-
stitute a meal. This is reflected in the absence of such an option from most
“combo” menus. So too, note that the subordination of fries and sodas to
burgers is one-way: ordering a drink does not typically evoke “Do you want a
burger with that?” Such is the nature of jugational subordination in Σ.5
One aspect of the jugational category of Σ that will soon become apparent
is that it does not admit of the same straightforward binary model as zoemics.
Furthermore, a number of uncertainties exist at the jugational level, primarily
owing to ambiguities in the Priestly texts themselves; we will discuss these as
they arise in examples.
Before proceeding, two final points of clarification are in order. The first is
that we intend this chapter, like , to act as an introduction to the
 booklet; for this reason, the discussions here are meant to cover
many representative examples rather than to exhaustively provide the same
rules that are set forth in the . The second is that jugation is not
intended to describe every tiered relationship; when several animals together
make up a festival offering, for instance, this relationship is hierarchical and
not jugational, and will be treated in the next chapter.
3
This triadic structure is not an innovation of P. See Wenham 1979:128; Urie 1949, who
reconstructs vVeSeN+HX Ni +OYM
mvM im $o as the basic West Semitic sacrificial scheme; also see, for
Lv
example, 1 Sam 1:24; 10:3, and Fleming 2000:233–93. Generally speaking, the cereal and libation
offerings in the Israelite system were significantly less diverse than those of the Hittite and the
Mesopotamian cultures––though the data are not easily comparable. By way of example, a
single Hittite ritual text (CTH 476) mentions no less than ten types of bread––a greater variety
than is found in the entire inventory of grain offerings in P (on which see below). P mentions
only the libation of wine (and perhaps of another type of libation, ale [RK
m$ao], on which see n. 000).
4
On salt, see §3.9.2.
5
On the combination of grain (in the form of bread) and wine without meat in “sacred
meals” that are not strictly speaking sacrificial (and thus are not part of Σ) but that may be
modeled after certain rituals associated with the Jerusalem Temple, see Marx 1994:163–5.
11:50:13:06:14 Page 64
Page 65
Jugation 65
3.2 STANDARD EXAMPLES
3.2.1 Subordination to Zoemes: Numbers 15
The law in Num 15:2b–16 states that every wholeburnt or wellbeing offering,
whether votive (RDeNa), volitional (HB Dv), or calendric, must be accompanied
mmN
by (wheat) semolina (TL eSo )6 with oil (IM
ee$o),7 and a libation of wine (IYiYx), as
shown in Figure 5.
According to the law in Numbers 15, the precise quantities of wine, semo-
lina, and oil required with certain animal sacrifices vary from one zoeme to
the next, according to the scheme shown in Table 2.8 This table will prove
Fig. 5. Basic jugational pattern according to Numbers 15.
Table 2. Table of jugation according to Numbers 15
Semolina Wine Oil
B웧 3/10 ephah 1/2 hin 1/2 hin

S웧 2/10 ephah 1/3 hin 1/3 hin
s웧 9 1/10 ephah 1/4 hin 1/4 hin
6
Where the offering is barley (as at Num 5:15), P states so explicitly. See n. 000.
7
As Milgrom (1990:239) notes, there are, strictly speaking, two types of oil: the oil for the
tāmîd and the candelabra is beaten or ‘“pressed in a mortar,” hence “pure oil,” and oil for all
other occasions is “unbeaten.” Concerning the status of oil in relation to semolina, see the
discussion on p. 000.
8
Note that the quantities of wine and oil are identical, though in the present scheme they are
not, strictly speaking, co-jugates: wine is a B-level jugate, whereas the oil is a C-level jugate. The
fixed relation between the ingredients can be expressed by a simple algorithm: 10y = 5 − 1/x,
where x = oil (equivalently, wine) and y = semolina. This is clearly not how these quantities
came into being, nor how they were conceived by practitioners. Rather, they presumably repre-
sent values of standard measuring cups (different for solid and liquid materials), with an
attempt to create a mixture of a more-or-less fixed consistency. Concerning Lev 23:13, see
Milgrom, Leviticus 2.1988.
9
On this zoeme, see §3.7.1.
11:50:13:06:14 Page 65
Page 66
useful in the discussion of some of the more complex patterns discussed in

what follows. At present, it should suffice to note that there is a very close
correlation between the table laid out in Numbers 15 and the lists (and
precise quantities) of subordinate jugates in the calendric public offerings
listed in Numbers 28–9.10 Outside these two texts, however, several Priestly
sacrificial texts do not reflect these proportions, and wineB is entirely absent
from a large number of texts.11
3.2.2 Independent Cereal Offerings: Leviticus 2
P also mentions jugations that pertain to entirely non-animal offerings. For

example, Leviticus 2 describes several voluntary offerings of semolina that do
not accompany animal sacrifices. In the case of raw semolina, frankincense
(HN v ) is required as well. Thus, the voluntary A-level (i.e., standalone)
mWoBL
offerings of raw semolina (Lev 2:1–3) are of the form shown in Figure 6 (on
the place of “oil” in this scheme, see immediately below).12 Scholars disagree
Fig. 6. Jugational pattern of semolina offering in Leviticus 2:1–3.
10
Several scholars (Noth, Nihan, Achenbach) consider these two texts very late strata in the
Priestly literature, or even post-Priestly. See n. 000.
11
See Seebass, Numeri 2.139; Rendtorff 1967:15–16, 169–73. There is a degree of circularity
in Rendtorff’s claim that passages outside Num 15 and 28–9 ignore the “table of jugations,”
since this is supported by the assumption that even where wineB is mentioned, it is an inter-
polation. However, independent evidence is occasionally found for the secondary nature of such
passages (e.g., Gray 1903:67–8 on Num 6:15b; Rendtorff 1967:15 and Milgrom 1990:239 on
Num 28:7). Note that translating DX mAeH m as “the first” (v. 7, JPS) rather than “per each” renders
the syntactic problem less pronounced but does not solve it.
12
In contrast to the stipulations for standalone HX mvMNi offerings, P does not require the
jugation of frankincense to a cereal offering that is itself jugated to a zoeme offered as a
calendric offering. In fact, there is no clear evidence for the jugation of frankincense to an
animal sacrifice in P’s sacrificial system. Haran asserts that such jugation “certainly” occurred in
ancient Israel (Haran 1978:230), and while he does not cite his sources, several late Second
Temple texts suggest precisely this. See ADL 8:6, where the frankincense is physically added
on top of the flesh–cereal–wine complex. See also Gen. Apoc. 10:15–16; ADL 9:15–16, the
discussion in Schiffman 2005:197–8, and, concerning Jubilees, Himmelfarb 2004:116–21.
Rabbinic tradition in this case adhered to the letter of the text and, like P, exempted frankincense
from subordinate cereal offerings (m. Men. 5:3).
11:50:13:06:14 Page 66
Page 67
Jugation 67
as to whether the laws in Leviticus 2 pertain exclusively to standalone freewill

grain offerings, or also to grain offerings that are themselves subordinate to
zoemes,13 a controversy that also finds precedent in antiquity.14 We will return
to this observation below, §3.8.
3.2.3 Oil: A Slippery Subordinate
Some clarification is necessary concerning the relationship between oil and

cereal offerings. The term HXmvMNi (when used in the specific sense of “cereal
offering” or “grain offering”) often appears to denote only the grain com-
ponent, most often semolina, while the oil merely serves as an “additive,”
subordinate to the semolina (Lev 2:1; 7:10; 9:4).15 Elsewhere in P, however,
the term HX Ni denotes the grain offering in its totality, including the oil and
mvM
the semolina (and perhaps other components as well, as at Lev 2:2; 6:8).16 We
are therefore faced with a degree of ambivalence in the Priestly texts: some-
times oil seems to be a subordinate jugate of semolina, but other passages
suggest that semolina and oil contribute equally to the cereal offering.17
13
E.g., Knobel 1861:76; Seebass, Numeri 2.139 holds a middle position: the laws in Lev 2
would pertain to subordinate grain offerings as well, to the exclusion of the requirement of
frankincense. Wenham 1979:128 assumes that a handful of each cereal offering was burnt (even
in the case of B-level grain offerings), the rest given to the priest. There is no evidence that
1 Cor 9:13, which Wenham cites in support of this argument, refers to subordinate grain
offerings.
14
For example, Jubilees, Josephus, the Temple Scroll and Tannaitic tradition seem to agree
that 2:2b–3 pertains to A-level offerings only, whereas B-level grain offerings are turned to
smoke in their entirety upon the altar. However, the Sadducean halacha referred to in the
scholion to Meg Ta’an suggests that some authorities in antiquity viewed this law as pertaining
to B-level grain offerings as well. See Regev 1996, and cf. Noam 2003:250–4; concerning MMT
see Qimron and Strugnell 1994:150–2.
15
In Lev 2:1, HX Ni refers primarily to the semolina, to which the oil and frankincense are
mvM
subordinate (see also vv. 14–15, which deal with a special type of calendric grain offering). Lev
7:10 and 9:4 (IM ee$
dox
B HL miWLBv HX Ni ) actually imply that the grain component is the HX
mvM Ni , to
mvM
which oil is added. See also 14:10; Num 28:9, 12, 13 (IM ee$
dox
B HLmiWLB
d HX
v mvMNi TLeSo ). The common
translation of HX mvMNi as “cereal offering” (such as Milgrom Leviticus, 1.177) implies this inter-
pretation. Ezekiel, at any rate, is unambiguous: the oil is a subordinate adjunct of the semolina
(46:14).
16
See also Exod 29:41 and Lev 2:5, 8–9. Lev 6:8 is unequivocal: the phrase HX mvM
NdixH TLedS
oMi
HmM
d Nv $ iiW (“from the semolina of the HX
xdoM Ni and from its oil”) suggests that the oil is an integral
mvM
part of the HX Ni , whereas HX
mvM mvM
Ndix
H LE x R$eoAq HN
mBov
d
LxH (“the frankincense that is upon the
HX Ni ”) suggests that the frankincense is not subsumed under the term HX
mvM Ni . This accords
mvM
with the physical distinction between the oil and the frankincense: the former is mixed in with
the semolina, whereas the latter is placed on top. See, however, Lev 2:2, 16 (d HT No vL-LK
mmB m LE
d x,
“including all its frankincense”).
17
At times, both schemes are found in a single passage (e.g., Num 28:9, 12, 13, 20, etc.),
suggesting stylistic variation. A similar type of variation is found at Num 15:4, 6, 9 (where LW i LB
d
m
probably modifies IWoR$ od iE, and, if MT is correct, TL
m eSo should perhaps be taken as common;
note that the phrase HL miWLBd in 15:6 could modify either HX
v Ni or TL
mvM o , contra Seebass, Numeri
eS
2.132–3).
11:50:13:06:14 Page 67
Page 68
In discussing further examples, I will conventionally classify oil as a

subordinate jugate of the semolina, since even those passages that imply that
the semolina and oil together constitute the HX Ni often imply that the
mvM
oil component is subsidiary to the semolina, since the phrase is consistently
IM
ee$dox miWLB
B HL d TL
v eSo (“semolina mixed with oil”), never TL
edS
od
x i LB
B LW d
m IMeo (“oil
e$
mixed with semolina”). Note, too, that while it is possible to find a HX mvM
Ni
consisting of a grain component without oil (as at Num 5:15), an offering of
oil alone is not considered a HX Ni in P.
mvM
3.3 IDENTIFYING SUBORDINATION IN GENERAL:

SOME LINGUISTIC CONSIDERATIONS
In the above sacrificial complexes involving zoemes, the jugational relation-

ships between animal and non-animal sacrificial materials within each
sacrificial complex are quite clear: they are, as indicated, of the form {zoemeA,
cerealB, libationB}. The formulation of the biblical laws clearly reveals that the
non-animal materials are offered as “adjuncts”18 to (that is, subordinate
jugates of) the animal materials.
This subordination is often expressed linguistically: a possessive pro-
nominal suffix, referring to the zoeme, is suffixed to HX Ni or v
mvM Ve
SeN, creating
the common forms WoTX Ni , d
mvM HTmm Ni , OT
XvM mm Ni , WodKv
XvM SiN, d
Hdm
KvSiN, H
m YK
eSmvN, and
OHed
a SiN/OH
Kv e YK
av
d SiN.19 The pronominal suffix refers either to the bare zoeme
(that is, to a single animal) or to the entire sacrificial complex (to several
zoemes), as in the following two examples:
WodKv
SiW
Nv oWTX NiiW . . . HL
mvM m EoL
v DX
mAe RQ
md
m ie RP
B-IB d (Num 15:24)
x
“one bull for a wholeburnt offering. . . and its grain offering, and its libation”
HT
d mm NiiW $oD
XvM eX
ox x Eo (Num 29:6)
H TL
“the wholeburnt offering of the New Moon, and its grain offering”
18
Milgrom’s term, Leviticus 1.200.
19
Rendtorff (1967:15–16) justifiably expresses his perplexity over the fact that the injunc-
tions pertaining to the libation are so irregular. Note that there exists a slight asymmetry
between HX Ni and v
mvM VeSeN: a single zoeme always has one HX Ni and one v
mvM VeSeN, expressed by
Hd
d m
KvSiN/WodKv
SiN or dHT mm Ni /WoTX
XvM Ni ; when several zoemes are referred to in conjunction (for
mvM
instance, when they form a sacrificial complex of a higher level), the terms used are invariably
OT mm Ni and OH
XvM e YK
av
d SiN, not OHe YT
a WoXN
vMi and OK
mv
d SiN. This is simply due to the fact that whereas vVeSeN
(though it can be used in the singular, sometimes collectively) has a plural form, OYK imSvN, HX
mvM
Ni
usually has none (contra Milgrom, Leviticus 2.2006, who seeks to find some logic in this).
However, there are a few exceptions, such as Num 29:39 and Amos 5:22. Ps 20:4 is another
possible exception, according to the vocalization of the Masoretes. The same is true of the
masoretic vocalization of IWoHT vm Ni (“their cereal offerings [pl.]”) in Ezra 7:17 (rather than
XvM
IWoHTvxXvM Ni (“their cereal offering [sgl.]”)).
11:50:13:06:14 Page 68
Page 69
Jugation 69
In the former passage, the pronominal suffixes (3rd m sg) refer to the zoeme
RPd; in the latter, the pronominal suffix (3rd f sg) refers to the sacrificial
x
complex $oDeXoxH TLx Eo, which happens to consist of ten zoemes (2B웧, S웧, and
7s웧, according to Num 28:11).20
Subordinate jugation may also be expressed by the preposition -Lv , as in the
law pertaining to the daily regular offering:
i
DX
mAem
H o$e
Bd
eKxL IYiY
m IYH
ix i iBRvv
H TYE VeSaW
Nv IYH
ixH EB
x Re TYT
id
m
K IM
e$
eoB
d LWLB
v d
m TL
e oS IR
omo$
d iEv
W
(Exod 29:40)
“And a tenth of a measure of semolina, mixed in a fourth of a hin of beaten oil,
and a libation of a quarter hin of wine for the first21 lamb”
This preposition may also act in a distributive sense, implying that the non-
animal material is offered “per” zoeme, as in the following passage pertaining
to the New Moon offering:
IYiY
m o$e
Bd
eKxL IYH
ixH TE
i YB i LYiA
i RvW xmL IYH
ixH T$
i oYL
iv$ioW RP
m
dxL HY
eH
viY IYH
ixH YC
iq
X OH
e YK
a
dvSiW
Nv
(Num 28:14a)
“Their libations should be: half a hin per bull, one third of a hin per ram, and a
quarter of a hin per lamb––[all these libations being] wine.”
Thus, both possessive pronominal suffixes and the prefix -L v unambiguously
express the subordination of non-animal to animal materials.
x to connect a cereal offering and a zoeme (that is,
By contrast, the use of LE
cereal offering X offered ‘al [“upon”] zoeme Y) does not in and of itself imply
subordinate jugation of X to Y.22 Admittedly, in some instances LE x is used
v , as the form “X ‘al Y” is used in cases where X is undoubtedly a
like -L
subordinate jugate of Y.23 However, in many sacrificial contexts (as elsewhere
in BH), LE x may also denote “with” or “in addition to,”24 implying sacrificial
20
Note that even if the semolina and wine are interpreted as subordinate (that is, jugated)
to the entire sacrificial complex, there is nevertheless clear “agreement in number” with the
zoemes, in that the amounts of grain offerings and libations correspond to the individual
zoemes. Hence, it is appropriate to speak of jugation to a particular zoeme.
21
Note the translation of DX mAeHm o$e KxL as “for the first lamb” (NRSV) rather than “per
Bd
e
lamb,” as context implies (cf. JPS; note the use of DX mA eHm in v. 39 and the repetition of this law
in v. 41). In the phrase d HdL-H$
m oe q
ETx d
d Hdm
KvSiKNviW RQ edoBx
H TX xvM
NidK (literally, “like the HX
v Ni of the
mvM
morning and its [f.] libation you shall offer with it [f.],” Exod 29:41), the pronominal suffix in
Hd
d m
KvSiN clearly does not refer to the HX Ni . The feminine pronominal suffix here, as in d
mvM HmL,
refers elliptically to the HL m oE (i.e., -H$oe q
ETx [RQ
d edoBxH TL xE o v
Ve NviW =] d
SeK Hdm
Kv NviW RQ
SiK edoBx
H TL
x oE TX
xvM
Nid
v
K
[OYiB
d
x RvE
xH
m IYB
da /BRee
EHm TL xEo vL=] d
HdL). In the Samaritan, which reads H$ET WKSNKW RQBH TXNMK
m
WL, both pronominal suffixes probably refer to o$e K (which is also mentioned earlier in the
Bd
e
verse), or else to RQ H. This may have resulted from an attempt to harmonize this verse with the
edoBx
parallel in Num 28:8, which reads H$ i
oe qETx WodKv
d SiK
Nv W RQ edoBx
H TX xvM
NidK.
v
22
Contra Levine 2002:125–35, who juxtaposes Lev 7 (where in his opinion the cereal
offerings are adjuncts) with Lev 23:20 (where in his opinion the animal materials are adjuncts to
the vegetable materials).
23
See for example Num 15:5.
24
HALOT 2.826 (6.b); note in particular Exod 12:8, 9, and compare Exod 35:22 with 35:29.
11:50:13:06:14 Page 69
Page 70
coordination––that is, a situation in which two materials are co-jugates.25 In

fact, a comparison of Lev 14:20 and 31, passages that pertain to the offerings
of wealthy and indigent persons (respectively) afflicted with scale disease,
suggests that the use of LE x versus simple conjunctions is merely a matter of
stylistic choice.26 In such passages, “X ‘al Y” can be equivalent to “X and Y.”27
Thus, it appears that the term LE x, unlike possessive suffixes and -L v , does not
in and of itself imply subordination, as it may also denote coordination, or
even spatial relation (another obvious usage of LE x).28 As a result, it is some-
times difficult to ascertain whether a cereal offering is a primary sacrifice
(that is, semolinaA) or an adjunct to a wholeburnt or wellbeing offering
(semolinaB).29
Having ascertained this, one may begin to schematize the diverse
jugational structures found in P. A number of examples follow in schematic
form, proceeding logically from the least contested to the most equivocal
passages.
3.3.1 Numbers 6: The Law of the Nazirite
Let us first consider the law of the Nazirite (Num 6), as it proffers the clearest
example of two grain offerings on different levels (here, A and B) within a
25
The coordinating sense is similarly manifest in the concatenation––expressed through
x ––of various pairs of coordinate sacrificial complexes (such as the Sabbath sacrifice offered
LE
DYM
iT mx
d H TL x, Num 28:10 et passim), which will be discussed in the next chapter.
x Eo LE
26
Lev 14:19–20 HXNMH-TAW HLEH-TA IHKH HLEHW . . . TAUXH-TA IHKH H$EW
HXBZMH “and the priest shall offer the purification offering. . . and offer up the wholeburnt
offering and the cereal offering upon the altar”
Lev 14:30–1 HXNMH-LE HLE DXAH-TAW TAUX DXAH-TA . . . H$EW “and he shall
offer. . . the one as a purification offering and the other as a wholeburnt offering with (LE) the
cereal offering”
Since the two rites of the purification are equivalent in terms of their praxemics, except for
the obvious differences resulting from the use of different zoemes (birds replacing quadrupeds),
and since it is hardly likely that the bird is physically placed “on top of” the cereal offering
(another obvious denotation of LE), it appears that the phrase “X LE Y” may be equivalent to
“X and Y.”
27
On the evidence of Lev 23:18–20, one is tempted to argue that the units preceding and
following LE x are interchangeable. However, this is so only in the final form of the text, which is
a syntactic monstrosity. See Milgrom’s detailed treatment of the history of these verses in
Leviticus 2.205–8.
28
Semantic shifts in the Hebrew language led to a situation in which coordination could
no longer be expressed by LE x. Compare, in particular, the difference between dHMmiE (expressing
d
coordination) and H m YL
emE (expressing spatial relation, as well as subordination) in TS 20:9:
HNWBL HYLE BRQ R$A HXNM LWKW UP$MK VSN HME BRQ R$A HXNM LWK (“every grain
offering accompanied by a libation according to the rule, and every grain offering upon which
frankincense is offered”). The cereal and libation offerings are clearly coordinate, whereas the
frankincense is undoubtedly a subordinate jugate of the cereal offering.
29
On the distinction between coordinate and subordinate grain offerings, see Gray 1903:405.
Albeck’s rule of thumb (quoted in Milgrom, Leviticus 1.200), according to which a cereal
11:50:13:06:14 Page 70
Page 71
Jugation 71
Fig. 7. Relation of jugates in Numbers 6.
single context. The list of zoemes in verse 14, {s웧wholeburntA, s웨purificationA,

S웧wellbeingA}, is followed by two types of cereal offering (15a), and finally by the
formula OH e YKav
d SiW
Nv OT
mm NiiW (“and their cereal offerings and libations”).
XvM
Though this formulation leaves much unstated,30 when read in conjunction
with v. 17 (TWodCx M
dxH LS
x LE x HWHYLx OYMimLv$o XB
xeZ H$
oe q
ExY LYiA
xHm TA
evW, “he shall
offer the ram as a sacrifice of wellbeing to , together with the basket of
unleavened bread”), at least this much becomes clear: whereas the two types
of cereal offering are offered in conjunction with the S웧, another cereal
offering, as well as a libation, is subordinate to the same S웧, as in Figure 7
(note that the category “Nazirite’s offering” is not part of the jugational
pattern; structures of this type––represented also in Figures 8 and 9 as
superordinate categories––will be discussed in ).31 It has been
suggested that vv. 15b and 17b are not original, and thus that the mention
of jugates subordinated to the wellbeing ram reflects a later literary
development. In any case, it is clear that if they are in fact glosses, then the
interpolator(s) considered this structure permissible.32
offering listed before animal offerings is independent (as at Num 18:9–14 and Ezek 44:29) but a
cereal offering listed after the offerings is an adjunct (as at Lev 14:10; Num 6:14–15; Ezek 45:17,
24, 25; 46:14), is rightfully rejected by Milgrom as inadequate. For example, as Milgrom notes,
this rule does not apply to Lev 9:3–4, 17: here, the HX Ni is listed after the animals, and yet v. 17
mvM
implies that it is not an adjunct of any of them.
30
It is especially unclear what the pronominal suffixes of OH e YK
a
dvSiW
Nv OTmm NiiW refer to. From a
XvM
purely formal point of view, the possibilities are: (1) the wholeburnt and wellbeing offerings
only (in accordance with Num 15); (2) the purification offering as well (contrary to the general
trend in biblical law––see §3.7.2) but in accordance with the “grammar” of certain post-biblical
traditions; or (3) the unleavened cakes and wafers as well (highly unlikely, though syntactically
possible). Note that 17b makes particular reference to the subordinate jugates of the S웧wellbeing, but
no specific mention is made of the subordinate jugates of the s웧. In all likelihood, these pro-
nominal suffixes do not refer back to the grain offerings referred to in 15a: the wording of verse 17
(WodKvSiN . . . WoTX Ni , not OH
mvM e YK
av
d SiN . . . OT
mm Ni ) makes it perfectly clear that it is the zoeme (in this
XvM
case, S웧wellbeing), not the grain offerings, that entails subordinate jugates of grain (and wine).
31
Note that each item of the A-level unleavened breads has a subordinate B-level jugate of
oil according to v. 15 (not indicated in Figure 7).
32
See, for instance, Gray 1903:67–8 (with some hesitation); Rendtorff 1967:33; Noth
1968:56; but cf. Ashley 1993:146–8.
11:50:13:06:14 Page 71
Page 72
A general rule of jugation can therefore be formulated tentatively, based on

this example: non-animal materials are not bound to a particular jugational
level––that is, the same material may be offered on different jugational levels
even within a single sacrificial complex. In this case, the zoemes and the grain
offering of wafers and cakes are A-level co-jugates, while the raw semolina is
a (B-level) subordinate jugate of one of the zoemes. Furthermore, note that
while the two grain offerings are on different jugational levels, the one is not a
direct subordinate of the other.
3.3.2 The Person Purified from Scale Disease
In light of this analysis, it appears that the “three-tenths of a measure of

choice flour” listed alongside the zoemes {s웧wholeburntA, s웧reparationA, s웨purificationA}
required of the person purified from scale disease (Lev 14:10) are not sub-
ordinate (B-level) cereal offerings but A-level coordinates of the zoemes.
At first glance, it is tempting to interpret these “three-tenths of a measure
of choice flour” as subordinate jugates: each zoeme, according to this inter-
pretation, requires one-tenth of semolina mixed with oil, in accordance
with Num 15:4, which assigns precisely this measure of cereal offering for
ovines (with the exclusion of rams).33 This interpretation appears to be
supported by the fact that a leper of insufficient means is allowed to offer
only one lamb, one-tenth of a measure of semolina, and two birds.34 Yet
upon closer examination, it becomes clear that this law has nothing to do
with the lists in Numbers. According to Numbers, purification and reparation
offerings require no subordinate cereal offering whatsoever, so that the
wealthy person would have been required to bring only one-tenth to begin
with.35 Additionally, even the wealthy person’s wholeburnt offering is not
subsumed under one of the categories listed in Numbers 15 (volitional,
votive, or calendric) and therefore would not entail a cereal offering.36
Compounding the issue, the absence of a libation in Leviticus 14 would
remain unexplained.37
33
See also Milgrom, Leviticus 1.845–6, who nevertheless states that these three-tenths of
semolina are still considered an independent offering.
34
See b. Men. 91a; the birds, we recall, do not entail jugates, according to the law in Num 15.
35
See Marx 1994:36–7, who nevertheless considers the grain offering to be subordinate (to
the wholeburnt offering alone). See also Milgrom, Leviticus 1.845, and §3.7.2.
36
But see Num 15:24, where a non-calendric mandatory wholeburnt offering is also believed
to have “proper” subordinate cereal and wine jugates.
37
Following Rendtorff’s suggestion (1967:15) that the mention of libations sometimes
reflects a later text-historical development (in Num 28–9), one may speculate that Lev 14
reflects an intermediate stage, when the subordinate grain offering, but not the libation, had
been required in priestly circles. Though highly conjectural, this process would not be without
parallel: see n. 000.
11:50:13:06:14 Page 72
Page 73
Jugation 73
These considerations militate against the interpretation of the three-tenths

of an ephah mentioned in verse 10 as three separate portions of one-tenth
of an ephah,38 each jugated to a different zoeme. This conclusion is also
corroborated from the point of view of the sequence of ritual acts, since the
cereal offering is not divided and attached (in the most physical sense) to
the flesh of each zoeme, but offered as a whole along with the wholeburnt
offering (v. 20). We must therefore conclude that the cereal offering of the
person purified from scale disease has nothing to do with the laws of sub-
ordinate jugation laid out in Numbers 15. Rather, it is clear that this offering
is, like some of the cereal offerings of the Nazirite (Fig. 7), an A-level cereal
offering, which, as in the case of the Nazirite, is closely linked to one of
the zoemes but is still not subordinate to it. The particular zoemes differ
from one sacrificial complex to the next––an s웧wholeburntA in the case of the
wealthy leper, a /wholeburntA for an indigent leper, and an S웧wellbeingA for the
Nazirite––but all three situations involve independent grainA offerings that
are co-jugates of their respective zoemes.
3.4 MORE COMPLEX EXAMPLES
3.4.1 Exodus 29 and Leviticus 8
The prescriptions concerning the ordination offering complex (Exod 29) and
the consequent descriptions of this complex (Lev 8),39 which list three zoemes
and three types of cereal offerings (Exod 29:2; Lev 8:26),40 can be schematized,
as seen in Figure 8 (the oil accompanying each cereal offering is not indicated
in the diagram). The three types of cereal offerings are clearly closely
linked from a praxemic point of view, since they are placed in a single
basket. This is indicated in Figure 8 by the square brackets including the
cereal offerings. These in turn are praxemically linked to the S웧ordinationA
(immediately adjacent to the cereal offerings in Fig. 8), since in the course
of the ritual they are physically placed together with parts of this zoeme
(Exod 29:22–4; Lev 8:25–7).
38
BH OYN iR
o$ov e
E H$ $o, like modern English “three-tenths,” can mean both “three one-tenth
moLv
portions” or “one three-tenths portion.”
39
See Rendtorff 1967:12; Milgrom, Leviticus 1.545–9 and Nihan 2007:111–12, concerning
this ritual and the textual relation between Exod 29 and Lev 8.
40
Note that these types of cereal offerings are here listed slightly differently from Exod 29:23;
see Chavel 1982:292. Concerning the bull mentioned in Exod 29:36, see Propp 2006:469,
Milgrom, Leviticus 1.562.
11:50:13:06:14 Page 73
Page 74
Fig. 8. Relation of jugates in Exodus 29 and Leviticus 8.
A parallel example can be found in Leviticus 9.41 As in the two previous

passages, there is no evidence that the cereal offering mentioned in Lev 9:4, 17
is subordinate to the wholeburnt offering of v. 16 (or to any other zoeme
listed in the chapter).42 As demonstrated in the example from Exodus 28 and
Leviticus 8, cereal offerings defined as coordinates with animal sacrifices may
nonetheless be more closely linked to one zoeme than to another, just as two
zoemes within a sacrificial complex may be more closely linked to each other
than to other coordinate zoemes (see also the B웧purification and 웧purification at
Lev 16:18–19).43
It is noteworthy that when this occurs, the subordinate jugates are linked
to a wholeburnt or wellbeing offering,44 just as cerealB is always subordinate,
according to Numbers 15, to wholeburnt and wellbeing offerings.
41
See Rendtorff 1967:13–14; Marx 1994:107–108; Milgrom, Leviticus 1.569–91.
42
Contra Ibn Ezra on 9:4 (DXAW DXA LKL TLS AYH, Shachter 1986:34–5: “the appropriate
amount of flour for each one of the sacrifices”). See Marx 1994:108, who––though not entirely
systematic in his terminology––uses the phrases “en conjunction avec un holocauste” and states
that “l’offrande végétale [est] associée à l’holocauste” when referring to the relation between
this grain offering and the zoeme (emphasis mine). The mention of the cereal offering between
the offering of the wholeburnt offering and the phrase RQ edoBx
H TLx Eo DB
xv
dLMi may be purely
coincidental, since RQedoBx
H TL
x Eo DB
xv i , syntactically a dangling phrase, may have been inserted in
dLM
the wrong place here. In fact, the HX Ni is listed in the very last place on the “list of ingredients”
mvM
(9:4aβ), after the wellbeing offering, and within the procedural list between the wholeburnt and
wellbeing offerings (see Rainey 1970 on three types of listing, and Milgrom, Leviticus 1.573–4).
There is, however, one slight indication that it may have been conceived as subordinate: the lack
of specification concerning the amount of cereal. This omission may suggest that the author
assumed the reader’s familiarity with Table A (Num 15) or with a comparable scheme (though
the verse does not specify UPm
dv$oMid
dxK). See Milgrom, Leviticus, who also notes that according to the
list of jugations in Num 15, these jugates are applicable only in the post-conquest period (15:2).
43
See Gane 2005:221–2. From the manipulation of their blood, it is clear that the two zoemes
are more closely related than either is to the coordinate S웧wholeburnt.
44
For jugates linked to a wholeburnt offering, see Lev 14:20, where the wholeburnt offering
and the cereal offering are both left for the very end of the ritual. For jugates linked to a
wellbeing offering, see Num 6:18–19, where parts of the two are physically conjoined (see also
v. 31). See also Exod 29:22–4 and Lev 8:25–7, where the ordination sacrifice explicitly serves as
a prototype for future wellbeing sacrifices (Exod 29:28).
11:50:13:06:14 Page 74
Page 75
Jugation 75
3.4.2 Leviticus 7: The Thanksgiving Offering
Having witnessed A- and B-level cereal offerings, we are now in a position to

look at an example of a simple structure that nevertheless contains such
torturous language that establishing its basic meaning requires substantial
attention.
In Lev 7:12–15, concerning the law of the thanksgiving (HD d ) offering,
mWoT
the multiple uses of the preposition LE x in describing a single sacrificial rite
yield a particularly difficult text.45 Consider first vv. 12–13a:
TWodCM
x YQ
aYQ i IM
i RvW ee$
dod
x o iWLB
B TL d TWodCM
v x TWodLxX HD
mWoT W . . . (12)
dxH XB xe-LE
Z x BYRiQ v iHv
OXeeL TL
ox
d X-LE x (13a) 46 IM
em$
dod
x o iWLB
B (TL d TL
v ox
d X) TK ie RvM
eB u TL
eSovW IM
e$mdod
x
B OYX iu$oM
v
WoNB
d
m RvQ
m BYRiQvx Y JMam
X
(12) “Upon” the sacrifice of thanksgiving, he shall offer unleavened cakes mixed
with oil, unleavened wafers spread with oil, and choice flour well soaked in oil.
(13a) “Upon” cakes of leavened bread he shall offer his offering [i.e., the cakes
and wafers].
Some scholars assume that the vegetable materials in this ritual are sub-
ordinate to the zoeme.47 However, this is far from clear. We have already
seen that the word LE x does not always imply subordination. If LE x were
consistently used to express subordination here, verse 12 would imply that
the unleavened cakes and wafers are adjuncts of the zoeme,48 thereby contra-
dicting 13a, which would imply that the same unleavened cakes and
wafers are adjuncts of the leavened cakes rather than of the zoeme. Matters
would be further exacerbated by the fact that the verse concludes with
WYM
mmL$
vo TDxWoT xeZ-LE
d XB x (“ ‘upon’ his thanksgiving sacrifice of wellbeing,” 13b),
which would again imply that the unleavened cakes and wafers are adjuncts
of the zoeme.
45
On the problematic use of this same preposition in the beginning of v. 12 see for example
Rendtorff 1985:249–50.
46
The words TL o iWLB
d TL
v ox
d X, here in parentheses, are absent from LXX and are likely the
result of a scribal error (Elliger 1966:82), possibly due to vertical dittography (compare
TLo iWLB
d TWodCM
v x TWodLxX in 12a). I have emended the JPS translation quoted here accordingly.
Concerning the translation of the terms QYQ i Rmand HLmx
d X, which may be analogous to ninda-gur4-
ra and ninda-sal-la respectively, see Levine and Hallo 1967:56–7. See also del Olmo Lete 2007 for
a new interpretation of Ugaritic dgt and gdlt.
47
This is apparently the view of Rendtorff, who considers the vegetable material in this ritual
a “Zusatzopfer” (1967:169; 1985:250–1). See his use of this term in 1967:169–70 to refer to grain
offerings that are undoubtedly subordinate.
48
There are of course other uses of LEx, even in this passage (note that HD
mWoT
d LEx should be
translated as “for thanksgiving” or “on the occasion of thanksgiving”). In some cases, the
vegetable materials are offered literally “on top of” (LE x) the animal materials (e.g., Lev 8:26).
This in itself cannot be taken as indication of subordination (even though intuitively it may
seem reasonable) any more than the fact that OYB imLq
XxH are offered TWoZX
meH LEx suggests that the
internal organs are subordinate to the breasts (Lev 9:20).
11:50:13:06:14 Page 75
Page 76
One might argue that the unleavened cakes and wafers are subordinate
both to the zoeme and to the cakes of leavened bread, or that they are sub-
ordinate to the cakes of leavened bread that are, in turn, subordinate to
the zoeme.49 However, P does not state this, and the situation would be
unparalleled. Thus far, we have encountered cooked grain offerings only as
A-level offerings, never as subordinate jugates.50 Therefore, it stands to reason
that the cooked vegetable materials here are also coordinate jugates of the
zoeme, in which case we should figure the thanksgiving offering in P
(Lev 7:12–14) as Figure 9 shows.51
Fig. 9. Relation of jugates in Leviticus 7.
It is unclear whether the law in its present context requires that the entail
subordinate jugates (of cereal and wine) as well. In other words, it is unclear
whether the coordination of four different cereal offerings (Fig. 9) replaces
the “ordinary” subordinate jugation of Table A. We shall see in Section 3.7.1
that the law represented in Table A above was composed with calendric offer-
ings primarily in mind. Therefore, it is likely that the law in Leviticus 7
required only the coordinate jugation of bread for wellbeing offerings in
the context of thanksgiving, and apparently no jugation whatsoever for
other freewill wellbeing offerings.52 Note that the bread accompanying the
49
Perhaps WYMmmLv
$o TD xWoT
d XBxeZ LE x R$ q would have expressed this if it qualified OX
eoA ee
L TL
ox
d X
JM
amX, but certainly not WYMmmLv
$o TD xWoT
d XB x.
xeZ LE
50
I thank Yehuda Kraut for this insight. Raw semolina, on the other hand, can be either an
A-level jugate (offered alone or as a coordinate of a zoeme) or a B-level jugate subordinated to
a zoeme.
51
Note that each of the A-level unleavened cereal offerings entails a B-level jugate of oil
(not indicated in Fig. 9).
52
See Levine 1974:43 on the diachronic development of shelamim and toda. Rabbinic
tradition determines that the thanksgiving offering entails subordinate jugates, like most
wholeburnt and wellbeing offerings, in addition to these coordinate jugates, which are called
HX Ni neither in P nor in the rabbinic texts (see b. Men. 46b concerning the “loaves of thanks-
mvM
giving offering” and Maimonides HMQ 9:14).
11:50:13:06:14 Page 76
Page 77
Jugation 77
thanksgiving offerings is not termed a HX mvMNi and consequently should not be

considered as such.
This hypothesis is further supported by the fact that elsewhere in P
(specifically Lev 8–9) the offering of (cereal) coordinate jugates along with
zoemes may sometimes preclude subordinate jugates (like cereal and wine).
One might even conjecture that in a more ancient form of priestly tradition,
subordinate and coordinate jugations of cereal offerings were mutually
exclusive.53
However, elsewhere in P this restriction breaks down. Admittedly, the
author who composed Num 15:3, expanding the scope of the law represented
in Table A to non-calendric freewill offerings, may not have been aware of
the law in Leviticus 7;54 if this is the case, the problem as such did not arise
for this author at all. However, when both laws were compiled in the same
work––however far apart––there arose the possibility of “ritual conflation,”
in the form of offering a single zoeme with coordinate and subordinate cereal
jugates.
Furthermore, it is clear that later contributors to P saw no inherent contra-
diction between cerealA coordinate with and cerealB jugated to the same
zoeme. In Num 6:15, it appears that such a mix of subordinate and co-
ordinate cereal offerings is not impossible.55 However, since the words
OHe YK
av
d SiW
Nv OTmm NiiW in this verse are likely the result of textual accretion,56 this
XvM
passage may belie a diachronic development in the grammar of P’s rituals.
Specifically, one might venture to generalize the following diachronic
development within P’s system:
Either coordinate or subordinate jugation of cereal to zoeme (i.e., cerealA and
cerealB are mutually exclusive).
Both coordinate and subordinate jugation of cereal to zoeme (cerealA and cerealB
are not mutually exclusive).
As we shall presently see, diachronic shifts like this one are not uncommon in
Σ. Before turning to these, however, a final comment is in order concerning a
rare case in which a zoeme appears to be subordinate to non-animal material.
53
Note, however, that the lack of subordinate jugates in texts such as Lev 8–9 could also be
explained on literary-narrative grounds: the Israelites have not yet entered Canaan, and Num 15
suggests that the “table of jugations” applies only after that event.
54
Amos 4:5 may testify to the antiquity of the tradition of offering leavened cakes in the
context of thanksgiving, but the precise composition of the offering referred to in the verse, as
well as the relation between animal and non-animal materials within this offering (if such a
relation exists at all), cannot be ascertained. See Levine 2002:132, which is somewhat
speculative.
55
See p. 000.
56
See nn. 000 and 000.
11:50:13:06:14 Page 77
Page 78
3.4.3 Subordination of a Zoeme to a Non-Animal Offering
There is probably no text more unusual from the point of view of jugation
than Lev 23:18–20, prescribing the materials to be offered on the occasion of
the first fruits (OYRiW
idKiB) of grain. The text, which likely suffers from a con-
d
voluted scribal history,57 is exceptional in giving rise to the possibility that
zoemes need not be independent materia sacra.
Whether or not this jugational pattern was the intention of the text at any
historical stage cannot be determined with certainty. It stands to reason that
the zoemes mentioned in 23:18 (7s웧, 1B웧 and 2S웧) are subordinate to
the loaves baked from the new grain, since the new grain is the immediate
occasion for the celebration.58 The matter cannot be ascertained on purely
linguistic grounds precisely due to the multivalence of the preposition LE x
demonstrated above: the zoemes are said to be offered in addition to (LE x) the
two loaves, but the word LE x, as we have seen, may imply coordination or
subordination. In any case, as wholeburnt offerings these zoemes in turn
entail subordinate cereal and wine libations (18b), as in Figure 10 (oilD is not
indicated in the diagram).
The jugational relation between the loaves of bread and the zoemes
mentioned separately in verse 19 (G웧purification, 2s웧wellbeing) is unclear; matters
are further complicated by the formula OY$ oi m
Bd
vK YN
a$ vo-LE
x in v. 20, in which
LEx may denote physical relations.
However tentatively, I would suggest that the zoemic materials mentioned
in Lev 23:18–19 are in fact subordinate to the bread mentioned in v. 17––and
that this ritual presents a precedent for subordinating zoemes to non-zoemic
material. With even greater hesitation one might suggest that if the “sheaf”
mentioned in vv. 10–12 is considered an offering, then the lamb mentioned
in v. 12 may be another example of a zoeme subordinate to non-zoemic
material. Here, too, the rationale would be that the presence of new grain is
the immediate occasion for the offering of zoemes, and not the other way
around. There is, however, little evidence that in P this “sheaf” is in fact an
offering.59
57
For treatments of the text history of Lev 23:18–20 in general and v. 20 in particular, see
Elliger 1966:302; Milgrom, Leviticus 2:2005–8.
58
See Rendtorff 1967:178 (claiming that the two grain offerings are on different levels);
Levine 2002:130–1. See the controversy in m. Men. 4:3, where subordination of jugates is
formulated in terms of causality––whether the absence of one jugate entails the cancellation of
another.
59
These precedents were apparently picked up by the authors of the Temple Scroll, who
exploit it on numerous occasions, e.g., with reference to wine and oil (19:11–20:8; 21:12–22:8).
However, since these examples depend to some extent on textual conjecture they cannot serve as
a basis for this , as this might lead to circular argumentation.
11:50:13:06:14 Page 78
Page 79
Jugation 79
Fig. 10. Jugational pattern in Leviticus 23:18–20.
3.5 DIACHRONIC JUGATION
3.5.1 Typical Subordination to Zoemes
As in zoemics, not every aspect of the jugational category remains fixed, some
features proving to be less stable than others. Despite the numerous trans-
formations that Σ underwent in the late Second Temple period, the “Table of
Jugation” (Table 2) did not undergo significant alteration. Admittedly, Ezekiel
46, which may or may not have been composed with an awareness of the
Table of Jugation in Numbers 15,60 contains its own “table of jugation.” But
there is little indication in the ancient post-biblical sources (including
late Second Temple texts and rabbinic literature) that different jugational
60
Ezekiel offers alternative measurements at 45:18–46:18. The Aramaic Document of Levi
elaborates on P’s table of jugations (9:1–18). The Temple Scroll (13–14, 18) also contains
enough fragmentary information to allow a plausible reconstruction of its system, which was
(apparently) essentially identical to that of P as far as the measurements are concerned. All of
these systems have been reconstructed in tabular form and discussed in detail in the scholarly
literature; see Greenfield et al. 2004:41; Schiffman 2005:190–9; and Himmelfarb 2004.
11:50:13:06:14 Page 79
Page 80
measurements were adopted. The texts referencing these details––namely

ADL, the Temple Scroll, and several rabbinic sources61 ––reflect the quantities
of Numbers 15 in all essentials.62 The differences between ADL and P are
either due to mathematical approximation (0.3 ≈ 1/3 in ADL 9:10)63 or to an
expansion of P’s “Table of Jugation” to include other non-animal materials,
namely salt and wood. The measurements, one might say, have a strong
grammatical resiliency from a diachronic point of view. Other features prove
less resilient to change, such as the applicability of subordinate jugation to
purification offerings, or the jugational laws of frankincense and salt, as we
shall presently see.
3.6 THE STRENGTH OF LINKS AMONG JUGATES
3.6.1 Evidence from Jugational Patterns
The occasional omissions from jugational lists can be quite telling. As we shall
see, the texts of P present jugational patterns in which a wine libation often
goes unmentioned, in stark contrast to the cereal offering. This suggests the
relative weakness, from a diachronic perspective, of the link between zoemes
and their subordinate libations. Moreover, later texts also contain evidence
that the link between zoemes and libations is weak even when wine is present.
In P, it is difficult to determine whether the B-level cereal offering and
the B-level libation are equally subordinate to the zoeme. Insofar as their
formulation is concerned, both the cereal offerings and the libations above
appear to be on equal ground, but it is noteworthy that the cereal offering is
always mentioned first in such formulas.64 Additionally, in Num 8:8, a whole-
burnt offering entails a cereal offering but no libation, another indication of
the greater strength of the cereal offering’s link than that of the libation.
Note that the priestly narrative in Num 7:87 states that grain offerings,
but not libations, are donated along with zoemes intended for wholeburnt
offerings, further suggesting that these zoemes entail only grain offerings.65
61
See the description in Sheq. 5:3. LXX and other translations that accord with P can hardly
be considered independent evidence in this case.
62
See Schiffman 2005:197.
63
Thus Schiffman 2005:196 n. 75, referring to Himmelfarb 2004 (Schiffman refers to p. 192,
but this is apparently a misprint, as the reference should be to pp. 114–15).
64
For example, Num 29:21, 24, 27, et passim. For an extensive review of the relevant passages,
see Rendtorff 1967, particularly pp. 15–16, 167, 170.
65
See n. 000. Note that LXX reads κα α- σπονδα α.τν, implying that wine was offered
as well. However, this is probably motivated by a harmonistic drive: note that wine is not
mentioned anywhere in the earlier sections of ch. 7.
11:50:13:06:14 Page 80
Page 81
Jugation 81
However, the evidence here is uncertain, since wine could go unmentioned

due to narrative considerations––for example, considering that in the story
the Israelites are in the wilderness where wine is difficult to come by––though
wine would still be expected to be jugated with the zoeme at the time of its
offering.66
More importantly, the Temple vision in Ezekiel consistently refers to B-
level grain offerings that are jugated to zoemes (for instance, 45:24, 46:5, 7,
11, 14, 15), along with C-level oil, enumerated separately as a subordinate
jugate of the semolina (45:24b, 46:5b; 7b, 11b; and 46:14aβ). Yet Ezekiel is
silent concerning B-level libations. In fact, except for one verse (45:17), the
Temple vision makes no mention of libation whatsoever. In this verse, it is not
provable that the libation in question is a B-level libation rather than an
(independent) A-level libation. Though it is difficult to imagine why the
author would refrain from mentioning B-level libations,67 their absence from
his jugational lists is too consistent to be coincidental.
Though it is perhaps unlikely that the author had P documents before him
when composing his own fantastic cultic world,68 it is evident that he knew of
a tradition where cereal and libation offerings were jugated to zoemes. If not
from P, he might have known this from a host of other ancient Israelite
sources where this triad is found (1 Sam 1:24; 10:3; Hos 9:4), or from hearsay
of ancient practice. In any case, it can hardly be doubted that he knew of it––
this is immediately evident from 45:17.
The Temple vision of Ezekiel systematically rejects the triadic model
{zoemeA, grain offering B, wineB} and (with the exception of 45:17) replaces it
with an alternative, linear model: {zoemeA, semolinaB, oilC} (Figure 11).69
Fig. 11. Ezekiel’s linear jugational model.
66
Recall that subordinate jugation in general, according to Num 15:2, does not apply in the
wilderness; see n. 000.
67
One must not conclude from Ezek 20:28 (OH e YK
av
d SiN-TAe O$ moiWKYSid
d W “they poured out their
xYx
libations there”) that Ezekiel considers this particular form of worship offensive, since the verse
mentions animal offerings as well. Furthermore, the one slip indicates that Ezekiel did not
abolish A-level libations.
68
See Haran 2008, 2009a.
69
Note that in Ezekiel this entails a semantic shift in the meaning of HX Ni . In Ezekiel, the
mvM
term almost always means “cereal” in the most limited sense of the word (i.e., to the exclusion of
the oil that is mixed in with the semolina). In fact, the terms HX Ni and TL
mvM eSo appear to be
interchangeable (see 46:14a, but cf. 14b, where HX mvMNi must include the oil). See particularly
45:25, IMem$
dox
KvW HX
mvM
Nd
ixKv
W HL m . As we have seen, P is much less systematic than Ezekiel on this,
m WoEK
d
and it appears that the whole mixture may be termed HX Ni , as well as the semolina alone, since
mvM
it is the main (solid) component in the mixture. See p. 000.
11:50:13:06:14 Page 81
Page 82
Haran’s solution––namely, that Ezekiel did not have the details of P’s
system before him when composing his own ritual system––does not account
for the numerous direct quotations of P’s and H’s systems (not necessarily P
and H documents) throughout Ezekiel.70 Thus, it is more likely that the
author was aware of the traditional system reflected in (but older than) P,71
but that he replaced it with a different system: much of what was associated
with the ancien régime was abhorrent and had to be changed––first and
foremost, the architectural structure of the Temple and the detailed laws of its
cult (43:10–11).72 In this case, Ezekiel instituted a new hierarchic scheme, as
well as new measurements of semolina and oil, and obliterated the wine
libation.
While it is impossible to ascertain why Ezekiel obliterated the wineB
libation rather than, for example, the grain offering, the “grammatical” analy-
sis offered here suggests one solution. We have seen that the link between a
zoeme and its subordinate wine libation is weaker than the link between
a zoeme and its subordinate grain offering. Since the wine libation was
traditionally less tightly linked to the zoeme, it is possible that it was more
easily detachable from the zoeme.73
A wider overview of priestly jugational patterns in general––including
coordinate jugation as well as subordinate jugation––reveals a similar situ-
ation: whereas cereal offerings are often coordinated to zoemes alone,
that is, without libation offerings,74 the reverse situation (only a libation
accompanying a zoeme, without a cereal offering) is basically unheard of in P.
In sum, it appears that a large number of sources from different periods
consider the cereal offering to be more closely linked to the zoeme than is the
70
See Haran 2009:166–8, 171, and nn. 22 and 34. A comparison of Ezek 45:21 with Num
28:16–17 should suffice for the sake of illustration, though this example alone does not
determine the direction of borrowing. (In Ezek 45:21, TE xB$ is to be considered original,
misconstrued in MT as TWoEB$; see Haran 2008:217 n. 9 and Zimmerli 2.481.)
71
See n. 000.
72
See Zimmerli 2.412. In this verse, TWodX Qu and TR d are not limited to architectural
o WoT
blueprints, but include cultic practices as well. That this is so is evident not from the use of the
verbiW$oEmvW, which may also refer to architectural constructions (e.g., WoTWoo$m
Ea d , 43:18), but
H OWoYB
v
from the fact that the cultic practices are introduced as TWodX Qu (43:18; 46:14) in the sections
immediately following this general statement in 43:11–12. See also 44:5, 24 (TWodX Qu and TR d ).
o WoT
73
See also Rendtorff 1967:167, 170 concerning the secondary nature of several of the
references to libations in Num 28–9.
74
As noted above at least twice, in the cases of the thanksgiving offering and of a person
purified of scale disease. Rabbinic tradition, which viewed the cereal offerings of the person
purified of scale disease as a cerealB offering, assumed (by analogy, though it is formally derived
from Num 15:5 in a midrash halacha, b. Men. 91a) that it is also accompanied by a libation. The
same is true for the s웧 offered by a parturient (Lev 12:6) and the S웧 offered by the high priest
according to Lev 16:3, which are wholeburnt sacrifices (see b. Men. 91b and Maimonides HMQ
2:2, 6). In contrast to these, which––though historically unlikely––are at least formally derived
from Num 15:5, the rabbinic statement that the cereal offerings of the person purified from
scale disease are B-level jugates is not derived even formally from a verse but attributed to
revealed oral tradition (HEWM$H YPM).
11:50:13:06:14 Page 82
Page 83
Jugation 83
wine. In terms of our fast-food simile, it would be possible to order a burger

with fries and no soda, but never a burger with soda and no fries.
3.6.2 Evidence from Praxis
There is also evidence suggesting that, at least in certain periods, the libation
was offered at a greater physical distance from the zoeme than was the sub-
ordinate cereal offering. If praxis is an indication of jugational tiering, this
may further corroborate that the jugation of wine is weaker than that of
cereal. The evidence in question comes from the present text of Num 28:7,
which, according to some scholars, suggests that the libation jugated to the
tāmîd complex was physically removed from the offering of the meat, taking
place in a separate chamber.75
Even when offered at the same altar, as appears to be the case in most
biblical passages, it is possible that the wine libation was still farther away
from the flesh than was the cereal offering. Unambiguous evidence of this is
already found in Sirach 50:15: σπεισεν ξ α0µατο σταφυλ1 ξ#χεεν ε+
θεµ#λια θυσιαστηρου (“He made a libation from the grape-blood [i.e.,
wine, and] poured [it on]to the foundations of the altar”).76 By contrast, the
cereal offering––or the portion of it that was offered77 ––was undoubtedly
placed on the fire (see explicitly Lev 14:20, 31), along with the flesh (or suet as
the case may be) of the zoeme to which it was subordinate.
The evidence from Sirach is corroborated by Josephus and rabbinic
tradition, suggesting that such practice was common in the latter part of the
Second Temple period.78 Whether this was the practice in earlier times is
difficult to determine, though one early text seems to suggest that the wine
75
See the opinion of Strack cited in Gray 1903; Milgrom, Leviticus 1.611–12; 1990:240;
contra Ashley 1993:564–5. This understanding is based on an interpretation of $oDeQ d
od
xB as
“inside the sanctuary” in the formula HWHYL x RK ao v
m$ Ve
Se Nv Va
Sx
d H $oDeQ
d
od
xB (Num 28:7). It is further
supported by the fact that P mentions golden libation vessels stored on the Table of Presence
(IH
adBv
m Vd
xSu
Y R$ q , Exod 25:29; 37:16; and v
eoA VeSd
mNx o vQ, Num 4:7). Note the specific prohibition of
H TWo$
offering libations on the inner altar (Exod 30:9). Contrast Haran 1978:216–17. As Milgrom
notes, the term RK ao is a cognate of Akkadian šikaru/šikru, a standard temple libation of “beer
m$
(made from grain), fermented alcoholic beverage” (CAD 17.420), i.e., “ale.” The author of
the Temple Scroll apparently viewed RK ao as a type of IYiYx, suggesting that the terms in Num 28:7
m$
and Exod 29:40 refer to the same substance: $DX IYY RK$ VSN VSNL (“to offer a libation of
RK$, new wine,” TS 21:10).
76
Note, however, that Sirach still terms this offering, which is not placed on the fire, 3σµ4ν
ε.ωδα 5ψστ7 παµβασιλε, as pointed out by Gray 1903:175.
77
See Regev 1996; Noam 2003:250–4, as well as n. 000 below.
78
See Josephus, Ant. Jud. 3.234–5, who indicates that the wine is applied περ τν βωµν
(around the altar, in contrast to the cereal offerings that are offered π τ βωµ). For the
rabbinic traditions see the texts cited in nn. 000, 000. See Dillmann 1886:82 (following Knobel),
who suggests a diachronic development from wine-application atop the altar (in ancient times)
to wine-application at the foot of the altar (as in Sirach and Josephus).
11:50:13:06:14 Page 83
Page 84
was offered on top of the flesh in Judea in the eighth century  (2 Kgs
16:12–13).79 Several other Second Temple sources, such as Jubilees (6:2–3;
7:3–6; 21:7–9)80 and ADL (8:6, where the frankincense seems to be placed on
top of the meat+cereal+wine),81 suggest that this practice persisted into the
last few centuries .
The above linguistic, statistical, and praxemic arguments strongly motivate
the following general rule:
The link {zoemeA, cerealB} is stronger than the link {zoemeA, libationB}
Read: A zoeme is more strongly linked to its subordinate cereal jugate than to its
subordinate libation jugate.
3.7 THE SCOPE OF SUBORDINATE JUGATION
There appears to have been an important difference between the wholeburnt

and wellbeing offerings in terms of their subordinate jugational patterns,
though the law in Numbers 15 mentions the jugation of both in one breath.
There is at least one case in P where a subordinate grain offering, namely
grainB, is more closely linked to wholeburnt (zoemic) offerings than to
wellbeing offerings. This is found in Num 7:84–8, in the summary of the
donations brought by the chieftains. Although the materials mentioned in
Numbers 7 were probably not sacrificed on the spot but rather reserved for
future use,82 the Priestly author thought of the grain offerings presented in
the silver bowls and basins as related to the zoemes designated for wholeburnt
79
Contra Wenham 1981:128. See already Knobel 1861:76, and more recently Seebass,
Numeri 3.254–5. King Ahaz, according to 2 Kgs 16:12–13, first ascends to the top of the
(Damascene-type) altar and there offers his zoeme(s), cereal offering, and libation. His
instructions concerning the “bronze” (Judean) altar seem to corroborate this (v. 15). However,
there is very little information about the praxemics of this sacrificial complex, and in fact it is
only a conjecture (though reasonable) that the ritual involves subordinate jugation.
80
Jubilees is not consistent in this respect. In one case (ch. 7) it appears that the wine is
offered directly on the fire (not on the flesh), but elsewhere it seems that it is on the flesh that
is on the altar. Jubilees 21:7 may reflect the conception {zoemeA, cerealB, wineC}, which would be
unique. The preservation of the Hebrew in 4Q219 (4QJubd) and 4Q220 (4QJube) does not
enable one to ascertain whether this structure is in fact envisioned. In praxemic terms, Jub 6:3
suggests that the grain offering is placed on the flesh, the wine on the grain offering, and the
frankincense on top. Jub 7:5, on the other hand, reflects a strong praxemic link between
the grain offering and flesh but seems to imply that the wine and the frankincense are placed
directly on the fire (perhaps on a different spot). See Himmelfarb 2004:119.
81
This evidence pertaining to physical proximity is supported by another aspect, pertaining
to temporal proximity: it appears that the wine is always offered after the cereal offering. In
Jubilees, the order is zoeme–cereal–wine–frankincense (7:3–6; 15:2; see ADL 8:6). See TS 34:13,
which suggests (if WNMM is the correct reading) that some of the wine is placed on the flesh, some
beside it.
82
See n. 000.
11:50:13:06:14 Page 84
Page 85
Jugation 85
offerings rather than to those designated for wellbeing or purification

offerings. This is clear from v. 87a, where the grain offerings are referred to
as OT
mm Ni :
XvM
-OYN
a$ vo HN
m$mo½YN
aBd OY$
v oi m
Bd
vK R$
om m
E OYN
a$vo OL
i YA
a OYRiP
m
d R$
om m
E OYN
a$vo HL
m EoL
m RQmd
m
BxH-LKm
d
OT mm NiiW R$
XvM o m
E
“Total of large cattle for wholeburnt offering––twelve bulls (B웧); of rams (S웧),
twelve; of male lambs (s웧), twelve; and their cereal offering[s].”
The pronominal suffix clearly refers back to the zoemes designated for
wholeburnt offerings only, not to the zoemes designated for purification
or wellbeing offerings, which are mentioned only subsequently (87b–8), and
where OT mm NiiW is absent. This piece of evidence, it must be admitted, is
XvM
83
atypical.
More to the point is the evidence from a non-Priestly text, 2 Kings 16.
According to this narrative, the offering of (possibly subordinate) libation and
cereal offerings is primarily linked to the offering of wholeburnt offerings,
both those made on the Damascene altar and those on the Judean altar.
Ezekiel seems to follow the same pattern, in that subordinate jugation
appears to be more applicable to wholeburnt offerings than to wellbeing
offerings. Whereas the Temple vision is explicit about the fact that whole-
burnt offerings entail subordinate jugation, it is unclear whether any laws of
jugation apply to wellbeing offerings.84
3.7.1 The Scope of the Law in Numbers 15
When considering the jugational scheme reflected in Num 15:2b–16,85 two

facts should be taken into account:
1. This scheme accords perfectly with the list of calendric offerings in
Numbers 28–9, where the legislators repeat the same ratios approximately
83
Similarly, if 2 Chr 29:35 (HLm EoL
m OYK
im Sd BiW OYM
vNx imLv$ox
H YBaLveXv
d
B BR
om m Eo-OG
L HL xWv) is translated (as
in JPS) “beside the large number of burnt offerings, there were the fat parts of the sacrifices of
wellbeing and the OYK imSvN for the burnt offerings,” then it would appear that only wholeburnt
offerings, not wellbeing offerings, entail OYK im SvN. It is possible that in this verse the term OYK imSvN
includes not only wine libations but, as in Qumranic and Mishnaic Hebrew, the complex
{libations + cereal offerings}. See p. 000. In any case, this piece of evidence is problematic, since
the verse is replete with syntactic uncertainties. See Japhet 1993:911, 931.
84
Ezek 46:4, 5, etc. The instruction TB d
mx $
dox
H OWoYB d H$
v oe q
ExY R$
eoA
qd
xK (“as he does on the Sabbath
day,” v. 12) clearly indicates that the same jugational rules of the calendric offerings in vv. 4–5
also apply to these non-calendric offerings. Whether one should go so far as to deduce that
even the wellbeing offerings mentioned in 46:12 entail subordinate jugates is difficult to
determine.
85
For a discussion of this passage in its current setting as well as its text history, see Seebass,
Numeri 2.137–41, Levine, Numbers 1.388–93; Achenbach 2003:517–18.
11:50:13:06:14 Page 85
Page 86
ten times verbatim.86 However, it does not accord with lists outside
Numbers 28–9, even within P and H. One example is the s웧 entailing
0.2 ephah (and 0.25 hin) of Lev 23:13 rather than the 0.1 ephah required
at Num 15:4. Another example occurs in Num 7:87, where only the
wholeburnt offerings of the chieftains entail cereal offerings (and,
according to LXX, libations) of unspecified quantities. This stands in
contrast to the requirement in Num 15:3 that volitional wellbeing
offerings also entail a cereal offering and a libation.87 Furthermore,
these jugates are mentioned nowhere in the laws of the wholeburnt and
wellbeing offerings in Leviticus 1–7, laws pertaining first and foremost
to non-calendric offerings. Had the legislators assumed that a non-
calendric wholeburnt offering entailed these jugates, it would be quite
striking that this fact is not at least implied.
2. The law in Numbers 15 in its present form contains a blatant dis-
crepancy. The protasis (2b–3)88 apparently refers to all quadrupeds
(birds do not entail subordinate offerings), as is evident from the
inclusive formula IAC do x
H½IMi WoA RQmd
mBx i ( or , large or small cattle).
H½IM
However, the detailed enumeration of the apodosis (the “table of
jugations,” vv. 4–10) mentions only three zoemes, o$e K (seemingly s웧),
Bd
e
x (S웧), and RQ
LYiA B-IB
md
m ie (probably 웧), completely ignoring goats,
female ovines, and (probably) female bovines.89
Some commentators have noted this problem.90 Levine, for example,
writes, “The present law includes goats in its provisions whereas v. 5 only
86
The ratios of wine are repeated once in full (28:14a, for all three zoemes) and once
concerning s웧 (28:7); the ratios of semolina and oil are repeated in full no less than six
times (28:12–13, 20–1, 28–9; 29:3–4, 9–10, 14–15) and twice concerning s웧 (28:5, 9, on
which see below). From the second day of the seven-day festival of the seventh month
onwards, the legislators enumerate neither the libations nor the grain offering, stating instead
UPm
dv $oM
id
dxK ORmmd
PvSi
MvdB . . . OH
e YK
av
d SiW
Nv OT mm NiiW or the like.
XvM
87
There is another discrepancy between vv. 3 and 24. The B웧wholeburnt mentioned in v. 24 is
said to require subordinate jugates “according to the fixed law” (UP m
dv$oMid
d x
K WodKvSiW
Nv WoTXNiiW,
mvM
apparently referring to the list in v. 9), yet this bull is not subsumed under any of the categories
listed in Num 15:1–16: it is neither votive, nor volitional, nor calendric.
88
This is clearly the protasis, though grammatically 2b alone could be the protasis and 3 the
beginning of the apodosis. Here, OT e Y$
oi q
ExW should be translated “and would [like to] make . . .”
(JPS “would present an offering. . .”). Compare 14a (protasis, analogous to 2–3a; in both cases
H$ om m W and OT
Ev e Y$
oi q
ExW are part of the protasis; see also Exod 12:48, where H$ om m
EvW is part of the
protasis) and 14b, where H$ oe q
ExY IK
a Woo$q
d ETx R$
d eoA
qdK implies that the same laws of jugation apply to
x
the offering of the RG as they do to that of the XRmvA Ze .
89
The two possible interpretations for RQ md
m
B½IBie are and 웧. As noted in nn. 000 and 000,
x-IB ie entails only “member of the group x,” not young age or male sex. However, where female
specimens are optional, P simply refers to the animal as RQ B, never RQ
md
m md
m ie . If RQ
B-IB mdm ie in this
B-IB
verse refers to females as well, it would therefore be the only case where this occurs.
90
Many commentaries simply ignore the problem (Wenham 1981:128; Seebass, Numeri
137–9) or misconstrue the zoemes implied in the passage (Noth 1968:114; Keil and Delitzsch
1.4.100 claim that o$e BdeK in v. 5 means “sheep and goats.” This, we have seen in , is an
11:50:13:06:14 Page 86
Page 87
Jugation 87
mentioned sheep. This discrepancy. . . is hardly significant in cultic terms.”91

It must be argued, however, that this example is precisely a “significant dis-
crepancy in cultic terms”––in Priestly literature, there is hardly an aspect of
ritual detail that could be more significant than the choice of a zoeme.
Furthermore, as we see here, the discrepancy is more serious than the omis-
sion of goats: other classes of animals are left out as well. Therefore, it seems
that the verses pose a serious problem that requires attention.
If the law in Numbers 15 had been formulated originally to include votive
and volitional wholeburnt and wellbeing offerings (as v. 3 in its present form
indicates), then the omission of goats (and of female ovines and bovines)
would be particularly odd: did Israelites never offer these animals as votive or
volitional offerings?
Note that it is only in the summary of the apodosis (v. 11), in what appears
to be an afterthought or an attempt at harmonization, that goats (of both
sexes), female ovines, and female bovines are suddenly resurrected from
oblivion: WoA OY$ oi m
Bd
vKx
B H$
odxL-WoA DX
mAeHm LYiAxm
L oWA DX
mAeH
m RWo$
dL
ox H$oe m
Ed
aY HKmK
m
d
OYZ
d
iEim B (that is, , S웧, and the rest of ).92
The conclusion is ineluctable: the table of jugations in Num 15:4–10 was
formulated with only the calendric sacrifices of Numbers 28–9 (or another
very similar text) in mind. Had vv. 4–10 been intended to cover the individual
volitional and votive wellbeing and wholeburnt offerings, the omission of so
many zoemes would be inexcusable.
It is now clear why female ovines are not mentioned: though acceptable as
wellbeing offerings, and probably offered as volitional and votive wellbeing
offerings, they are not part of the “zoemic inventory” of calendric sacrifices
as formulated in Numbers 28–9. The same holds true for female caprines,
which are not mentioned in vv. 4–10, and of female bovines, which––as we
have seen––are probably also unmentioned.
The fact that male goats were also left out is quite revealing. Although the
RYEi$om (웧) is often mentioned in Numbers 28–9 as part of the inventory
of the calendric sacrifices, it is nevertheless left unmentioned in the laws
of jugation in Num 15:4–10. The reason for this is now clear: the zoemic
inventory of Numbers 28–9 consists of only four zoemes: B웧, S웧, s웧, and
웧. The lists in Numbers 28–9 enumerate the precise quantities of wine
and cereal offerings for B웧, S웧, and s웧, but not for 웧. The author of the
error). Dillmann 1886:82, following Knobel 1861, assumes that the measurements that apply
to rams apply to mature male caprines.
91
Levine, Numbers 1.392 on 15:11.
92
The reader is reminded that RWo$o in P always denotes . The fact that the interpolator of
v. 11 found it necessary to replace RQ md
m
B½IBie with RWo$o when applying the laws of vv. 4–10 to
all types of cattle may imply that this interpolator understood RQmdm ie to denote male bovines
B½IB
only, to the exclusion of female bovines. Otherwise, the interpolator could have left the
word RQ md
m ie as it stands, and would have been content with replacing o$e
B½IB BdK with OY$
e oi m
Bd
vKx
B H$
oe
OYZ
d
iEimB WoA.
11:50:13:06:14 Page 87
Page 88
original “table of jugation” of Num 15:4–10 understood that the 웧 (always

a purification offering in the calendric cycle) requires no subordinate
materials whatsoever, and therefore that Numbers 28–9 does not list any
quantities of wine, cereal, or oil to accompany the goat. The author who
specified XB m oWE in v. 3 clearly agreed with this reading, since this
xeZ WoA HL
excludes the goats of Numbers 28–9 from the laws of subordinate jugation, as
they are all purification offerings rather than HL xeZ-type offerings.93
m WoE or XB
Whether or not this author’s interpretation is correct––that is, whether or
not the law in Numbers 28–9 really excludes the 웧purification from subordinate
jugation––was a matter of great controversy in antiquity and will be discussed
in the following section. At present, it suffices to note that, according to
the ancient law––prior to the reinterpretation of Num 15:4–10 to include
non-calendric sacrifices––votive and volitional wholeburnt and wellbeing
offerings were not subject to subordinate jugation with libations and cereal
offerings. It is therefore possible to describe a diachronic shift as follows:
Only calendric wellbeing and wholeburnt quadrupeds require jugation accord-
ing to the table of jugations
All wellbeing and wholeburnt quadrupeds require jugation according to the
table of jugations
It is impossible to ascertain whether there were other jugational rules that
applied to votive or volitional offerings. Presumably there were such rules,
but P does not preserve them. It is also impossible to determine when and
why this transformation took place. However, it appears that the change is
part of a more general pattern of extending the scope of subordinate jugation
in biblical and post-biblical traditions. This pattern is reflected also in the
following section and summarized schematically in the  (Jugation,
rule 10).
In conclusion, there is good reason to believe that Num 15:2b–16 was
composed by an author who had the calendric offerings of Numbers 28–9 in
mind, probably with the intention of systematizing the jugational table
implicit in that text and applying it to private volitional and votive sacrifices
as well. Before the composition of Numbers 15, it seems that an individual’s
volitional and votive sacrifices were not considered subject to subordinate
jugation of this type: rather, they seem to have consisted either of a bare
93
Rendtorff (1967:15) decidedly claims that (a) the passages in Num 28–9 pertaining to
libations are later than those pertaining to grain offerings; and (b) the passages pertaining
to purification offerings are later than those pertaining to wholeburnt offerings. This text-
historical reconstruction gives rise to another possible conjecture as to why male caprines
are not included. If, as Rendtorff claims (1967:15), the references to purification offerings in
Num 28–9 are systematically secondary, then it could be conjectured that Num 15 was written
on the basis of Num 28–9 at a stage when the latter did not yet mention the goat. However, while
significant evidence for (a) is adduced (see nn. 000 and 000), the evidence for (b) is insufficient
(e.g., Rendtorff: 16 n. 4; cf. the similarity between the day of purgation and the other occasions,
noted on pp. 115–16).
11:50:13:06:14 Page 88
Page 89
Jugation 89
(non-jugated) zoeme or of a zoeme jugated with non-animal materials of an

unspecified quantity. Note that in this case, the diachronic development we
have postulated depends on an observation pertaining to zoemics. This
demonstrates the extent to which characterizing the properties or processes in
one part of the grammatical framework occasionally demands knowledge of
another category or categories as well.
3.7.2 An Issue in Numbers 28–9: Does the 웧 Offered as a

Purification Offering Entail Subordinate Jugates?
In terms of the jugational patterns of calendric offerings, the data in Numbers

28–9 are quite straightforward and unambiguous, despite a small number of
scribal errors. However, a simple reading of these verses yields results that are
unlikely from a historical point of view, preventing commentators from
accepting the plain meaning of the verses as they stand.94 Admittedly, the
literal reading laid out here is highly unlikely. It stands to reason that, as
Rendtorff and others observe, the text’s present form is the result of a com-
plex text-history, which led to some combinations that were perhaps not
intended by any of the authors or interpolators.95 However, it is important to
first state what it is that we are rejecting as unlikely.
Stated simply, the list of calendric purification offerings in Numbers 28–29
reflects two models (Figures 12 and 13). Model A (Figure 12) is assigned
for New Moon celebrations (28:15),96 the 15th–21st days of the first month
(28:22–4),97 the Day of the First Fruits (28:30–1),98 the celebration on the first
94
See, for example, Rendtorff 1967:15–16, who argues that the abbreviated references to the
libations in Num 28–9 (e.g., OH e YK
av
d Nv) are interpolations that were inserted unsystematically for
SiW
an unknown reason.
95
See Rendtorff 1967:15–16 but compare Milgrom 1990; on the relation between this text
and Lev 23 (as well as Ezekiel) see Achenbach 2003:604–7 and the bibliography cited there.
96
The pronominal suffix of WodKv Nv must refer back to the RYE
SiW i$om , as LXX correctly under-
stands. Due to the grammatical structure WodKv SiW
Nv H$oe q
EaY½DYMiTmx
d H TLx WoE-LEx (literally “in addition
to the regular offering it should be offered, and its libation”), it is highly improbable that
this pronominal suffix refers to any other noun in the sentence (contra Milgrom, Leviticus
1.845). Even if one accepts the variant OH e YK
av
d Nv (with the Samaritan version), the law would
SiW
still require a libation for the RYE $, and (apparently) no cereal offering (the only difference
i om
being that according to the Samaritan version, no mention is made of the cereal offering
accompanying DYM iTmx
d H TLx WoE).
97
Vv. 23–4a seem parenthetical and may very well be an interpolation. Therefore, the
formula WodKv SiW
Nv H$
oe q
EaY½DYMiTmx
d H TLx oWE LE x must refer to the RYE om , as everywhere else in Num
i$
28–9. This is evident from LXX as well. The variant found in the Sam version, OHYKSNW W$EY,
may (but need not) be interpreted as referring to the other zoemes as well (2B웧, 1S웧, 7s웧,
1 웧).
98
Even if one accepts the MT WoTX NiiW DYM
mvM iT mx
d H TL x Eo DB
xv i rather than the more likely
dLM
HTXNMW (Sam), the fact remains that the cereal offering must be subordinate to DYM iTmx
d H TLx Eo
(perhaps with the possessive suffix referring to DYM iT m ). However, LXX (τ4ν θυσαν α.τν)
d
seems to reflect OT i
mm Ni W, suggesting that the RYE
XvM i$om requires a cereal offering as well.
11:50:13:06:14 Page 89
Page 90
Fig. 12. Model A: male goat with subordinate jugation of wine.
Fig. 13. Model B: male goat without subordinate jugates.
of the seventh month (29:6),99 the tenth of the seventh month (29:11),100 the
16th of the seventh month (29:19),101 and probably the 15th of the seventh
month (29:16).102 In each case, the sacrifices are listed in the following order:
1. Wholeburnt offerings
2. 웧
3. Coterminous offerings (offerings that occur on the same calendric day,
such as the daily morning and evening wholeburnt offering, which
appears in all of the lists)
99
The formula d HTmm NiiW DYM
XvM iT mx
d H TL x EoWv d
HTmm NiiW $oD
XvM eX
ox H TL x Eo DB
xv i is parenthetical. This
dLM
example is most telling: the cereal offerings accompanying the calendric wholeburnt offerings
particular to this occasion (1B웧, 1S웧, 7s웧) are specified in vv. 3–4; the cereal offerings of the
wholeburnt offerings offered on the same day, which is also a New Moon, are mentioned in
v. 6 (dHTmm NiiW $oD
XvM eX
oxH TL x Eo DBxv i ), as are the cereal offering(s) accompanying the daily offering
dLM
(d
HT mm NiiW DYM
XvM iTmx
d H TL x EoWv); but the term OH e YK
av
d Nv, which appears at the end of the list (v. 6),
SiW
can hardly exclude the 웧, as it covers both the zoemes listed before this 웧 (1B웧, 1S웧, 7s웧)
and those listed after it ($oDeX ox
H TL x Eo, DYMiT mx
d H TLx Eo). Note that the monthly 웧 is not offered
on the first of the seventh month. It is probably considered to be replaced by the 웧purification of
this occasion. This is clearly implied by the fact that the verse states $oDeX oxH TLx Eo DBxvdLMi but not
$oDeX
ox H TAU xx
d X* DB xvd
LMi . We know that $oD eXoxH TAU xx
d X* is not a part of $oD eXoxH TL x Eo from
vv. 14–15 of the same chapter, where HN m$
mdoxH Y$aoD
vX
mvL Wo$oDvX
mvdB $oD eXo TLx Eo TAZo follows the whole-
burnt offerings but precedes the 웧.
100
Note that the suffix of d HT
mm Ni is singular, referring to DYM
XvM iTmx
d H TLx Eo only, whereas OH e YK
av
d SiN
is plural, referring both to DYM iTmx
d H TLx Eo and to TAU m
dxX RYEi$ ov ; the libations and cereal offerings
of the wholeburnt offerings are mentioned in the previous verses.
101
Here, too, the Sam version reads HYKSNW.
102
MT reads d Hdm
Kv Nv, Sam HYKSNW, with no significant difference in meaning. These versions
SiW
are logically unlikely: they would imply that the wholeburnt offerings on this day require
neither libations nor cereal offerings. Therefore the possessive suffix must have been plural, if it
were to include the wholeburnt offerings (the cereal offerings accompanying these wholeburnt
offerings already being listed in vv. 14–15, but not their libations). Therefore, OH e YK
av
d SiN must be
the correct reading here, and as soon as one reads OH e YK
av
d SiN, the grammar automatically includes
the RYE i$om . The Sam, perhaps aware of this but unwilling to have the RYE i$om included among the
offerings referred to in the suffix of OH e YK
av
d SiN, inserted OH e YK
av
d Nv at the end of v. 15.
SiW
11:50:13:06:14 Page 90
Page 91
Jugation 91
In all of these cases, the cereal offering accompanying the wholeburnt

zoemes is listed before the 웧, as are the sacrifices introduced with DB xv
d
LMi or
LEx (such as the daily wholeburnt offering), clearly implying that this jugation
does not apply to the 웧. However, in all of these cases, the libations are
mentioned after the 웧 (and in fact after the additional coterminous
sacrifices). This implies that the 웧 also entails a libation (but no cereal
offering). If one were to object, arguing (somewhat forcedly) that OH e YK
a
dvSiW
Nv
refers to the wholeburnt offerings only (meaning something like “and their
libations, if they require any,” as, perhaps, at Num 6:15), this is immediately
refuted by v. 15, where “and its libation” can only refer grammatically to the
웧.103 This argument is further supported by the frequent juxtaposition
of d
HTmm NiiW (singular) with OH
XvM e YK
av
d Nv (plural) in 29:6, 11, 19:104 it is perfectly
SiW
logical that the author specifies OH e YK
av
d Nv in these cases, since these libations
SiW
pertain both to the daily wholeburnt offering and to the 웧purification, whereas
only the daily offering entails a grain offering.105 The scheme implied in these
verses is therefore as follows:
{calendric-wholeburnt sacrifices with their cereal offerings, 웧,
coterminous wholeburnt offering with its cereal offering}
and their libations––that is, the libations of all of the offerings listed in braces.
Model B (Figure 13) is assigned for the 17th–22nd days of the seventh
month, where each list is invariably followed by the formula TL x Eo DBxv
dLM
i
i ) DYM H (the possessive suffixes of d
Hd
d m
KvSiW
Nv d
HT mm
XvM
Ni (W iTmx
d HTmm Ni and d
XvM Hd
m SiN are
Kv
singular, clearly referring to DYM
iTmx
d H TLx Eo only and excluding 웧).106
Historically speaking, it is unlikely that P made this hairsplitting dis-
tinction between the calendric purification offerings of the 17th–22nd of
the seventh month and all other calendric purification offerings. Note
that the distribution of models A and B follows (with one exception) the
progression of the text––with the dividing line between Num 29:19 and 20
(that is, in the middle of a festival, between its second and third days)––
rather than following any discernible logic of the festivals’ importance or
significance. This suggests that the distinction may be a byproduct of the
103
Contra Milgrom, Leviticus 1.845, who argues that it could refer to DYM iTmx
d H. If this were the
case, one would expect the word order H$ oe m
EaY oWdKv
SiW
Nv DYMiT
mx
d H TLx Eo LE x.
104
Note that vv. 28:15, 24 do not imply that DYM iTmx
d H TLx WoE is exempt from subordinate
jugation. In these verses, DYM iT mx
d H TLx WoE certainly means “the daily HL m WoE including all of its
entailments.” Conversely, the author refers to HT mm NiiW DYM
XvM iTmx
d H TLx WoE in some verses (e.g., 28:31;
29:6), implying that DYM iTmx
d H TL x WoE here denotes the bare zoemes (without their jugations).
105
See also 28:31, which implies the same. Verse 29:16 is probably an error; see n. 000.
106
The word H m YK
eS Ni in v. 31 may be a scribal error or a stylistic variant (recurring in Sam)
mvW
and in either case it has no bearing on the problem. Note that DYM iT mx
d H TLx Eo naturally has more
than one libation, inasmuch as it is offered twice daily; these can be referred to collectively as
Hd
d m
KvSiN, or in the plural H
m YKmvN.
eS
11:50:13:06:14 Page 91
Page 92
text history of Numbers 28–9, not the result of a development in the world of
sacrifice.107
There is, of course, one seemingly strong argument against the recon-
struction of Model A: one does not know what libation would accompany a
웧 offered as a purification offering, since even in the final form of Num
15:1–16, no libation is assigned to this zoeme when offered as a purification
offering. Recall that Num 15:3 refers only to votive, volitional, and calendric
wholeburnt and wellbeing offerings. However, this is not as great a problem as
might seem at first glance. First, one does not know what libations accompany
either the B웧 mentioned in Num 15:24 (since it is non-calendric) or the
Nazirite’s offering (Num 6:15b), and yet it is undeniable that they require
subordinate jugation. Second, if the 웧 offered as a purification offering
should require a libation, one might conjecture that this would be identical to
that required for a 웧 offered as a wholeburnt or wellbeing offering.108
It is not surprising that none of the documented traditions outside
Numbers 28–9 reflect this explicit but unlikely jugational pattern. The Temple
Scroll, on the one hand (and, at least from a formal point of view, the Aramaic
Levi Document)109 requires that every purification offering be accompanied
by both a libation and a cereal offering. So central is this to the authors of the
Scroll that they note it no less than twelve times in the surviving passages.110
The rabbinic tradition, on the other hand, is explicit in denying all
calendric purification offerings accompanying libation and cereal offerings.111
As we have seen in the previous section (3.7.1), this interpretation is
107
Note, however, that the dividing lines between Models A and B do not converge with
further text historical divisions that have been identified in the chapter on linguistic grounds,
suggesting the text-history of these chapters may be even more complex than claimed in the
past––see Knohl 1987; Milgrom 1990:244, 326; Achenbach 2003:604–7.
108
A forced reading of Num 15:3 would yield precisely this result. This would involve placing
a major pause before HL m Eo and reading the verse as if it referred to (1) votive or volitional
wholeburnt and wellbeing offerings and (2) all calendric offerings. The fact that only x XYRaH$adA
oi
HWHYL x x i offerings are included would not refute this “midrash,” since the purification
XoXYN
offerings in Num 28–9 are also sometimes considered HWHYL x xXXo YN
i x
XYRa H$aoi , though
dA
unsystematically so (see Num 28:24; 29:6). According to the more natural reading, correctly
noted by the Masoretes, the verse refers to the following categories: wholeburnt and wellbeing
offerings offered as either (1) votive or volitional offerings or (2) calendric offerings.
109
See 9:1 QYLS YHWDWXLB ABRT OAW (“and if the suet alone is offered”) and 9:16 (π
στ#ατο, “upon the suet”), which could theoretically refer to all offerings other than the whole-
burnt offering. No qualification is made in ADL indicating that the table of jugations that
follows applies only to wellbeing offerings and not to purification or reparation offerings.
110
See in particular 18:4–6; 25:5–6, 12–15.
111
In fact, rabbinic tradition exempts almost all purification offerings, including the
Nazirite’s offering, from subordinate jugation. The only exception is the purification offering of
a person purified from scale disease. See b Sotah 15a. According to Marx (1994:155–6), Jubilees
implies that the purification offerings are exempt from subordinate jugates; the evidence,
however, is inconclusive (see 6:3; 7:5, which are ambiguous as to which offerings the jugates
accompany; and 21:7–9, which is irrelevant).
11:50:13:06:14 Page 92
Page 93
Jugation 93
suggested even in the core of the “table of jugations” (Num 15:4–10), as well
as in the final form of Num 15:1–16.
A similar view appears in Ezekiel 40–8. This vision, which is concerned
primarily with calendric offerings and never mentions non-calendric purifi-
cation offerings, provides detailed rules of jugation for calendric and non-
calendric wholeburnt offerings (Ezek 46:1–15)112 but is silent on the jugation
of calendric purification offerings. The laws of the seven-day festival of the
first month (45:23–5) offer a telling example: the calendric offerings for
this festival include {7B웧wholeburnt, 7S웧wholeburnt, 웧purification} per day, but only
the bulls and the rams entail subordinate jugation.113 Note, however, that
Ezekiel’s vision also differs from the rabbinic tradition, as well as from all
other traditions mentioned here, inasmuch as it never speaks of calendric
libations––with the exception of 45:17––and apparently does not require the
subordinate jugation of wine to zoemes.
It is not difficult to imagine why rabbinic tradition rejects the possibility
that some calendric purification offerings must entail a libation, despite
several verses which, we have just seen, appear to state the opposite. They
simply follow the author of Num 15:1–16 in assuming that none of the
calendric purification offerings require subordinate jugation.114
The reasons for the Temple Scroll’s rejection of the literal meaning of
Numbers 28–9 and its replacement of Model B with the standard triadic
scheme {zoemeA, grainB, libationB} can be understood better after another
grammatical property is described, namely “hierarchics.” We shall return to
this problem below (§6.3), as it pertains to the interrelation between several
operative categories in the grammar.
3.8 JUGATION INDUCED BY JUGATION
In some cases, the form of a jugate––for example, whether it is raw or

cooked––and its status within a chain––for example, standalone or subordin-
ate––may affect the structure of its own subordinate jugates.
112
Calendric wellbeing offerings do not exist for Ezekiel, with the possible exclusion of 46:2
(which can hardly be termed calendric) and the abstruse instruction in 45:17. Non-calendric
wellbeing offerings are only briefly mentioned, as in 43:27, and (with specific reference to the
AY$o N) in 46:12, on which see n. 000.
113
The B웧 mentioned in v. 22 is not part of this passage: it is to be offered on the 14th day,
and therefore the general ruling in vv. 24–5 does not apply to it.
114
Other considerations may have played a role here as well. For example, the scheme that
is found in the final form of Num 28–9 probably seemed as illogical to the tannaim as it
does to modern readers. For other considerations, see b Men. 91a, according to which the only
purification offering that requires subordinate jugation is that of the person purified from scale
disease; see Milgrom, Leviticus 1.845–6 and below, p. 000.
11:50:13:06:14 Page 93
Page 94
For example, if Leviticus 2 in its present form applies only to independent

grainA freewill offerings (see §3.2.2), then there exists a distinction between
raw grainA offerings on the one hand and grainB offerings (which are always
raw) and cooked grainA on the other hand. The former entail subordinate
oil and frankincense, the latter only oil.115 In a way, it is as if the cooking
“replaces” the frankincense––one is expected to spend energy (cooking) or
money (frankincense), so to speak, to make the offering pleasing.116
In Tannaitic tradition, on the other hand, a different distinction is made,
this time between A-level cereal offerings and B-level cereal offerings: the
former entail subordinate jugates of oil and frankincense, whereas the latter
entail only subordinate oil. In other words, the existence of frankincense
depends on the jugational position of the cereal offering (A- versus B-level).
Thus, a ruling in m Men. 5:3 decrees that a cereal offering offered as an
A-level freewill offering (cerealA) always entails subordinate jugation of the
type {oilB, frankincenseB}. According to this mishna, this jugational pattern
applies to grainA regardless of the form of the grain offering (raw or
cooked).117
By contrast, a cereal offering that is offered as a B-level offering (that is, as
a subordinate jugate to a zoemeA), but that is otherwise identical to the
former, always entails subordinate jugation of the type oilC, but never
entails frankincenseC. Note that in this, Tannaitic law seems to disagree with
several late Second Temple texts that imply a third scheme, requiring that
frankincense be added to grainB offerings as well.118
Stated more generally, the ruling of the mishna in Menah.ot may be
formulated as follows:
115
See Lev 2:1–10 and Num 15:1–16. See also TS 20:9–10, which distinguishes between
VSN HME BRQ R$A HXNM LWK on the one hand and HBRX WA HNWBL HYLE BRQ R$[A] HXNM
on the other hand. This seems to imply that grainB offerings do not require frankincense C, as
in Mishnaic law, but neither does every grainA offering, only those that are offered raw. This
accords with a literal reading of Lev 2.
116
See p. 000.
117
Note that this rule applies even to the “sheaf of elevation” (Lev 23:10–11, RM m , termed
e EoH
i NTH RMEo in v. 15). Tannaitic tradition considers the sheaf of elevation a type of HX
HPW Ni . Note
mvM
that RMe EoH
m TX Ni here must refer to the A-level barley “sheaf” itself, not to the grain offering
xvM
referred to in v. 13, which is a subordinate jugate of the zoeme. By contrast, this law does not
pertain to A-level cereal offerings brought on account of particular sins, where P explicitly states
that no jugation applies (Lev 5:11; see also Num 5:15).
118
See ADL 9:15–16, which distinguishes between grainB offerings, which are referred to
as π στ#ατο (“upon––i.e., jugated to––the RDeP d”) on the one hand, and grainA offerings,
e
referred to in the clause 8<ν> α9ν προσαγάγ: µνον (“alone,” as an independent offering) on
the other hand. Note that though ADL lists the quantities of frankincense per zoeme, it clearly
implies that the frankincense is physically jugated to the grain offering both in the case of a
grainA offering and in case of a grainB offering. Similarly, Gen. Apoc. (1Q20 10:15–16) mentions
semolina that entails oil and frankincense (AXNML ANWBL OE X$MB ALYP APY$N TLWS) in
a context that implies that the grain offering in question is subordinate to a zoeme. See also
Schiffman 2005:197–8 and Himmelfarb 2004 concerning this evidence and further evidence
from Jubilees.
11:50:13:06:14 Page 94
Page 95
Jugation 95
Rule 1: Subordinate jugation of frankincense does not apply to cereal offerings
that are themselves subordinate jugates of a zoeme.
The “grammatical” property of subordinate jugation is determined by the
location of the cereal offering within a jugational scheme.119
In theory, this could be explained on the following grounds: when offered
with meat, there is no need for frankincense with a grain offering, since
the meat itself is sweet-smelling. However, the same pattern is repeated in
one rabbinic tradition, in a suspiciously similar manner, with regard to
the jugation of salt to wine libations. The very fact that what is at stake is the
salting of wine suggests that we are dealing here with implicit jugational
patterning, a development within the grammar of sacrifice that departs from
the usual use of condiments in a meal. Recalling the fast-food analogy, it is
natural to salt one’s burger or one’s fries, or both; but the beverage is not
normally salted. This example will therefore prove useful in demonstrating
how Σ is conducive to extrapolation from the practical to the imaginary.
We have already mentioned in passing (§3.2.2) the requirement to add salt
to certain offerings. As we shall see in the following section (§3.9.2), some
controversy existed in antiquity concerning the scope of the requirement to
salt offerings. However, even if one posits that all offerings must be salted, the
salting of wine need not present a special problem: since the wine is offered
praxemically alongside the meat and the cereal offering, which are salted, one
might say that the wine is somehow salted as well.120 In rabbinic tradition,
however, wine is not automatically salted with the meat, since wine may be
offered independently (wineA). Moreover, even when offered as a B-level
jugate subordinate to a zoeme, wineB may be praxemically separate, as it may
be offered at a time and place that are different from the flesh and grain com-
ponents.121 Thus, a problem of salting arises: should wineA and wineB be salted?
One rabbinic tradition decrees that the subordinate jugation of salt to wine
depends on the location of the libation within a larger jugational scheme.
According to this tradition, wineA requires saltB, but wineB (a libation offered
as a subordinate jugate of a zoeme) does not require saltC. Hence the
following rule can be formulated (see Jugation rule 15a.iii in the ):
Rule 2: Subordinate jugation of salt does not apply to wine libations that are
themselves subordinate jugates of a zoeme.122
As with Rule 1, it is clear that in the rabbinic system jugational patterning is
at play, and not some other kind of logic. If one were to argue that wineB is
119
It is unclear whether this distinction in the mishna reflects actual practice. The schematic
formulation of this mishna in a tetralemma suggests that some of the distinctions in it might be
artificial.
120 121
Such a procedure is reflected in ADL 8:6. See also Sir 50:15 and §3.6.2.
122
See n. 000.
11:50:13:06:14 Page 95
Page 96
exempt from the jugation of saltC because it is somehow conceived of as salted

along with the flesh and cereal offering that it accompanies, even if it is not
in physical contact with them, this would be refuted by the fact that in the
rabbinic system, grainB offerings still require salt, even though the flesh that
they accompany is also salted.
Rules 1 and 2 are strikingly similar, perhaps similar enough to justify the
formulation of a third rule:
Rule 3: If jugate a is located on Level A, subordinate jugational pattern b will
be found on Level B; if jugate a is located on Level B, jugational pattern b will be
modified to b*.
Rule 3 holds for the values a = semolina, b = oil and frankincense, b* = oil
(but not frankincense); as well as for a = wine, b = salt, b*= no salt.
Considering the fact that the salting of wine is mentioned only in texts
composed after actual sacrificial procedures in Israel were a matter of the
past, it is quite possible that such a custom was never carried out in practice.
However, it demonstrates the potential of jugational patterning inherent in Σ,
even after the sacrificial procedures ceased to be performed.
3.9 GENERATIVE JUGATION
Thus far, we have only examined “real” jugational chains. By “real,” I do not
mean to imply that these jugations ever had an ontological existence outside
of the texts in which they appear, though it stands to reason that some of
them are not purely imaginary.123 Rather, they are “real” inasmuch as they are
laid out in a straightforward manner in the texts describing them. Resorting
to the analogy of language, these jugations are like sentences that were at
least expressed by speakers of a language: not only are they grammatical, but
they have also actually found expression at a certain time and place. “Real”
jugations, like “real” sentences, offer direct evidence of usage, but they only
teach us a limited amount about the rules that govern licit formations.
3.9.1 Unreal Jugational Patterns
We will now consider some jugational patterns that are “unreal,” inasmuch as
they most certainly were not verbally conceived in any text but are licit
according to the grammar of sacrifice. These are analogous to grammatical
sentences that have never been expressed, either due to pragmatic reasons or
123
See §1.4.3.
11:50:13:06:14 Page 96
Page 97
Jugation 97
because they have never come up in conversation. Consider sentence A, of the

latter class:
A. Xavier Klugman is sitting under this baobab tree.
This sentence has likely never occurred to anyone before, not because of any
practical difficulty involved in its production, but purely because the topic of
Xavier Klugman under a baobab has probably never arisen.
Among the more informative cases are sentences that cannot be expressed,
either because they are of unlimited length (sentence B), or because they are
too complicated for most speakers to follow (sentence C):
B. Here is the buffalo that sees the god who sees the buffalo that sees the god. . .
C. Do you know what the reason he knows what the color of the buffalo whose
horn this is is is?
Sentence B is of unlimited length: it is grammatical but literally ineffable
(one is reminded of Patañjali’s mention of the mahāsattras, sacrificial
sequences that would last a thousand years). Sentence B is grammatical but
pragmatically unlikely, since it is too Dr Seussesque to follow.
The distinction between “real” and “unreal” jugational patterns thus
resembles both the distinction between parole and langue as conceived by
Structuralists and the distinction between performance and competence as
conceived by Generativists. The “real” jugational patterns are the sum total
of all sequences that have ever been expressed (either in text or in practice)
in Israelite sacrifice; the “unreal” ones exist only potentially, inasmuch as
they are grammatical, licit patterns.
Due to the nature of our texts, we are of course far removed from the world
of sacrifice, both because some texts that serve as the basis of the present
grammar are utopian to begin with and because even these texts only allow
for, but do not envision, the following examples of jugational patterns.
Nonetheless, just as the above outlandish sentences’ grammaticality is an
important aid in understanding the English language, the identification of
sacrificial patterns is essential for an understanding of Σ. Let us now turn to
one such example.
3.9.2 Infinitely Long Jugation: To Be Taken with a Grain of Salt
Leviticus 2:13 reads, XL xM

e TYBi$
d voT x AL
ovW XLmMvTi XL
d xeM
dd
xBmVTvm
XvM
Ni IBd m-LK
x RvQ mWv
XL
xeM BYRiQ
vTx
d m
VvB
Nd
m R
vQ
m - LK
m
d LE
x m
V T
e m
XvM
N i LE
x a
M m
V YH
eoL f
A TYR
iB
d
v (“You shall salt every
cereal offering; you shall not omit from your cereal offering the salt of your
covenant with your God; with all your offerings you must offer salt”).124
124
On the translation of this verse, see immediately following. Concerning the “covenant,”
see Schwartz 2004:210.
11:50:13:06:14 Page 97
Page 98
This verse introduces certain elements that must be addressed in terms

of jugation. Note first that many of the jugational diagrams may now be
modified to accommodate the biblical reference to salt. For example, Figure 5
(p. 000) could be presented more fully as follows (according to reading (b)
and (c) in Figure 14).
Fig. 14. Basic jugational pattern including salt.
Although the salt and oil are both C-level jugates in this sacrificial complex,
the wording of the biblical text implies that their respective links to the cereal
offering are not equally strong: the salt, it appears, is less an integral element
and more an additional requirement. This classification is evident from a
physical point of view, inasmuch as the offering is not thoroughly mixed with
salt (as it is with oil)125 but only sprinkled with it.126 In fact, according to
several rabbinic traditions, it appears that only the portion of cereal offering
burned on the altar was salted,127 implying that salt is not an ordinary jugate,
but rather an “additive condiment.”
Several difficulties present themselves in Leviticus 2:13: first of all, it is
unclear how many discrete instructions the verse contains. Grammatically,
it consists of two (13a, 13b), possibly three (13aα, 13aβ, 13b) discrete
125
Hence the appositive formulations such as IM ee$
dox miWLB
B HL d TL
v eS o HX Ni , Num 15 et passim;
mvM
see Lev 2:5, 7 (IM
ee$dod
x
B TLeSo ), but see also Lev 2:4 (IM
em$dod
x
B OYX
iu$oMv ), which implies a less intimate
relation between the oil and the semolina.
126
Haran (1978:242 n. 23) has demonstrated this on linguistic grounds: P uses the pi‘el
conjugation to denote the mixing with salt ($oDeQ o RWoHUm XLmM
d uM v , Exod 30:35) but the qal to
denote sprinkling with salt (XL mMvTi XL
d xeM
dd
xBmVTvm
XvM
Ni IB
d m-LK
x RvQ mWv, Lev 2:13).
127
See, for example, b Sot 14b HIM 5:11, HMQ 13:12, according to which it is usually only
the HRmdm Zx (“token”) of a cereal offering that is salted, whereas a cereal offering of a priest is
KvA
salted in its entirety, since it is offered in its entirety (in accordance with Lev 6:16). This would
mean, for example, that in the case of a OYM im $o, only the suet was salted.
Lv
11:50:13:06:14 Page 98
Page 99
Jugation 99
sentences.128 However, these sentences may all express a single ruling,

formulated in a pleonastic fashion found elsewhere in P and often in H.129
The solution to this first problem is linked to another problem, the scope of
m
VvB
Nd m-LK
m RvQ m in 13b. If taken literally as “every offering,” then the verse contains
d
at least two discrete commands: salt every cereal offering (13a) and salt every
offering (13b). This redundancy may be resolved in part by a hypothesis of
diachronic accretion: 13b may be a later interpolation, attempting to broaden
the scope of 13a to include other offerings as well.130
However, since this diachronic reconstruction does not solve the problem
but only transfers the responsibility from the author to the interpolator, we
will consider the verse in its final form, bearing in mind that this may also
have been the original form. As it stands, the verse allows for at least three
possible readings of the phrase XL xMe BYRiQ
vTxm
d VvB
Nd m-LK
m RvQ m LE
d x (“with all your
offerings you must offer salt”):
(a) a contextually limited reading: only cereal offerings mentioned in the
section to which this law is appended (Lev 2––so only freewill cerealA
offerings) require salting;131
(b) a less contextually restricted reading: all cereal offerings require salting
(including statutory and freewill cerealA and cerealB offerings);
(c) a more literal reading: all offerings, including all meat offerings as well
as all cereal offerings, require salting.
Since the law of Lev 2:13 is appended to a passage pertaining to cereal
offerings, it is most likely, though not provable, that the law in 13b was
intended to refer only to cereal offerings, in contrast to a more literal sense
of m
VvB
Nd m-LK
m RvQ m.
d
The earliest traditions reflect a reading somewhere between (b) and (c).
From Ezekiel (a text that may or may not reflect awareness of this priestly
injunction), one may deduce that the salting of sacrificial meat (43:24) was
128
The former is probably more correct, with the phrase m VYH
eoLAf TYRiB
d XL
v xMe TYBiv
d $oT
x ALov
W
m
VTemXvMNi LE
xMa a subordinate clause, the force being “lest you omit from your cereal offering the
salt of your covenant with your God.” See GKC §§197q, 109g; cf. J.–M. §168i.
129
Within P, see Lev 6:5aα, 6 (see also 2bβ), where the negative injunction HB iev
KT o (twice)
i AL
contains no new information not found in WodB DQ i i
xWT dxXd
a
BvM
Zd
ixH $oA
avW, WodB-DQ
xWTdxXd
a
BvM
Zd
ixH-LE
x $oA
aHmvW,
x
Xda
BvM
Zd
ix H-LE
x DQ xiWT
d DYM
iT a ; and within H, see Lev 11:42b–43 (see also 44b). As Milgrom
m $oA
d
(Leviticus 1.683–4) has demonstrated, all of these commands refer to a single act, the ingestion
of prohibited flesh. See also Paran 1989:49–73, 163–4.
130
Note, however, a stylistic device used in this verse, perhaps implying single authorship.:
m
VTvmXvMNi IB
d m (13aα) first appears as a construct chain, then split into m
m RvQ VT
em Ni (13aβ) and
XvM
m
VvB
Nd m (13b). The precise location of vv. 11–13 in Chapter 2 suggests that the law of salting
m RvQ
the cereal offerings was not intended to include the “cereal offering of first fruits” (14–16), since
vv. 14–16 may have been appended only after the concluding rules of 2:11–13 were included in
the text.
131 idKiB TX
More precisely, with the exception of OYRiW d xvM Ni (see previous note).
11:50:13:06:14 Page 99
Page 100
known in some priestly circles. However, the salting of meat is mentioned

only once in Ezekiel, in the context of the wholeburnt offering. A careful
reading of ADL 9:6–9, which specifies the precise amounts of salt for each
zoeme, may similarly lead to the conclusion that only wholeburnt offerings
require salt,132 a notion that Josephus also implies.133 The requirement to salt
the showbread, found already in the LXX version of Lev 24:7, probably
reflects a rejection of reading (a) in favor of reading (b) or (c), and an
understanding that the showbread is a cereal offering.134
Rabbinic traditions are even more inclusive in their reading of m VvB
Nd m-LK
m RvQ m
d
in Lev 2:13b,135 offering an extremely literal reading:
(d) an ultra-literal reading: all offerings––not only meat offerings and
cereal offerings, but also such offerings as wine and oil––require salting.
In P, there is hardly any practical difference between readings (c) and (d).
Sacrificial materials that are neither cereal offerings nor meat offerings
(such as wine, oil, or frankincense) are not discussed as independent A-level
offerings in P, and when wine, oil, or frankincense are offered as B-level
jugates, they are offered along with salt anyway, since the meat itself or the
cereal offering is salted (this is particularly true if salt is sprinkled on the top
of the whole offering).
However, later traditions contain such structures for which readings
(c) and (d) play out differently. As noted above, the list of offerable
materials was gradually expanded in the ancient Israelite sacrificial system
(following the initial restriction of P). Thus, several materials that are not
found in P as independent offerings re-entered the sacrificial repertoire of
post-P sacrificial systems as acceptable A-level offerings, so that wineA,136
132
There are two main reasons for this. (1) ADL consistently lists the amount of wood
required for the suet of a zoeme alongside the (much larger) amount required if the animal as a
whole is offered. By contrast, ADL lists only the quantities of salt required for whole zoemes and
is silent about the quantities of salt required when only the suet of these zoemes is offered. (2)
ADL stresses that one may use the rest of the salt for the salting of the hide (see Lev 7:8).
Furthermore, quantities of salt are noted only in the context of male zoemes (this last fact is less
decisive, since it is true of other subordinate jugates in ADL as well).
133
See Josephus, Ant. Jud. 3.227 (also noted by Milgrom, Leviticus 1.192). Reading Jub. 21:11
in its immediate context suggests that not only wholeburnt offerings are implied, but at least
other meat offerings as well (Marx 1994:155–6)––but the data here are uncertain.
134
This understanding is reflected also in m Men 5:3.
135
See also Mark 9:49, which according to some manuscripts reads κα πα̃σα θυσα α;λ
α;λισθ&σεται (see GNT 158). For a discussion of this text and the late Second Temple sources,
see Schiffman 2005:194 and n. 68.
136
See TS 20:14 and the opinion of Shmuel in b Zeb 91b, according to which wine can be
offered on the altar fire (OY$YAH YBG LE WPLZMW AYBM IYY BDNTMH); but compare the opinion
of R. Akiba in t Men 12,10. The offering of wineA may hark back to pre-P traditions, as in the
case of frankincense (see below). Concerning wineA in P, see perhaps Num 29:39, and compare
Ezek 45:17.
11:50:13:06:14 Page 100

Page 101
Jugation 101
oilA,137 frankincenseA,138 woodA,139 and even saltA140 offerings appear to be

grammatical according to later traditions.141
This led to the possibility of a much wider application of the phrase “all
your offerings,” which now included the salting of materials that had not
been conceived of as offerings in P. Thus, in rabbinic literature, the law of
Lev 2:13b came to be regarded much more inclusively as a generic law
covering diverse materials, and in fact rabbinic sources refer explicitly to the
salting of oilA, frankincenseA, woodA, and wineA.142 Let us therefore depart for
the time being from the original intention of P and examine the development
of Σ within the rabbinic traditions pertaining to the subordinate jugation of
salt. From these rabbinic texts, it appears that an offering can be exempt from
salt as a subordinate jugate only under the following conditions: (1) it is not
considered fit for the altar fire in the first place; (2) it is exempted by a specific
verse; or (3) it is exempted by an explicit, otherwise unrecorded divine
decree.143 This implies that, in this (obviously fantastic) system, wherever
137
See TS 21:14–16. According to the opinion of R. Tarfon recorded in t Men 12:10 and b
Zeb 91b, oilA is offered on the fire of the bronze altar. The log of oil supplied by a person purified
from scale disease (Lev 14:10, 21) is irrelevant here because none of it is offered on the altar.
138
See t Men 12:15, Sifra (Weiss 12a), and see Milgrom 1979 (compare also Jub 3:27). The
offering of frankincenseA probably also harks back to pre-P traditions, and some forms of
independent incense burning are found in P. See Haran 1978:230–45. It is also likely that the
reference to frankincense in an Elephantine text, Cowley 33:11 (A4.10:11, Porten and Yardeni
1986:78) pertains to frankincenseA, as the frankincense precedes the grain offering in what
appears to be a list of independent items.
139
This is perhaps evidenced in Nehemiah (10:35, 13:31, OYC iaEHm IBd u) and Jubilees
x RvQ
(21:12–13), and certainly in ADL (8:7, note the use of BYRQHM) and TS (23:9–24:16,
OYN$ OYCEH TA WBYRQY “they shall offer the two pieces of wood”). The phrase HLWEL OYCE is
found in 4Q365 23, l. 5 (DJD XIII, 291) and in 4Q409 (4Q Liturgical Work A, context missing;
see DJD XXIX 64–5); and see t Zeb 1:11 (OYCEW OYRBYAW OYBLX RUQH “the turning-to-smoke
of suet, body parts, and wood”). Note that Maimonides appears to reject the original denotation
of OYCEH IBRQ, arguing that it actually implies a zoemic offering (HKM 6:9).
140
See t Zeb 10:4. See also the summary in HTM 9:11.
141
The examples cited here are the clearest cases. Eberhart, however, rightly notes that the
mere use of the term IB d m does not in itself imply that an object is, strictly speaking, sacrificial
m RvQ
(Eberhart 2011:24; and see Num 31:50).
142
See Sifra (Weiss) 12a–b, b Men 20a, Maimonides HIM 5:11, and the commentaries.
Surprisingly, Rashi on Lev 2:13 (Chavel 1982:326) does not reflect the maximalist rabbinic
reading, but simply states ILWK OY$DQH LK YRWMYAW FWEW HMHB TLWE LE, “upon the whole-
burnt offerings of quadrupeds, birds, and the suet of all qodashim,” seemingly omitting such
offerings as frankincenseA and wineA, and reflecting reading (b).
143
For example, wineB is exempt from the subordinate jugate salt since it is not to be offered
on the upper surface of the altar in the first place (see n. 000); blood is exempt due to a special
midrash halacha (VMD LEM ALW –– VTXNM LEM, “ ‘from upon your mincha’ [Lev 2:13]––not
‘from upon your blood’,” b Men 20a following Sifra [Weiss 12a]); and woodA is exempt from saltB
only due to an (otherwise unrecorded) explicit divine exemption. See HIM 5:11 and compare
HMQ 16:14 with the commentaries. Note that a different logic for exempting both salt and
wood from subordinate jugation of saltB is rejected in the Sifra. Note, too, that a special scrip-
tural maneuver is required for exempting blood from subordinate jugation of salt, suggesting
that unlike wineB it is in some sense “fit for the fires”––i.e., rabbinic tradition retains a faint echo
of the conception that blood ought to have been offered atop the altar. See on this Meshel 2013b.
11:50:13:06:14 Page 101

Page 102
Fig. 15. Jugation of saltB to saltA .
Fig. 16. Never-ending jugation of salt.
an ad hoc divine exemption is absent, anything that is considered a IB d m in

m RvQ
theory requires a subordinate jugate of salt.
There is one possibility that the rabbis did not entertain, though a rigorous
application of the logic of sacrifice would require taking it into account: since
saltA is a grammatical offering in the rabbinic system––as noted above,
the rabbis viewed salt as a sacrificial material in its own right and not as an
additive144 ––and since no special divine (scriptural or oral) decree exists for
exempting saltA from saltB, it appears that the jugation of saltB to saltA is not
only possible but necessary.
This in itself causes no serious rupture in the system, as one can imagine
an offering of the form seen in Figure 15. However, in the rabbinic system,
subordinate B-level jugates also require salting (except for wineB, which is
exempt).145 Now, since there is no divine decree exempting saltB from the
subordinate jugation of saltC, it appears that––if the generative grammar
of ritual is taken to its extreme limits––a sacrifice of salt would be a never-
ending enterprise of the form (saltA (saltB (saltC ( . . . as in Figure 16. Such a
chain would be created according to the following law: every jugate entails an
immediate subordinate jugate of salt.
144
Moreover, the offering of salt is referred to as BYRQT (“you shall offer,” Lev 2:13). This
appears to be a crucial point; see the opinion in b Zeb 91b that the biblical verse BYRiQ vTx IYiYxWv
d
(“and wine you shall offer,” Num 15:10) is evidence that wine is an offering–– IB d m ––and
m RvQ
therefore ought to be spattered on the fire. Note that the conception that anything offered on
the fire must be salted is based on the word of the law, XLM BYRQT VNBRQ LK LE (“on every
offering of yours, you shall offer salt,” Lev 2:13).
145
In the rabbinic texts, a distinction is made between wine offered as a B-level jugate of a
zoeme on the one hand and wineA offered as an independent offering on the other. Whereas
the former, the rabbis decreed, is exempt from subordinate jugation, the latter requires a
subordinate jugate of saltB (contrast Maimonides HIM 5:11 with HMQ 16:14). The formal
reason for this is that although neither is offered on the altar fire, wineA is considered “fit for the
[altar] fires” OY$AL YWAR ––it is only due to certain external restrictions that it is not poured
on the fire (to refrain from extinguishing the altar fire), whereas subordinate wineB is not
considered “fit for the fires” in the first place. See b Men 21a and HMQ 2:1; 16:14, with the
commentaries.
11:50:13:06:14 Page 102

Page 103
Jugation 103
In fact, having established this, it is now possible to state that every

sacrificial material in the rabbinic system––not necessarily a saltA offering––
should entail an endless chain of subordinate jugates as soon as it requires
one subordinate jugate of salt.146 This situation would create a serious rupture
in sacrificial procedures, since it would essentially prevent any sacrificial rite
from being completed.
This loophole belies either a loose end in the rules or a rupture, marginal
but potentially fatal, in the sacrificial system. Of course this loophole is the
result of a mental game, and it still remains to be seen what it can teach us
about the grammar––not to mention the meaning––of sacrifice (see
§6.2.5.2). However, it should be remembered that in order for a generative
grammar to be comprehensive, it must allow for the generation of every
conceivably possible grammatical form and prevent the generation of all
ungrammatical forms. Therefore, the subordinate jugation of salt to any
offering is a rule that must be part of this grammar, even though––like many
rules in the grammars of natural languages––it allows for the generation of
infinitely long (and hence pragmatically impossible) sequences.
146
Furthermore, it is possible to return to the Priestly text and find that the same situation
may be found even in Leviticus, if Lev 2:13b is understood according to reading (d) above. Since
saltB is, formally, an offering (note the wording XL xM
e BYRiQ
vTx ), there should be no reason to
d
exempt the subordinate jugate saltB from its own subordinate jugate saltC. The reason I have
chosen to present this rupture in the system from the rabbinic point of view, though it may be
found in the priestly system as well, is that in P it would depend on a forced, ultra-literal reading
of Lev 2:13b, whereas in the rabbinic texts cited here the paradox can hardly be suppressed: it
appears to be inherent in the system.
11:50:13:06:14 Page 103

Page 104
Hierarchics
4.1 INTRODUCTION
In , we mentioned another kind of tiered structuring different from

jugation, which we shall call “hierarchics.” Unlike jugational relationships,
wherein one sacrificial element is added to another and therefore subordinate
to it, hierarchical relationships arise when two components constitute or
make up a third. Recognition of this property may assist in solving a number
of long-standing cruxes in biblical sacrificial passages and––since the
phenomenon extends to post-biblical (rabbinic and Qumranic) texts as
well––provide a better understanding of Σ in general. The examples are
presented in logical rather than chronological order, proceeding from the
simplest attestations to the more complex cases.
As demonstration that the “hierarchic” category is real, and not merely
some misunderstood form of jugation, note that zoemes in the Israelite
system are never joined to one another in subordinate jugation, for offering
one zoeme does not entail the offering of another, subordinate zoeme.
Therefore, two A-level zoemes offered together in a single ritual context are
more properly characterized as being in a constitutive, hierarchical
relationship.
It is important to clarify exactly how these hierarchies differ from the
patterns described in . In the process of jugation, A-level sacrificial
material entails (subordinate) B-level sacrificial material, these jugates in
turn entail C-level sacrificial material, and so forth. Hierarchics, by contrast,
involves sacrificial materials on some Level x that together constitute an entity
on Level x + 1. In other words, jugation models the relationships between
a sacrificial element and its ancillary or adjunct materials, represented
graphically as a right-branching tree, whereas hierarchical relationships
are compositional and are represented graphically by a vertically oriented
structure, as in the case where several smaller offerings compose or constitute
a single larger offering.
By way of example, consider the law in Num 28:11–15, which specifies
the calendric sacrifices to be offered on the New Moon (Figure 17). In this
11:50:13:06:14 Page 104

Page 105
Hierarchics 105
Fig. 17. Hierarchic structure of wholeburnt offerings in Numbers 28:11–15.
Fig. 18. Hierarchic structure for the New Moon ceremony.
case, P explicitly names the Level +1 category (Num 28:14, Wo$oDvX

mv
dB $oD
eX
o TL
x Eo
1
HNm$
mdox
H Y$
aoDmvL; and 29:6, $oD
vX eXox
H TLx Eo). This ritual complex, together with
another, much smaller coterminous offering consisting of only one zoeme
(a 웧 offered as a TAU m
dxX), generates the hierarchic structure shown in
Figure 18.2
Adopting the convention of labeling these individual zoemes as Level 0
sacrifices, we then denote such a junction of several independent jugates
(in the present example, these jugates are all zoemes) as Level +1. When
1
Note dHTmm NiiW $oD
XvM eX
oxH TLx Eo, “the ‘olah of the New Moon and its [the ‘olah’s] cereal offer-
ing,” clearly implying that the HX Ni is subordinate to the HL
mvM m Eo here (not WoTXNiiW $oD
mvM eX
ox x Eo,
H TL
which would denote “the cereal offering of the New Moon”).
2
As in the case of zoemes, the fact that certain sacrificial categories are unnamed does
not preclude that they were indeed conceived of as relevant categories. It is clear from Num
28:14–15 that this 웧0purification is not subsumed under $oDeX oaH TLxEo . Note that while the
wholeburnt+1 offering is explicitly named $oDeX ox x oWE in 29:6, the purification+1 offering
H TL
(theoretically $oDeX
oxH TAUd
xxX, consisting of a single zoeme) is not explicitly named in the text
(see n. 99 in Chapter 3).
11:50:13:06:14 Page 105

Page 106
several Level +1 complexes are joined, they constitute a Level +2 complex, and
so forth. The study of the complexes that result from these junctions is here
referred to as hierarchics.
4.2 DUAL USAGE IN THE HIERARCHIC SYSTEM
It is commonly held that the terms used to designate sacrificial types––such

as HLm WoE (“wholeburnt offering”), TAU m
dxX (“purification offering”), OYM imLv$o
(“wellbeing offering”), and O$ moAm (“reparation offering”)––are mutually
exclusive: an individual animal, for example, can be a TAU m
dxX or an HL
m WoE,
but not both at the same time. Contrary to this assumption, we will argue
that a single animal can, in fact, be offered as more than one sacrificial type
at one and the same time.3
In order to better understand the system of hierarchics, one must take into
account a linguistic property of Israelite sacrificial texts, which––though
too obvious to have gone unnoticed––has not been fully recognized by
many commentators. We will refer to this phenomenon as “dual usage.” It so
happens that the terms designating types of sacrificial categories are used
within Priestly traditions with two distinct denotations. Though this will
likely generate some consternation for modern readers, it appears not to have
been a source of confusion for the authors of the ritual texts.
4.2.1 Higher Level Shares Its Name with All of Its Constituents
The term HLm oWE (“wholeburnt offering”) offers the simplest example of this
phenomenon. At least two denotations of this term are consistently dis-
tinguished within P:
(a) wholeburnt0 : a sacrificial animal offered in accord with a given
sequence of actions in which the flesh of the animal is burnt in its
entirety on the altar
(b) wholeburnt+1 : a collection of two or more wholeburnt0 sacrifices,
together constituting a single ritual
For example, DYM iT
mx
d H TL x Eo (Num 28:6, wholeburnt+1) is a single ritual
comprising two TWoLWoE (each a wholeburnt0) separated by many hours: in
Priestly literature, the first is called RQ
edoBx x Eo (the wholeburnt offering of
H TL
3
See Meshel 2013b.
11:50:13:06:14 Page 106

Page 107
Hierarchics 107
the morning, as at Lev 9:17), while the second is presumably BRee EmH TLx Eo.4
When referring to calendric sacrifices offered on special occasions in
addition to (LE x) this double daily offering, P may employ a phrase such
as DYM iTmx
d H TL x Eo LE x (“in addition to the daily wholeburnt offering
[wholeburnt+1]”).
Similarly, the phrases WoT ddxBx
$ov
d
B TBd
x$ xo TLx Eo (Num 28:10) and $oD eXoxH TLx Eo
(29:6; or Wo$oDvXmvd
B $oDeX
o TLx Eo, 28:14), though grammatically singular, all refer
to complexes greater than a single wholeburnt0 offering: two lambs on the
Sabbath, and two bulls, a ram, and seven lambs on the New Moon.
This phenomenon is not limited to wholeburnt offerings: P likewise
mentions an offering termed OE mmL R$eoA
q OYM imLv
$dox
H XBxeZ (Lev 9:18) that con-
sists of two zoemes, each offered as a wellbeing-type offering {웧0wellbeing,
S웧0wellbeing},5 as in Figure 19. The authors of P are evidently undisturbed by
this type of dual usage. This is apparent from the phrase RQ edoBx
H TL x Eo
DYMiTm
d x
H TLx EoL
v R$eoA
q (Num 28:23), referring to the regular daily sacrifice
of the fifteenth of the seventh month, where the two denotations of HL m Eo,
“wholeburnt0” and “wholeburnt+1,” are found side by side (see Figure 18).
The term TAU m
dxX offers another telling example. Like HL m WoE and OYM imLv$o,
the term TAU m
dxX has at least two distinct denotations:
(a) purification0 : a sacrificial animal offered in accord with a given
sequence of acts that include sprinkling or daubing the sacrificial
animal’s blood on the sancta, turning designated internal parts to
smoke on the altar, and having most of the flesh consumed either by
the officiating priest or by fire.6
(b) purification+1 : a collection of two or more purification0 sacrifices
together constituting a single ritual
The phrase OYRiPuiKx
dd H TAU xx
d X (“the purification offering of purgation”),
mentioned twice in connection with the sacrifices performed on the tenth
Fig. 19. Hierarchic structure in Leviticus 9:18.
4
The second is not explicitly named, but a reasonable guess would be that the authors of P
thought of it as BReeEHm TLx WoE; it is also thus termed in 2 Chr 31:3 (the phrase OYiBd
x RvE
xHm IYB x WoE,
a TL
d
if it were grammatical in BH, would denote the timing more precisely, at least for P).
5
Since parts of their flesh (and possibly their blood) are physically placed together, it is likely
that these constitute a tightly linked sacrificial complex. The OYM imLv xeZ, in turn, is embedded
$o XB
in a larger sacrificial complex.
6
Milgrom 1976a.
11:50:13:06:14 Page 107

Page 108
Fig. 20. “Purification offering of purgation”: minimal hierarchical structure.
day of the seventh month, is commonly thought to refer to more than one
animal.7 It is evident that TAU m
dxX in these verses is used in sense (b.), denoting
a complex that includes at least two purification0 sacrifices, a B웧 (from
Lev 16:3) and a 웧 (16:9),8 as in Figure 20.
Some scholars have argued that TAU m
dxX in v. 5 is a super-category of the
type purification+1 consisting of two (identical) zoemes: one 웧 offered as a
purification0-type sacrifice (the 웧 “for ”) and one 웧 dispatched
to the wilderness,9 though the latter goat is not, strictly speaking, a sacrifice.10
If this interpretation is correct, then it further substantiates our argument for
the existence of two distinct denotations of the term TAU m
dxX.
Thus far, we have witnessed HL m WoE, OYMim $o, and TAU
Lv m
dx X sacrifices that
appear to function in the same manner: several Level 0 zoemes constitute
a Level +1 sacrificial complex of the same sacrificial type as the zoemes.
The examples we have seen so far are of a kind, and they find regular
enough attestation to suggest a hierarchic rule H1, which appears to be
operative in P:
H1: x+1 = [(x0)a , (x0)b, . . . , (x0)n] is grammatical
Read: a sacrificial complex of type x on hierarchic Level +1 comprising n offerings of
the same type (x), each on hierarchic Level 0, is grammatical.
x+1 designates a Level +1 sacrificial complex (wholeburnt+1, purification+1
or wellbeing+1, consistently)
x0 designates Level 0 sacrifices (wholeburnt0, purification0, or wellbeing0,
consistently)
a, b, n represent serial index numbers
7
Exod 30:10 (where the date is not mentioned) and Num 29:11; both texts hark back to Lev
16. For possible reconstructions of these passages’ text history, see Milgrom, Leviticus 1:1061–3.
8
This is almost explicit in Gane 2005:221, who still uses the term “collectively,” but it does
not appear that he means “collective” in the strict linguistic sense.
9
See Schwartz 1995:18 and n. 57. See also Gane 2005:252–3, who cites this opinion along-
side several alternatives. It is unclear whether, according to this view, the 웧 dispatched into
the wilderness is a TAU m
dx X in its own right (though one of a very different nature from most
purification offerings) that combines with another 웧, which is a purification0 in its own right
(this is explicit in v. 9), to create a third, higher-level TAU
m
dxX; or whether the former 웧 merely
complements the latter.
10
See Eberhart 2011:30–1.
11:50:13:06:14 Page 108

Page 109
Hierarchics 109
Fig. 21. Rule H1.
Or, schematically, the structure shown in Figure 21 is attested in P. At first

glance, the distinction between different levels of a particular type may seem
somewhat artificial, since each Level +1 complex in the above examples
entirely comprises Level 0 constituents of the same name.11 One might
suspect that wholeburnt+1, for instance, is simply a collective form of the
same noun that designates wholeburnt0.12 However, it is more likely that these
are not collective forms but ordinary singular nouns that belie a far more
complex system, as will presently become clear.
4.2.2 Higher Level Shares Its Name with Some of Its Constituents
4.2.2.1 Hierarchics “above” the Level of the Single Animal
Consider the following, slightly more complex structure, taken from the
Temple Scroll, concerning the tenth day of the seventh month, as follows
(25:12–13):
HEB$ HN$ YNB OY$BK DXA LYA DXA RP HWHYL HLWE WB HMTBRQHW
You shall make a wholeburnt offering thereon to , one bull, one ram, seven
male lambs in the first year of their lives.
This law does not differ from the law in Numbers in the hierarchics it implies.
In accordance with rule H1, a wholeburnt+1 (l. 12, see 27:4) consists of 1B웧,
1S웧, 7s웧, every one of which is presumably offered as a wholeburnt0.
The law further enjoins one 웧 for a purification offering (OYZE RYE$
TAUXL DXA, ll. 13–14), again in accordance with the law in Numbers,
noting (along with Num 29:11) that these sacrifices are in addition to the
purification offering of purgation (OYRWPKH TAUX, l. 15). However, the
Scroll differs from the law in P in its description of this OYRWPKH TAUX
(25:15–16):
11
See, for example, Maimonides HYH 1:1, who notes TWLWE OLWK, concerning these lists
in Numbers.
12
Thus Nihan 2007:117 n. 33. For HL m oWE as a collective noun, see Lev 6:3, 5 (HL m meaning
m EoH
all TWoLWoE), 18, etc.; as well as Ps 51:18 etc. A telling example is 2 Chr 35:1–16, where HL m Eo and
TWoLEo are interchangeable, as are XS xeP and OYX
d imSvP. Similarly, the phrase OYM
d imLv xeZ could be
$o XB
read as referring to many animals without resorting to a hierarchic theory: XB xeZ may also be used
collectively, and the noun OYM imLv$o is plural regardless of the number of animals offered.
11:50:13:06:14 Page 109

Page 110

iii
WYLE LWDGH IHWKH BYRQY DXA HLWEL OYN$ OYLA WBYRQT OYRWPKH TAUXLW
[. . .] WHYBA TYB LEW
And for the TAUX offering of purgation, you shall offer two rams as an HLWE ––
one shall be offered by the High Priest on behalf of himself and his family [. . .]
Although the text cuts off here, it is safe to assume that the succeeding words
would have referred to the other ram.13 And even without this conjecture,
the term TAUX is clearly used here as a superordinate category, consisting at
the very least of two rams (S웧) that are not themselves purification0-type
sacrifices but rather wholeburnt0-type sacrifices, and in all likelihood also
of two purification0-type sacrifices (a bull, B웧, and a he-goat, 웧) required
in Leviticus 16 and discussed in col. 26. Whether or not the second he-goat
(the “scapegoat”) was subsumed in the Scroll under this OYRPKH TAUX
i
cannot be determined with certainty, but it is unnecessary to assume that
it was.$
In sum, the hierarchic structure of the sacrifices according to the Scroll
includes (but is not limited to) the levels shown in Figure 22. (The formulas
and diagrams are simplified inasmuch as they do not reflect the order in
which the animals are offered; nor do they reflect internal hierarchic
arrangements, such as any special relationships that may exist between pairs
of animals, e.g., the ram and the bull or the ram and the he-goat.) A natural
question to ask in light of this model is, ‘How can a single zoeme be an ‘olah
and, at the same time, part of a hatta’t?’ Stated differently, how can a sacrifice
that in terms of its praxemics (that is, its physical manipulation by the priests)
m oWE ––it is dissected and burnt in its entirety, its blood is tossed (rather
is an HL
than daubed) on the altar, its hide is presumably handed over to the priest––
function as a component of a purification offering?
Fig. 22. The “purification offering of purgation” in the Temple Scroll:

hierarchic structure.
13
See 1983 2.114 and Werman 2006:101. Thus, the text may have read (IHKH BYRQY) DXAW
LHQH OE LWK LE (compare 18:7; 26:7, 9; Werman 104 suggests LHQH OE LE BYRQY DXAW).
11:50:13:06:14 Page 110

Page 111
Hierarchics 111
It appears that each S웧 is a wholeburnt-type sacrifice in terms of its

praxemics but part of a larger TAU m
dxX complex in some other sense. This
other sense is one of the mysteries of sacrifice, and is apparently related
to the desired effect of the ritual and, if desired effect plays a part in ritual
meaning, may pertain to the meaning attached to sacrificial rituals.14 The
formal structure of this ritual can be represented as follows:
H2: x+1 = [(x0)a ,(x0)b , . . ., (x0)m ,(y0)a , (y0)b ,. . ., (y0)n] is grammatical
Read: A sacrificial complex of type x on hierarchic level +1 comprising m offerings
of type x each on hierarchic Level 0, as well as n offerings of type y each on
hierarchic Level 0, is grammatical.
x+1 designates a Level +1 sacrificial complex (wholeburnt+1, purification+1 or
wellbeing+1, consistently)
x0 designates Level 0 sacrifices (wholeburnt0, purification0, or wellbeing0,
consistently)
x, y designate two Level 0 sacrifices (wholeburnt0 and purification0)
a. . . n represent serial index numbers
Schematically, rule H2 can be represented (Figure 23). The clearest cases of
the application of rule H2 are found outside of P,15 as in the example from
the Temple Scroll above, and in the following example from the Books of Ezra
and Nehemiah. However, as we shall see, rule H2 was operative in P as well.
Fig. 23. Rule H2.
14
On the schools of thought that equate ritual meaning with a ritual’s desired effect, see
§6.2.1.
15
If the author of Num 29:11 could be proven to have known of the ritual details in Lev 16––
and not merely of the existence and general nature of such a ritual––then the phrase
OYRiPu
di
dKxH TAU d
xxX in this verse could include wholeburnt sacrifices and accord with the opinion
of the Temple Scroll ’s author. This would further affirm our hypothesis on dual usage, as rule H2
would also be attested in P. However, this reading cannot be proven with certainty. Gane 2005:46
assumes that the author of Num 29:11 was aware of the details of Lev 16, but he suggests
that OYRiP u
di
dKxH TAU d
xxX here refers only to the inner-sanctum purification0 offerings. The term
OYRiPu
di
dKxH TAU dxxX at Exod 30:10 seems to be limited to purification0 offerings and not wholeburnt
offerings (compare Lev 16:3, 5); note that the blood of this OYRiP udi
Kx
d H TAUd
xxX is applied to the
horns of the golden altar. However, on the same argument, one may argue that OYRiP u
di
dKx
H TAU d
xxX
in Exod 30:10 includes all of the offerings of Lev 16, when in fact the formula TAU d
xxX ODxM
d i
OYRiPu
di
dKxH (referring only to part of its blood) is clearly limited to the blood of the inner-sanctum
purification offerings. This, however, is highly conjectural. See also Nihan 2008:184–5.
11:50:13:06:14 Page 111

Page 112
Let us first consider Ezra 8:35:

i
R$ E-OYN
om m a$vo OYRiP
m
d LA
a Rmov$iY YH
aoLALa TWoLEo WBYRiQ v iH HL Gx -YN
m oWdH aBv YB
iv $
dox
HM a OYA id
mBx
H
TAUm
dxX YRaYP
ivC HEmvB$
i oWv OYE
ivB$i o OY$
oi m
Bd
vK H$
mdo$
i oWv OYEiv
$oT
i OYL
d i YAa LA $iY-LK
a Rmov m -LE
d x
HWHYL x HLm Eo LK
do x
H R$ om m
E OYNa$
vo
The returning exiles who arrived from captivity offered wholeburnt offerings to
the God of Israel: twelve bulls for all Israel, ninety-six rams, seventy-seven16 male
lambs, and twelve he-goats as a purification offering, all this as a wholeburnt
offering to .
In what sense are these twelve TAU m
dxX YRaYP
ivC (“purification he-goats”)17 a
TAU m
dxX, and in what sense are they an HL m Eo? Authors and translators in
antiquity struggled with this text,18 as do modern commentators, usually
resorting to one of several solutions. Most commentaries simply ignore the
problem19 or hint in passing that the verse is difficult.20 Others unjustifiably
choose to omit the last three words of the verse, read them as if they refer
back to the bulls, rams, and lambs only,21 or offer complex and unwarranted
textual conjectures.22 A few have suggested that these TAU m
dxX goats were not
eaten––apparently, they were of the “inner” type, which is burnt outside
the camp (as at Lev 6:23)––and that hence they were somewhat like HL m WoE
offerings, which are not eaten.23 A much simpler solution may now be
offered, based on what we have seen in the Temple Scroll. In terms of their
16
1 Esdr. 8:66 mentions seventy-two lambs. Batten 1913:329–30 accepts this number as
original, but see Fensham 1982:122.
17
This is equivalent to TAU m
dxX YRaYEov (see also Dan 8:21, which probably implies that when
i$
the text was composed, the meaning of the term RYE i$ om was no longer clear, as it required
interpretation).
18
1 Esdr. 8:65–6 circumvents this problem by referring to the entire offering simply as a
θυσα, and specifically to the goats as τράγου 5π"ρ σωτηρου (perhaps the author has OYM imLv
$o
in mind). Most MSS of LXX on Ezra 8:35 retain the apparent contradiction found in MT by
offering the standard equivalents for TAU m
dxX and HL m WoE (περ α;µαρτα and <λοκατωµα/
<λοκατωσι, respectively). Josephus (Ant. Jud. 11:137–8, perhaps influenced by 1 Esdras)
partly solves the problem by omitting the concluding phrase “all this as a burnt offering to
,” but he refers to the bulls as being offered 5π"ρ . . . σωτηρα. See Marcus 1978:380. For
one talmudic solution, see n. 000 below.
19
See Blenkinsopp 1988:173; Levering 2007:96; Roberts 1993:134; Kidner 1979:67;
Kettlewell 1906; vom Orde 1997; Brockington 1969:105; Breneman 1993; Clines 1984.
20
Becker 1990:50; Allen and Laniak 2003:70. Vos 1987:69 notes that purification offerings are
burnt only in part.
21
Snaith 1971: 150–2, dismissing all previous solutions, suggests either omitting these words
or placing R$ om m
E OYN
a$vo TAU m
dxX YRaYP
ivC (“and twelve he-goats as a sin-offering”) in parentheses
to avoid the “impossible” suggestion that a sin offering could be part of a burnt offering. See
Fensham 1982:122, who suggests that the phrase in v. 35b refers back to the burnt offerings
because they were the largest part of the offering; note that the various text-historical
conjectures cited by Fensham on p. 121 do not purport to solve this problem. Also see Noss and
Thomas 2005:202.
22
Batten 1913:329, reconstructs HEW$T YRPC or HL$L OYRPC [sic] on the basis of LXXAB.
23
See Williamson 1985:114; Hartom 33. In b Hor 6a it is suggested that these he-goats were
“like a burnt offering, as a burnt offering must not be eaten, so were those sin offerings not to be
eaten”––and the context of the pericope clearly indicates that these goats were understood to
11:50:13:06:14 Page 112

Page 113
Hierarchics 113
praxemics, these he-goats were offered as purification0 offerings (that is, their
blood was daubed, not tossed, on the altar; only their suet was burnt up; etc.).
However, they were offered along with several other animals, each of which
was offered as a wholeburnt0 offering in terms of its praxemics. The Level +1
sacrificial complex as a whole functioned as a wholeburnt+1 offering. The
complex can be illustrated schematically (Fig. 24).
Fig. 24. Hierarchic structure in Ezra 8:35.
Thus, the text of Ezra should probably be interpreted as a simple example

of the application of rule H2: a sacrificial complex of hierarchic Level +1 (in
this case, wholeburnt+1) comprises nearly two hundred animals, twelve of
which are purification-type sacrifices (purification0) and the rest wholeburnt
sacrifices (wholeburnt0).
This conjecture suits the grammar of the verse perfectly: TWoLEo (plural)
refers to these animals individually (12·wholeburnt0), whereas HL
m Eo (singular)
refers to the complex as a whole (1·wholeburnt+1).24
4.2.2.2 Hierarchics “beneath” the Level of the Single Animal

Thus far, we have seen examples where Level 0 accords with the zoemic level
and Level +1 accords with a “sacrificial type” comprising several zoemes. The
have been offered in the fashion of an inner (= “burnt”) purification offering, not in the fashion
of a wholeburnt offering. See Keil 1888:113; Bertheau 1887:108; Ryle 1893:113. Some of the
classical Jewish commentaries, followed by Zer-Kavod, suggest that the goats atoned for the
Judeans though they were burnt in their entirety on the altar. The analogy to 2 Chr 29:21,
however, cited by Zer-Kavod as ostensible proof of this interpretation, seems to support the
suggestion (made below) that from a praxemic perspective the goats were treated as purification
offerings (compare iWQRvZ
vY
d
ixW, used three times in v. 22, with iWAU
v
dx Yx in v. 24). See Zer-Kavod
XvW
1980:57. Note that this approach does distinguish between praxemics and function but does
not accord with the talmudic solution mentioned above.
24
The form HLm oWE can be used in a collective sense with no change in meaning, and it could
have appeared in this verse instead of TWoLEo. However, the singular HL
m Eo in this verse could not
have appeared in the plural, as it designates a single sacrificial complex (one Level +1 HL m WoE
consisting of multiple Level 0 wholeburnt and purification offerings).
11:50:13:06:14 Page 113

Page 114
Temple Scroll, however, offers an interesting transformation in the conception

of the hierarchics of sacrifice that is slightly different (15:3–9; Yadin’s trans-
lation is emended according to Qimron’s reconstructions):
i
DXA LS OYAWLMH YLYA
o o
i
LWKL OXL YLSW OWYW OWY LWKL LYA LYA OYAWLMLW]
o
ii
OWYW OWY LWKL OYMYH T EB$L OYLSHW OYLYAH LW]K TA WCXW DXAH LYAL
i ii
HSKMH BLXH TA]W 25 LYAH IM HLWE HWHYL OYBYRQM WYHY HMHYT]WQLXMK
R$A BLXH TAW HMHYLE R$A] BLXH TAW TWYLKH YT$ TAW BRQH TA
WTXNMW DBKH TRTWY TAW HHYCE TMWEL] HYLAH TAW OYLSKH LE
UP$]MK WKSNW
3. And for the ordination, one ram for every [day,]

3a. [and] baskets of bread for all the ra[ms of the ordination, one basket for
e]ach
4. [ram.] And they shall divide a[l] l the rams and the baskets for the seve [n
days of ordination for every]
5. day, according to [their] divisions they shall offer to 
6. a burnt offering from the ram, and [the suet that covers the entrails, and the]
two
7. kidneys and [the] suet that is on the [m, and the suet that is on]
8. the loins, and the [entire] broad tai[1] near its backbone, and the caudate
lobe,
9. and its cereal offering and its drink offering, according to the ord[inance.]
Here it appears that certain parts of a single zoeme (apparently, not only the
right thigh but also the suet) are considered a wholeburnt offering, although
the animal in its entirety is an ordination offering (Figure 25).26
Fig. 25. Hierarchic structure of an ordination offering according to the Temple Scroll.
25
Yadin’s reconstruction of IYMYH QW$ TA here is not only reasonable but also expected
from a grammatical point of view, since BLXH TAW (and not BLXH TA) follows. To be sure,
according to Qimron’s collation (Qimron 1996) these words are not found in the text, but
the immediate context (see ll. 11–12: AYH HLWE . . . IYMYL R$A HMWRTH QW$ OE) proves
that these words are at the very least implied in LYAH IM HLWE.
26
See previous note and Schiffman 1995.
11:50:13:06:14 Page 114

Page 115
Hierarchics 115
What could it mean that these parts are an ‘olah? Apparently, the author
believes that they are a wholeburnt offering in their praxemics. The suet
and right thigh are not flayed, and their blood is not extracted and dashed
upon the altar, but the fact that they are burnt in their entirety––the most
prominent characteristic of wholeburnt offerings27 ––is sufficient grounds
for calling them HLm oWE in the Temple Scroll. Since they do not constitute a
complete zoeme, we will designate them as wholeburnt–1.
The grammar of the Scroll is consistent in this, as a comparison of this
last example and the following text demonstrates (16:6–13; Qimron’s
reconstructions, Yadin’s translation with slight modifications):
i
ii i i ii
LE R$A BLXH LWK TA IW$YARH RPH BLX TA RYUQHW XBZMH LE BRQYW
TAW HMHYLE R$A BLXH TAW TWYLKH YT$ TAW DBKH TRTWY TAW BRQH
i i
XBZMH WRYUQYW HMUP$MK WK]SN TAW WTXNM TAW OYLSKH LE R$A BLXH
WRWE TAW RPH R$B TAW vacat (?) HWHYL XWXYN XYR H$A AWH HLWE
HM$ TWAUXL LDBWM OWQMB $DWQH HNXML J]WXM WPWR$Y W$RP OE
WBLXM DBL HM$ WLWK WPR$W WYBRQ LWK OE WYERKW W$AWR LE WHWPWR$Y i i
o
AWH IHWK TAU X
6. [And he shall approach/sacrifice upon the al]tar and burn th[e suet of the
first bull]
7. [all] the suet that is on the entrails and th[e caudate lobe and the two]
8. kidneys and the suet that is on the[m] and th [e suet that is on]
9. the loins, and its grain offering and [its] liquid [offering, according to their
ordinance, and he shall bur [n them on the altar,]
10. a wholeburnt [offering] it shall be, an offering by fire, a pleasing odor
be[fore  vacat(?) But the flesh of the bull]
11. and its hide with its dung they shall burn out side the [holy encampment
(?)]
12. in a place set apart for the sin offerings; there th[ey] shall bur[n it with its
head and its legs,]
13. with all its entrails. And they shall burn all of it there, except its suet; it is a
[priest’s] sin [offering.]28
Here, too, the suet is considered an HL
m WoE (l. 10), although it is obviously
not flayed and its blood is not extracted and tossed upon the altar.
27
See Watts 2007:63–78.
28
While H$ i in Biblical Hebrew probably denotes “food gift” (see Milgrom, Leviticus
edoA
1.161–2), the translation “offering by fire” is retained here, since late Second Temple authors
probably assumed that H$ i was connected with $oA
edoA a . See, for instance, the juxtaposition of the
phrases H$A XBZMX LE LWKH RYUQ[T] (with reference to the wholeburnt offering) and
$AH LE RYUQT (with reference to the suet of the wellbeing offering) in 4Q220:5–6 (4QJube).
Similarly, while TAU m
dxX in Biblical Hebrew denotes primarily “purification offering” (see
Milgrom, Leviticus 1.253–4), for this late Second Temple author, TAUX may have implied a sin
offering. Note the evidence cited by Milgrom, and see also the pun employed by the author of
the Temple Scroll in 35:14–15.
11:50:13:06:14 Page 115

Page 116
Once again, the fact that it is burnt in its entirety on the altar––the most
prominent characteristic of wholeburnt offerings––whereas parts of the
bull are burnt outside the camp after the manner of “inner” purification
offerings, is sufficient grounds for considering the suet an HL m WoE in the
Temple Scroll.
Thus, a Level 0 purification offering (B웧0purification) is here conceived as
comprising a wholeburnt–1 offering (the suet), which is burnt in its entirety,
and a purification–1 offering (the flesh), which is treated like an “inner”
purification offering, as in Figure 26. Schiffman demonstrates this phenom-
enon clearly and also notes that the Temple Scroll perceives all purification
offerings as containing smaller wholeburnt offerings, namely, the portions
that are turned to smoke on the altar.29 Since the Scroll appears to be con-
sistent on this matter, it is reasonable to postulate that its author believed
every purification offering (at least of a quadruped) comprised both a whole-
burnt–1 offering that is turned to smoke on the altar and a purification–1
offering that is either consumed by the priest (in the case of an “outer”
purification offering) or burnt outside camp (in the case of an “inner”
purification offering).
This “grammatical” property appears to have occurred to R. Abraham
Ibn Ezra (twelfth century), who implies that the offerings of the wealthy and
indigent offerers described in Lev 5:6 and 5:7, respectively, should be inter-
preted as shown in Figure 27.30 Ibn Ezra notes the correspondence between
each bird and a different part of the quadruped. However, he does not
explicitly state that the praxemic order is preserved inasmuch as the
Fig. 26. Hierarchic structure of a purification offering according to the Temple Scroll.
29
See Schiffman 1995.
30
Ibn Ezra’s words are, concisely, TAUX YN$HW OYRWMAH DGNK DXAH$ YLA BWRQHW
UP$MK (on Lev 5: 7). Shachter 1986:15–16 translates, “The more straightforward explanation,
in my opinion is: One [bird] provides the equivalent of the fats of a sin-offering that are wholly
consumed, while the other is the actual sin-offering that is called for by law.” Note that Ibn Ezra
does not explicitly state that the second bird corresponds to the flesh and blood of the quadru-
ped, but he comes very close to doing so. Note that the zoemes indicated in the diagram (Fig.
27) follow P rather than rabbinic tradition.
11:50:13:06:14 Page 116

Page 117
Hierarchics 117
Fig. 27. Ibn Ezra’s interpretation of the offerings of the wealthy and the
indigent offerers.
m WoE follows the TAU

HL m
dxX in both cases. In the first case (Lev 5:6, if this
law indeed presupposes the laws found in Lev 4, such as vv. 30–1), the
purification–1 blood is applied to the altar, and only then is the wholeburnt–1
(consisting of the suet of the female sheep or goat, 웨) offered on the altar
(the flesh is presumably consumed by the priest). So too, in the second case,
the purification0 bird is offered, and only then is the wholeburnt0 bird offered
(finally, the flesh of the purification0 is consumed by the priest).
Thus, in a number of cases, the term HL m WoE, when interpreted as whole-
burnt–1, appears to designate the physical act of complete burning; but when
interpreted as wholeburnt0, it is the name of a sacrificial type. The distinction
is a fine one, but it is not without parallel: the term LYL idK appears to cover
m
a similar semantic field. In BH (see, for instance, Lev 6:15, and probably
Deut 13:17, despite its ritual overtones) the term serves, among other uses, as
an adverb designating “completely” (in the context of immolation by fire). In
addition, however, and probably by a specific semantic shift, it came to serve,
in BH as well as other Semitic languages,31 as a technical term for a sacrificial
type (Deut 33:10; Ps 51:21).32
The phenomena presented from Figures 17 to 27 are all of the same nature
31
The form kll apparently may designate a sacrificial type in other Northwest Semitic sacri-
ficial texts as well. On kll in Ugaritic and Punic, see Dietrich et al. 1975a, 1975b; Loretz 1975, and
the discussion in Pardee 2000:1.648–649 (on RS 24.260:9–10), who rejects the alleged attestation
of šlmm kll in Ugaritic, dividing the sentences (and the letters kll) differently. It is usually
assumed (mostly on etymological grounds) that the kālîl-type sacrifices were burnt in their
entirety, although in fact some parts of the kālîl were probably left unburnt. The Punic tariff
from Marseille (KAI 69) almost certainly suggests that part of the flesh of the largest quadrupeds
(ll. 3–6) that are offered as kll offerings becomes the property of the priests.
32
On Lev 6:15–16, see Milgrom, Leviticus 1.400–1.
11:50:13:06:14 Page 117

Page 118
Fig. 28. Hierarchies above and beneath the zoemic level.
Fig. 29. Rule H3.
in their essentials: several lower-level sacrificial types constitute a higher-level

complex, and the authors resort to dual usage of terms for sacrificial types.
They differ, however, in the specific animals involved, in the relevant sacri-
ficial types, and in the levels on which this hierarchic relation occurs––
“above” or “beneath” the level of the zoeme, as demonstrated in Figure 28.
Accounting for the structures attested in Figure 26 and in the left column
of Figure 27 simultaneously requires a more general rule:
H3: x0 = [(x–1)a , (x–1)b ,. . .,(x–1)m , (y–1)a ,(y–1)b ,. . .,(y–1)n] is grammatical
Read: A sacrificial complex of type x on hierarchic Level 0 comprising m offerings of
the same type (x) each on hierarchic Level –1, as well as n offerings of a different
type (y) each on hierarchic Level –1, is grammatical.
Or, this can be shown schematically (Figure 29). At this point, rules H2 and
H3 can be characterized as particular cases of a more general rule:
H4: xp+1 = [(xp)a , (xp)b ,. . .,(xp)m , (yp)a , (yp)b ,. . .,(yp )n] is grammatical
Read: A sacrificial complex of type x on hierarchic Level p+1 comprising m offerings
of the same type (x) and n offerings of a different type (y), each on hierarchic Level
p, is grammatical.
Or, this can be shown schematically (Figure 30).
11:50:13:06:14 Page 118

Page 119
Hierarchics 119
Fig. 30. Rule H4.
Now that we have understood the property of dual usage and identified
hierarchics as an operative category in the grammar of the Israelite system, we
are in a position to understand more clearly several biblical and post-biblical
sacrificial complexes. In the present discussion, the post-biblical examples,
which are more straightforward, will be discussed immediately below,
followed by the biblical examples on pp. 000–00.
4.2.3 Higher Level Shares Its Name with None of Its Constituents
Thus far, we have established that a term designating a sacrificial type may
refer to different hierarchic levels. In each case, we have assumed that the
designation of the lower level corresponds to the praxemic aspects, whereas
the higher level corresponds to another aspect. This other aspect we will
provisionally call a ritual’s “meaning,” with a necessarily broad sense of the
word in mind, including the task a ritual is designed to carry out, the function
that a ritual fulfills, and any symbolic correspondence that may exist between
a ritual’s elements and those of another similarly structured system; for a
discussion of these and other senses of “meaning,” see below, Section 6.2.
In all of the cases discussed so far, the higher-level type shares its name
with one or more of its lower-level constituents, but this does not need to be
the case. Let us now consider hierarchic structures in which the higher-level
type shares its name with none of its constituents.
4.2.3.1 Hierarchics within a Single, Bare Zoeme

The simplest of these cases––though perhaps the most counterintuitive––is a
single animal constituting two different sacrificial types, one on the praxemic
level and another on a higher level that, as noted above, relates to the ritual’s
meaning. Such a combination is apparently considered grammatical in
several theoretical post-biblical formulations. One example is found in
b Zeb 9b in the statement AYH HLWE IW$XN TAUX (“Nah.shon’s purification
offering is a wholeburnt offering”). In its context, this statement appears to
assume that the zoeme (웧) offered by each of the chieftains in Numbers 7
was, in reality, as shown in Figure 31. This very same model is applied to other
11:50:13:06:14 Page 119

Page 120
Fig. 31. Hierarchics within a single zoeme according to b Zebah. 9b.
Fig. 32. Purification zoemes in P.
offerings, like the purification sacrifices offered by a Nazirite and a person

purified from scale disease, of which is said WHNYN TWLWE YNH (“these are
wholeburnt offerings”). In all of these cases, the rabbinic authorities clearly
believed that every offering termed TAU m
dxX in the biblical texts is to be
manipulated as a TAU m X (that is, its blood is applied to the horns of the altar,
dx
only its suet is incinerated on the altar, etc.),33 and so the statement that “these
purification sacrifices are wholeburnt offerings” is not paradoxical at all:
whereas they are praxemically purification0, they function as wholeburnt+1.34
P clearly does not share this view, as nothing in P prevents the reader from
assuming that these purification sacrifices were like any other purification
offering (Figure 32). However, an analogous structure may indeed be found
33
As for the offerings of Nazirites and persons purified from scale disease, the biblical verses
clearly imply a praxemic distinction (Lev 14:19–20, H$ om m
EvW vs. HX
mda
BvM
Zd
ixH . . . HL
mfEe
HvW). In rabbinic
literature, the purification offerings of these two groups are praxemically purification offerings.
Since the pericope assumes that these two are analogous to the purification offering brought by
the chieftains in Num 7, it clearly assumes that the chieftains’ purification offerings were also
praxemically purification offerings.
34
This may also explain a troubling formulation that is found elsewhere in rabbinic
literature. For example, b Zeb 76b refers to a situation in which a person should stipulate
HBDN YML$ L$ HZ O$A WAL OAW. Note that the wording is not YML$ L$ WZ HMHB WAL OAW
HBDN, but HZ O$A. This may imply that this Reparation0 is offered––as is evident from
the immediate context––like a standard O$ moAm , so that praxemically it is an O$ m : it must be
moA
slaughtered to the north of the altar, its blood is tossed upon the altar, and its flesh is eaten by
priests only, and yet it serves functionally as a OYM im
Lv$o, since it is not offered as a reparation for
anything.
11:50:13:06:14 Page 120

Page 121
Hierarchics 121
in P, in two controversial verses. Let us first consider the less problematic case,
Lev 5:6a:35
HB
d
m$ov d
iK IAC H-IM
do x i HB
maQvN AU
mm
X R$
eoA
q oWTAU
m
dxX LE
x HWHYL
x oWM$ A-TA
moq e AYBiaHv W
TAU
m
dxXvL OYZ
diE
i TRxYEov -WoA
i$
“He shall bring his O$ m to  for the transgression that he has committed, a
moA
female from the flock––a female ovine or a female caprine––for a TAU m
dxX.”
For a reader accustomed to viewing sacrificial types (like TAUm

dxX, O$ m , and
moA
m WoE) as praxemically distinct and mutually exclusive sacrificial types, it
HL
seems incongruous that as one’s O$ m one must offer a TAU
moA m
dxX animal. One
wonders how this animal (a female sheep or goat, 웨) can be offered as
both an O$ m and a TAU
moA m
dxX: in particular, if it is a TAU
m
dxX, then its blood ought
to be daubed on the horns of the bronze altar (or applied inside the sanctuary,
in the case of an “inner” TAU m
dxX) and the rest of the blood should be poured
on the base of the altar; and if it is an O$ A, then its blood ought to be tossed
mom
on the altar, like that of wholeburnt and wellbeing offerings (see Lev 7:2).
Some previous interpreters have resorted to highly conjectural text-
historical reconstructions and to the assumption that some of the Priestly
authors used the terms O$ m and TAU
moA m
dxX synonymously, as if unaware of the
clear praxemic distinction between the two.36 Commentators aware of the
clear-cut distinction P draws between the two have resorted to an unlikely
interpretation of the noun O$ m in this verse as “debt” or “compensation” in
moA
a general sense, forcing an interpretation of the verse as follows: “He shall
offer his compensation to  for the transgression that he has com-
mitted.”37 However, this reading must be rejected. Although O$ moAm certainly
35
The laws permitting the indigent to offer birds or a cereal offering in Lev 5:7–13 are
intended to apply only to the specific case discussed in 5:1–6 (pace Milgrom, Levine, following
rabbinic tradition, and contra de Vaux, Wenham, and Péter-Contesse, who believe that they
apply to 4:27–35 as well, following Philo and Josephus; see Gilders 2004:111 and 215 nn. 5–6).
Note that P explicitly states that this law applies only to “one of these” (HL e
dAaM
a TX x ), at the
xA
beginning of the pericope (5:4–5, referring to the list in 1–4a) and then again at its end (5:13), as
distinct from HNY$ET AL R$A HWHY TWCM LKM TXA (5:17; NRSV: “any of the things that by
the Lord’s commandment ought not to be done”). See also Schwartz 2004:216, who expresses
some doubt on this matter.
36
For a representative example, see Wellhausen 1973: 75 and n. 2; Murphy 1874:77;
Gerstenberger 1993:51–2. Rendtorff 1967:207–9 posits a text-historical development which led
to the present confusion, but his analysis cannot be accepted because of its ad hoc ascription of
WoM$moqA to an editor in v. 7 but TAUm
dxXvL to an editor in v. 6, despite the fact that––as he admits in
209 n. 1––the she-goat is typically a TAU m
dxX, not an O$moAm.
37
See for example Milgrom, Leviticus 1.303 (“penalty”); Snaith 1967:48–9; Levine 1989:28;
Coffman 1987:42; Budd 1996:92–3; Wenham 1979:86 (“reparation”); Rooker 118 and n. 182;
and the translations of NEB and JPS. Rendtorff 1985:195 proposes “Schulddarbringung” but
unnecessarily insists that this is distinct from O$ m as a technical term for an offering type (as
moA
in 5:15). Some of the translations render the verse literally (and correctly), e.g., RSV (“guilt
offering”), JB (“a sacrifice of reparation”), but this leaves the incongruity unexplained. It is
possible that it is this passage that misled Philo to believe that the sacrificial type O$ m is a
moA
subcategory of the type TAU m
dxX. See Daniel-Nataf 2000: 2.269 and n. 271.
11:50:13:06:14 Page 121

Page 122
denotes “debt” and “compensation” elsewhere in BH (as at Num 5:7, 8 bis; 1

Sam 6:3, 4, 8, 17), it is highly unlikely that this is its meaning in this verse.
First, the analogy to Lev 5:11, which is very similar in structure, and pertains
to an offering that is a substitute for the offering mentioned in Lev 5:6,
suggests that O$ m in v. 6 is analogous to IB
moA d m in v. 11 (concerning the
m RvQ
grammar of v. 11 see immediately following):
(v. 6) TAU
m
dxXvL HB
d
m $oK
i IAC
d H-IM
do x i HB
ma QvN AU
mm
X R$ eoAq HWHYL
x oWM$ A-TA
moq e AYBia
HvW
OYZ
d
iEi TRxYE ov -WoA
i$ WoTAUm
dxX LE
x
(v. 11) TAU
m
dxXvL TL
eSo HP
maAH
m TRxY$
oi q
E AU
mmX R$
oe q
A WoNB
d m-TA
m RvQ e AYB
ia
HvW
He shall bring his reparation offering to  for the transgression that he has
committed, a female from the flock––a female ovine or a female caprine––for a
purification offering. (v. 6)
He shall bring his offering for having transgressed, a tenth of an ephah of
semolina for a purification offering. (v. 11)
Hence, it is likely that O$ m here is in fact a sacrificial type and not a “debt”
moA
or “compensation” in general.
Note that in contrast to v. 6, where AU mmX R$ q can logically be read as
eoA
modifying the preceding noun (WoTAU m
dxX), in v. 11 it is clear that the clause
AU mmX R$ q cannot logically be read as a relative clause modifying WoNB
eoA d m, but
m RvQ
must be understood adverbially as “for having transgressed” or “because he
transgressed” as––certainly an attested, though rare, meaning.38
Second, two distinct verbs are usually used with reference to the offering
of an O$ m in BH, corresponding to the two distinct denotations of this
moA
noun: O$ moA
m BY$ a and O$
i oH moA
m AYB iaH. The former is used invariably where
O$ m denotes “debt” or “compensation” in a general sense, as in the formula
moA
WoM$ A-TA
moq e BY$ a (Num 5:7, 8 bis; 1 Sam 6:3, 4, 8, 17; see also Lev 5:23),
i oH
and, in one case, to denote the reparation offering (Num 18:9). By contrast,
O$moA
m AYBiaH is used only where O$ m denotes a zoeme offered as a reparation
moA
offering (Lev 5:15, 25; 19:21; Num 6:12; see also Lev 14:23) and never where
O$ m denotes “compensation” in the general sense. This is hardly surprising,
moA
since WoM$ A-TA
moq e AYBiaH in this case is only a particular instance of the general
38
For R$ q with the force of “because” or R$
eoA eoA x, see BDB s.v. R$
q LE q 8c, p. 83b; KBL
eoA
s.v. R$ q B.b.; GKC §158b; J–M 170e; Gordis 1951:254, 213 (on WNXYNA$) and 286 (on 8:11);
eoA
and particularly Gen 30:18, Y$ i oYA
i vL YT
imXvP$i o YT
iT
d N-R$
xm eoA
q YRiK
m$
oi OYH
i oLf
A IT
xmN (JPS “for having
given”); Gen 31:49, RM xAm R$eoA
q HP m
dvCiM
dxHvW (JPS “and [it was called] Mizpah because he said”);
1 Kgs 3:19; 15:5; 1 Kgs 8:33 || 2 Chr 6:24. Note that JPS translates Lev 5:7 as “he shall bring to the
LORD as his penalty for that of which he is guilty” and elsewhere (e.g., 4:3; 5:6) consistently
translating AU mmX R$eoA
q oWTAUm
dx X as “the sin of which he is guilty,” always taking the R$ q of
eoA
AUmm X R$ q in a sense very close to R$
eoA eoA
q LE x (“for that which he has sinned”). See 1 Kgs 16:19,
H$oi mE R$
eoA
q oWTAUm
dxX.
11:50:13:06:14 Page 122

Page 123
Hierarchics 123
type WoNB
d m-TA
m RvQ e AYB H (or AYB
ia H + animal) standard in P (for example, Lev
ia
4:23, 28, 32; Num 15:25; see also Lev 5:18).39
These data indicate that if the verb BY$ a were used in Lev 5:6, it would be
i oH
reasonable (though not necessary, in light of Num 18:9) to interpret O$ moA
m
as a general “debt”; but since the verb AYB iaH is used, a verb used with O$moA
m
only to denote the offering of a sacrifice, it appears that O$ m in Lev 5:6 must
moA
denote a sacrificial animal.
Verse 6, it appears, supplies us with an unusual type of hierarchic strati-
fication, very similar to that depicted in Figure 31. This verse reflects the
hierarchic structure shown in Figure 33.40 This reading accords perfectly with
the logic of Lev 5:6–10, as we shall presently see.
4.2.3.2 Hierarchics with Several Zoemes

Let us now consider the more problematic case, Lev 5:7:
TAU
m
dxXvL DX
mAe HWHYL mWoY-YN
x HN aBv YN
a$ vo-WoA OYRiT
o YT
a
dv$o AU
mmX R$
eoA
q oWM$ A-TA
moq e AYB
ia
HvW
HL
m EoL
v DX
mAev
W
He shall bring his O$ m for having transgressed, two turtledoves or two pigeons
moA
for , one for a TAU m
dxX and one for an HL
m oWE.
Previous interpreters of this verse have resorted either to dubious textual

emendation of MT or to the unlikely interpretation of the noun O$ moA
m
in this verse as “debt” or “compensation” in a general sense, which has
39
2 Kings 12:17 is of a different order (the subject is the silver, TWoAU
m
dxX FS
eeW
Kv TWoM$
moq
A FS
ed
eK),
and it is therefore not analogous to the verses cited here. In any case, the passage would only
support the argument made here since it demonstrates that AYB H is never paired with “debt” in
ia
general, but only with the sacrificial type. Gen 26:10 is irrelevant here.
40
Remarkably, Murphy 1874:77–9, though offering a very confused description of the
praxemics involved, comes close to recognizing this structure, noting that “in these cases
the trespass-offering and the sin-sacrifice are united in the same victim” and “this particular
sin-sacrifice comes under the head of a trespass-offering.” Staubli comes even closer, stating
that the offering is “zwischen Sünd- und Schuldopfer: Der Form nach ist es ein Sündopfer,
inhaltlich geht es aber um die Wiedergutmachung einer Schuld” (Staubli 1996:66–7). Some
translations adhere to an ultraliteral (and thus correct) translation of this verse, such as RSV
(“guilt offering”) and JB (“a sacrifice of reparation”).
11:50:13:06:14 Page 123

Page 124
now been rejected.41 Note that the analogy to verse 11 is here even more
straightforward:
HWHYL mWoY-YN
x HN aBv YN
a$vo-WoA OYRiT
o YT
a
dv$o AU
mm
X R$
eoA
q oWM$ A-TA
moq e AYB
ia
HvW
(v. 7) HL
m EoL
v DX
mAevW TAU m
dxXvL DX mA
e
TL
eSo HP
mA amH TRiY$
oi q
E AU
mmX R$
eoA
q oWNB
d m-TA
m RvQ e AYB
ia
HvW
(v. 11) TAU m
dxXvL
He shall bring his reparation offering for having transgressed, two turtledoves or
two pigeons for , one for a purification offering and one for a wholeburnt
offering. (v. 7)
He shall bring his offering for having transgressed, a tenth of an ephah of
semolina for a purification offering. (v. 11)
Leviticus 5:7 reflects the two-tiered structure shown in Figure 34, which
is quite similar to those above. This sacrificial complex differs from the
examples discussed thus far in one detail only: both Level 0 sacrifices are of
different types from the Level +1 sacrifice that they constitute. This, how-
ever, should come as no surprise to anyone acquainted with the laws of the
reparation offering: the O$ m , according to Biblical law, consists of some
moA
HLm WoE aspects (such as tossing the blood) and some TAU m
dxX aspects (the fact
that the flesh is given to the officiating priest and that it expiates).
Some objections may be raised against this interpretation on grammatical
as well as textual grounds. Grammatically, one might argue that it is some-
what forced, since it requires interpreting AUmmX R$
eoAq not as a relative clause
modifying a noun, as in the immediately preceding verse (v. 6), but as an
adverbial clause, “because he sinned” or “for having sinned,” an admittedly
rare (though attested) meaning.42
However, this objection should be dismissed, since we have already seen
that in the same chapter (v. 11), an almost identical construction is found:
AUmm X R$eoA
q WoNB
d m-TA
m RvQ e AYB ia W. This verse supports the interpretation of
Hv
41 42
See p. 000. See n. 000.
11:50:13:06:14 Page 124

Page 125
Hierarchics 125
R$ q as “for having sinned” (like R$

eoA eoA
q LEx), since AU
mmX R$ q , as noted above,
eoA
cannot be a relative clause modifying WoNB d m but must be interpreted as an
m RvQ
adverbial clause.
The second objection is textual: in LXX, the equivalent for R$ eoA
q WoM$ A-TA
moq e
AUmmX is περ τ1 α;µαρτα α.το! = 8µαρτεν (“for the transgression
which he committed”), which at first seems to reflect a Vorlage that reads
AUmmX R$ eoA
q WoTAU m
dxX LE x.43 However, such a reconstruction is dubious,
since LXX often fails to reflect the nature of O$ m as a technical term for
moA
a sacrificial type, resorting to a host of periphrastic formulations to avoid
translating it as such.44
It appears that verse 5:7 is one such case: the Vorlage, like MT, read
AUmmX R$
eoAq oWM$ A-TA
moq e , but the translators misconstrued it.45
The reading of Leviticus 5 and 6–7 that is presented here not only offers
a solution to an ancient crux that could not have been solved without the
43
BHS suggests AU mmX R$eoA
q WoTAU m
dxX TA e , perhaps under the influence of the manner in which
LXX treated v. 11 (MT AU mmX R$ eoA
q oWNB
dm RvQ
m TA e ). Note that while AU mmX R$ eoA
q oWTAU m
dxX is common in
P (Lev 4:3, 23, 28 bis, 35; 5:6, 13), meaning “the transgression that he committed [or ‘trans-
gressed’],” TAU m
dxX in this formula never denotes the sacrificial type, but only the sin. By contrast,
AUmm X R$ eoAq oWM$moqA would be a hapax, if this were the grammatical construction here. While
both the sin and the type of sacrifice that often serves to expiate it share the name TAU m
dxX in BH,
the authors of P never confuse the two denotations (Hosea is probably punning on the two
meanings in 4:8; see Milgrom, Leviticus 1.286–7). In fact, if one accepts the LXX reading
according to BHS, then our verse would offer an explicit example of Ibn Ezra’s reconstruction in
Figure 27: a Level +1 TAU m
dxX comprising a Level 0 TAU m
dx X and a Level 0 HL m WoE.
44
For instance, consider the following three examples:
(1) Lev 5:15. MT O$ mvL . . . LYiA
moA x (“a ram. . . as an O$ m offering [i.e., a reparation offering]”)
moA
|| κριν . . . περ ο> πληµµ#λησεν (“a ram . . . for that which he trespassed [i.e., for having
transgressed]”); noun replaced by a verbal clause.
(2) Lev 5:19. MT HWHYL i H O$
x O$ o oA
m O$ o oA
m AW m (“It is an O$
moA moAm offering, he has trespassed
[inf. abs. + finite vb.] against ”) || πληµµ#λησεν γὰρ πληµµ#λησιν ναντι κυρου (“he
has committed [‘trespassed’] a trespass against ”); misinterpretation of the presumed
Vorlage, technical meaning of O$ i H O$
moAm ignored, AW m lacking equivalent.
moA
(3) Lev 7:1. MT O$ moA
mHm TRxWoT d TAZov W (“This is the law of the O$ m offering [i.e., reparation
moA
offering],” see HL m EoH d TAZo, HX
m TRxWoT mvMNd
ixH TRxWoT d TAZov W, etc., 6:2, 7, 18; 7:11) || κα ο>το < νµο
το! κριο! το! περ τ1 πληµµελεα (literally: “And this [is] the law of the ‘ram [offered] for
the trespass’ ”); here, the translators limit the reparation offering to rams, forgetting or ignoring
the injunctions of Lev 14:12, 21; Num 6:12.
The mistranslation in Lev 5:7 differs from the examples above in that the translators do not
use the verb πληµµελ#ω, but rather a noun, α;µαρτα. However, this is not surprising in light
of the fact that in two other verses (Lev 4:22 and 5:4) LXX resorts to a similar solution when
faced with the stative verb O$ aoAm : in the first case, the equivalent of O$ aoA
mv W is κα α;µάρτ: κα
πληµµελ&σ: (perhaps a doublet); in the second, simply κα α;µάρτ:. Cf. Wevers 1997:56, 64,
80; Harlé and Pralon 1988:101, 103.
45
It is likely that the translators were further confused by the juxtaposition of TAU m
dxX and
O$ m , as their appositive translation in v. 6 indicates. In v. 6, where the MT is unequivocal
moA
(AU mm X R$eoAq WoTAUmdxX LEx HWHYL x oWM$ moqA-TA e AYB iaHvW), LXX has κα ο?σει περ @ν πληµµ#λησεν
κυρ7 περ τ1 α;µαρτα = 8µαρτεν, literally, “he shall bring for [those things] that he
trespassed against , for the sin that he committed”; that is, LXX reads WM$A but miscon-
strues the verse, rendering it to mean AU mmX R$ eoAq WoTAU m
dxX LE x ,HWHYL x O$ moAm R$eoAq LEx AYBiaHv
W
etc.
11:50:13:06:14 Page 125

Page 126
identification of hierarchics within Σ, it also accords perfectly with the logic

of the entire passage, vv. 6–10: the individual has transgressed in a particular
manner and must offer a reparation offering (to be precise, a reparation+1
offering). This reparation may come in (at least) one of two forms, either
an 웨 offered as a purification0 sacrifice (from a praxemic point of view),
or 2/, one bird offered as a purification0-type sacrifice and one as a
wholeburnt0-type sacrifice. The following verses (11–13) imply that this
reparation+1 offering may also come in the form of a purification0 grain
offering, which is perfectly congruous with the hierarchic structure described
here, notwithstanding the obvious fact that the praxemics of a grain
purification0 offering differ considerably from the praxemics of an animal
purification0 offering.46
This last instance (Lev 5:11–13) is further complicated by the fact that,
aside from being a purification0 offering and, presumably, a reparation+1
offering (though this is not explicit in P), the offering of the extremely indi-
gent is also, in a very concrete sense, a grain offering (see Lev 2). Apparently,
P does not consider it a HX Ni , since in the law concerning its consumption P
mvM
notes that its laws are like those of the HX Ni (v. 13). Rabbinic sources, how-
mvM
ever (along with many modern scholars, perhaps erroneously) do consider
this offering to be a type of HX Ni ;47 in this case, this meager 0.1 ephah
mvM
becomes, from a “grammatical” point of view, the grandest offering––con-
stituting three different kinds of offerings at once: a TAU m
dxX, an O$ m , and
moA
a HX Ni .
mvM
46
Moreover, from the second half of the chapter (vv. 14–26, particularly vv. 15, 18, 25; cf.
19:21 [H]), it is evident that a reparation offering (O$ m +1)––a different one, to be sure––may
moA
appear in the form of a reparation offering (O$ moAm 0). The awkward statement that an O$ mAo may
m
appear in the form of an O$ mmoA, though counterintuitive, is supported by the grammar of vv. 15
and 25 (and is not contradicted by v. 18): in each case, the noun O$ m appears twice in the verse,
moA
first with a pronominal suffix designating its functional relation to the offerer (reparation+1),
and then followed by a vL-, corresponding to the Level 0 sacrificial type as defined by its
praxemics. Thus, the general structure of these verses is almost identical to that of Lev 5:6, 7. It is
therefore possible that here, too, a hierarchic structure is implied––only that in the case of the
O$ m , the zoeme is convertible into silver. This interpretation may point in the direction of a
moA
general theory of meaning in Σ, namely, that every sacrificial offering consists of a hierarchic
structure including a meaning-related (Level +1) tier and a praxemic (Level 0) tier––even where
no such tiering is explicated. According to such a model, even a layperson’s voluntary whole-
burnt0 offering would constitute a tiered structure––a wholeburnt+1 expressed as a wholeburnt0,
but as the praxemic type and the “meaning”-related type converge (they share the same name),
the hierarchic structure is not explicated. This model would need to be nuanced further in
light of the consideration that hierarchic structures may consist of more than two tiers (see
immediately following).
47
For example, m. Men. 5:3, which refers to this offering as AUWX TXNM, listing it alongside
the offering of the suspected adulteress (HNWBL ALW IM$ AL –– TWANQ TXNMW AUWX TXNM),
which is considered a HX Ni already in Num 5:15. See also the phrase ]M$AH TXNM in
mvM
Qimron’s reading of TS 38:8 (Qimron 1996:54).
11:50:13:06:14 Page 126

Page 127
Hierarchics 127
4.3 SOME GENERALIZATIONS
In order to account for the hierarchic complexes described in Figures 31, 33,
and 34, a more comprehensive rule should be formulated as follows:
H5: z+1 = [(x0)a , (x0)b ,. . .,(x0)m , (y0)a , (y0)b ,. . .,(y0)n] is grammatical
Read: A sacrificial complex of type z on hierarchic Level +1 comprising m offerings
of type x and n offerings of type y, each on hierarchic Level 0, is grammatical.
Or, the rule can be shown schematically (Figure 35). Note that the only
difference between rules H5 and H2 is that in H5, the sacrificial type on
Level +1 (z+1) need not share its name with any of its constituents (“x” or “y”).
Stated differently, H2 is merely a particular case of H5 in which z = x.
All five rules formulated thus far are in fact only particular cases of a
general rule H6 that accounts for all hierarchic structures encountered here:
H6: zp+1 = [(xp)a , (xp)b ,. . .,(xp)m , (yp)a , (yp)b ,. . .,(yp)n] is grammatical
Read: A sacrificial complex of type z on hierarchic Level p+1 comprising m offerings
of type x and n offerings of type y, each on hierarchic Level p, is grammatical.
Or, the rule can be shown schematically (Figure 36). Rule H6, it appears,
underlies all of the hierarchic structures that we have examined so far. In
other words, H1–H5 are merely particular cases of H6:
H1: x = y = z, p = 0
H2: x = z, p = 0
H3: x = z, p = –1
H4: x = z
H5: p = 0
Following the method exemplified in this chapter, it is possible to collect
further data––that is, more instances of ritual structuring––to examine the
operation of H6. Note that although rule H6 can be employed to generate
numerous “grammatical” (that is, licit) hierarchic complexes, certain
restrictions may apply to it, limiting the number of structures that it can
generate in a manner analogous to the negative formulations in Maimonides’
abstractions (see p. 000). Since an unattested structure is not necessarily
Fig. 35. Rule H5.
11:50:13:06:14 Page 127

Page 128
Fig. 36. Rule H6.
ungrammatical, such stipulations on H6 can only be conjectured at present,

but further instances and more ample attestation can more clearly delineate
what stipulations may be active on our postulated rule. For example, the
largest hierarchic structure presented here, from the point of view of the
number of animals involved, consists of less than two hundred animals
(Figure 24); similarly, no examples have been found in which a wellbeing
offering is subordinate to a wholeburnt or purification offering. It stands to
reason that there is nothing inherently ungrammatical within Σ about a
structure with two hundred or more animals. It may or may not be the case
that the subordination of a wellbeing0 to a purification+1 is “grammatical.”
One example from a late rabbinic text suggests that H6 can be applied to
ordination offerings, and also demonstrates that innovative application of H6
persisted in Σ even in purely imaginary contexts. Maimonides in HMQ 2:14
(following b Men 45a) suggests that the numerous wholeburnt and purifi-
cation offerings listed in chapters 43–46 of Ezekiel’s Temple vision, including
their subordinate libations, are in fact OYAiiWdLM
i (“ordination offering[s]”)––in
hierarchic terms, higher-level ordination offerings.
Note that in most of the hierarchic structures we have seen, the overall
number of types involved is limited to two or three, though it is possible to
imagine hierarchic complexes consisting of four or five sacrificial types on
the same hierarchic level. Moreover, we have seen that usually only two levels
are at play (Level 0 and Level +1, or Level –1 and Level 0), and none of the
structures presented above has exceeded Level +2. However, it is possible to
postulate the existence of Level +3 complexes consisting of sacrificial rites,
performed beside the altar and on it, and other rites that are not related to the
altar at all, such as the lighting of oil lamps.48 This requires a certain leap of
theoretical imagination and depends on the modern reader’s propensities.
It stands to reason that some Israelite authors were themselves of such
dispositions: the complex potentially formed by constituent sacrifices and
lamp-lightings was tagged by those priests DYM m .49
iT
d
48
Haran 1978:205–29 (“The Ritual Complex Performed Inside the Temple”).
49
The priestly formula DYM
iTmx
d H TLx WoE indicates that it is a noun, serving as an absolute in
a construct chain. The word DYMiT
dm is also clearly a noun in the late BH of Daniel (e.g., 8:11;
11:31).
11:50:13:06:14 Page 128

Page 129
Hierarchics 129
Moreover, the combination of hierarchics “below” and “above” the zoemic
level within a given context is to be expected in large Level +1 complexes
that contain more than a single purification0, which itself consists of a
purification–1 and a wholeburnt–1, as at Figure 26, viewed as part of the larger
ritual described in Figure 22. Such considerations suggest that hierarchic
patterns within Σ can become even more complex, including numerous
sacrificial types on several hierarchic levels above and below the level of the
single zoeme.
11:50:13:06:14 Page 129

Page 130
Praxemics
5.1 INTRODUCTION
The study of ritual from the point of view of physical acts is defined here as
praxemics. The noun, derived from πράξι, is not a complete neologism:
the term praxeme has been used in the past to denote the elementary unit of
ritual activity.1 In fact, previous contributions to the topic of ritual grammar
have taken this operative category as the central aspect of which all other
grammatical properties are, in a sense, derivatives.2
In this grammar of Σ, six components constitute a praxeme: (1) atomact
(compounded from atom+act), an elementary physical act in a series of acts;
(2) agent, the performer of an atomact; (3) object, the physical body that is
manipulated during this atomact, almost invariably materia sacra; (4) target,
to which the object is applied; (5) location, the spot where the agent is located
when performing the atomact; and (6) time-frame, the point in time at which
an atomact is performed.3 Every praxeme consists of an agent and exactly one
atomact,4 and at times one or more of the other four components.
1
According to Tambiah 1979:164, “praxeme” was coined by Catherine Bateson in 1974, but
in a different sense, pertaining to the “pragmatic” approach to ritual, according to which a ritual
is “examined not in terms of multi-level analysis (i.e. how lower-level units make up higher-
level units as adopted in descriptive linguistics) but in terms of its being experienced by par-
ticipants as made up of fused higher-level entities or whole chunks as single events,” or as a
“fused numinous experience.”
2
This is true even of the elaborate theory of Lawson and McCauley 1990. See also Michaels
2007:243 and Thomas 2006. This is not to say that previous grammatical analyses have viewed
ritual as mere action independent of zoemic, hierarchic, or jugational properties. However, since
these three categories were not identified as such, it has not been demonstrated that their diverse
inventories are subject to their own internal constraints.
3
Note that Haran (1978:1) enumerates four dimensions without which no discussion of
the cult is full: place, time, act, and person. The components referred to here relate to these
dimensions but contain finer distinctions. Note too, however, that by his own admission (p. 2),
Haran dedicates Temples and Temple-Service in Ancient Israel to “cultus,” not to sacrifice (a
subcategory of “cultus”). Therefore, the sacrificial offerings that took place in the courtyard do
not constitute a central part of Haran’s discussion, with the exception of the paschal f웧.
4
For a general theory of ritual, there is some advantage in this description over Lawson and
McCauley’s analysis, since they seem to imply (1990:92) that in religious ritual every action has
11:50:13:06:14 Page 130

Page 131
Praxemics 131
For example, the ritual activity corresponding to the offering of a

voluntary wellbeing offering includes a praxeme envisioned in the formula
BYBimSx Xd
aBvM
Zd
ixH-LE
x ODmx
d H TA e OYNiH
qdoKx
H IR
oqHAx ½YNd iWQRvZ
aB
v mWv (Lev 3:2, “and the sons
of Aaron, the priests, shall toss the blood upon the altar, roundabout”). The
atomact in this praxeme is expressed by QRxmZ; the agent by OYN iH
qdK
oxH IR
oqHAx ½YN
aBd;
v
the object manipulated by OD mx
d H; and the target by BYB im
Sx Xd
a
BvM
Zd
ixH-LE x.5 The
location is unexpressed in this formula; theoretically, it could be “atop
the altar” (if the priests are assumed to ascend the altar) or “near the altar.”
The time-frame is determined by the relation to other praxemes in the ritual
sequence; for instance, it is certain that this praxeme must follow the praxeme
indicated by the phrase DE aoWM LHeAo XTxed
P oWUX
m$vioW (v. 2, “and he shall slaughter
it [at the] entrance of [the] tent of meeting”).
I should note that several other components pertinent to sacrificial systems
are not included in the present praxemic description, such as the priestly
attire:6 though such components may render a ritual valid or invalid, they do
not in and of themselves differentiate one ritual from another. It is true, for
example, that a priest may be required to wear a different set of clothing
depending on his rank and the particular ritual activity in which he is
engaged; however, there is no case in Σ where the same acts, performed once
in set of clothing A and once in set of clothing B constitute two distinct, valid
rituals.
The six components that are included in praxemics will be discussed in
detail in Section 5.5. However, before this discussion can commence, the place
of praxemics within a grammar of ritual must be clarified.
5.2 PRAXEMICS AS ONE OPERATIVE CATEGORY

IN THE GRAMMAR OF RITUAL
The assertion that ritual sequences are composed of smaller building blocks is
not itself an innovation. This has been the underlying assumption of previous
an object that is acted upon (in their language, the structure of ACTION is always P + ACMPLX
+ P, never P + ACMPLX). This, however, may be false: reciting a mantra may also constitute a
ritual, wherein there might not be an object acted upon. The present formulation is also
somewhat more economic than that employed in Mishra 2010, inasmuch as “means” and
“manner” are subsumed under “atomact.”
5
On this praxeme, see p. 000.
6
As noted in n. 000, we do not discuss other dimensions (to which Haran refers extensively),
such as the genealogy of the agent, his personal history (e.g., whether he washed his hands and
feet prior to approaching the altar), or his bodily form. Some of these are subsumed under what
Lawson and McCauley term “embedding” (such as preliminary hand-washing); note that upon
their own admission (1990:112–13), “embedding” as they define it necessitates historical
assumptions that may be true or false, but do not seem to be necessary in some cases.
11:50:13:06:14 Page 131

Page 132
work on ritual grammar, and––as Michaels has argued7 ––of ancient ritual
manuals as well. In fact, a full praxemic description of a ritual system is
precisely what scholars from the Heidelberg Ritualdynamik circle have been
calling for explicitly in recent years.8 To be sure, praxemic analyses of discrete,
even extensive rituals have been carried out in several cultures.9 However,
these scholarly forays resemble analyses of selected sentences in a natural
language more than a systematic grammar: examples may be analyzed
according to their formal structure, their meaning, creative devices such
as alliteration, the use of palindromes, repetition, and a host of other
properties––but no sense of a whole language can be generated by such
restricted analysis. Thus, previous studies on the grammar of ritual have
been, as Michaels notes, more programmatic than proto-grammatical.10 What
is necessary––but has not actually been done––is the systematic listing of an
inventory of building blocks for a complete system, along with general rules
that determine which combinations of building blocks are grammatical and
which are not (rather than a list of grammatical combinations, which the
ancient legal texts themselves supply, and which represent only a subset of all
grammatical combinations).
This analysis is an attempt to offer a complete praxemics of an entire
system, beginning with an enumeration of all of its elementary units (as
identified and often named by the ritualists themselves) and proceeding
through a description of all of the grammatical sequences of these elementary
units in which their grammaticalities and ungrammaticalities are discernible.
More generally, it aims to formulate––inductively and empirically––the
systematic ordering of the given corpus in a comprehensive manner, and with
a sensitivity to the existence of differential levels of conceptualization––the
four operative categories identified here––and to rules of transformation
between these levels, on the basis of a logic that allows for certain sequences
but does not tolerate others. Thus, praxemics, a category that has been identi-
fied in the scholarship, combines in this study of the ancient Israelite system
with zoemics, jugation, and hierarchics––entire operative categories that have
not been previously recognized as such––to form the first full grammatical
description of a ritual system.
As noted (p. 000), previous contributions to the topic of the grammar of
ritual have considered what is here called praxemics as the central aspect
of ritual grammar. The underlying logic of these works, stated rather sim-
plistically, is as follows: just as languages are composed of words, so rituals are
7
Michaels 2007:243– 4 (referring to Gladigow as well).
8
Michaels 2007:244; see §1.3.4.
9
For instance, Humphrey and Laidlaw 1994 (Jainist); Lawson and McCauley 1990
(rituals in the West); Staal 1979 etc. (Vedic); Michaels 2005 (Nepali); Gane 2005 (Israelite,
Mesopotamian); and Payne 1999 (Japanese).
10
See Michaels 2007:241 and above, n. 000.
11:50:13:06:14 Page 132

Page 133
Praxemics 133
composed of actions.11 Since the grammar of language is a theoretical frame-

work for the analysis of words, the grammar of ritual ought to be a theoretical
framework for the analysis of deeds. However, the preceding discussion of
zoemics, jugation, and hierarchics should suffice to demonstrate that such
unilateral logic is faulty.
In other words, the situation in the study of ritual grammar today is some-
what analogous to a situation in which linguists are aware of one category,
for example phonetics, but do not acknowledge any other categories, thus
attempting to compose grammars of natural languages on the narrow basis
of phonetics. Such attempts would lead to grammars with a very low
explanatory power, if these could be termed grammars at all. Accordingly, a
successful grammar of a ritual system cannot be constructed solely on the
basis of praxemics.
Admittedly the praxemic description is less economical than zoemics,
jugation, or hierarchics, as there are many more components at the praxemic
level. Yet despite this limitation, it is possible to compose for Σ a praxemics
that is integral to the present .
5.3 SOME METHODOLOGICAL LIMITATIONS
There are a number of methodological issues that should be addressed with

special reference to praxemics. First, as opposed to zoemics and hierarchics,
which take place primarily in the worshippers’ minds, praxemics takes place
primarily in their limbs. As a result, praxemics could be expected to be more
directly observable and less theory-laden than other branches of a grammar
of Σ. Theoretically, a full picture of the praxemics of a ritual could thus be
attained simply by filming the rituals from a number of angles from beginning
to end. While this may be partially true for the grammatical analysis of
contemporary rituals, it is not true for the rituals examined here, since one is
entirely dependent on literary descriptions for a study of the Israelite system.
A more important limitation, inherent in ritual and not particular to Σ,
is present in praxemics. Unlike zoemes and jugates, which are by definition
discrete (it is easy to determine where one s웧 ends and the next s웧 begins),12
praxemes must be singled out by the observer through a process of dis-
cretization.13 The time from the beginning of the ritual to its end and the
11
This is of course a broad generalization. Utterances may be embedded in rituals, just as
body language interacts in complex ways with spoken language.
12
See, however, n. 000.
13
Michaels similarly points out that there is necessarily some degree of arbitrariness in
determining where one act ends and the next begins (2007:244).
11:50:13:06:14 Page 133

Page 134
participants’ movements are in constant flux, so that the ritual grammarian

must divide these into discrete units of action, such as “priest raises hand” or
“priest walks to bronze altar.” Therefore, the reader should keep in mind that
whereas in this study, praxemic sequences appear to comprise several discrete
praxemes, rituals in reality are often performed as uninterrupted (at times
overlapping and even simultaneous) streams of activity.14
However, it is at this point that the ritual grammarian treating systems
recorded in texts has an advantage over the direct observer: ritual texts already
contain ready-made praxemes that are usually explicitly named and thus
discretized and differentiated from one another.15 In the Israelite system, one
might say that the authors of the Priestly texts did much of the meticulous
work of discretization from which the modern ritual grammarian benefits
(see p. 000). Admittedly, this is an advantage only from a purely emic point of
view. Authors of ritual texts are themselves interpreters of rituals, and it is
possible that different individuals would discretize the same ritual in different
ways. Examples of such difference of opinion reveal that differences in dis-
cretization can in fact have significant ramifications.16 Yet the advantage is
an important one from a methodological point of view. As Goody notes, a
theoretician may attempt to impose an anthropological scheme upon live
footage, or even upon oral tradition, at the risk of freezing a contextual
statement into a system of permanent oppositions; and while such a scheme
simplifies reality for the observer it is often at the expense of a real under-
standing of the actor’s frame of reference.17 But as Goody himself suggests,
a theory developed on the basis of written documents avoids the problem of
decontextualization. For a “grammarian” working on P, frozen units of ritual
are, in a sense, prefabricated by the authors of the text.
The nature of P also imposes a severe methodological limitation that is
particularly evident in praxemics. The information in P’s sacrificial laws is
14
See Lévi-Strauss 1981:671–5, who likens ritual to a film that presents continuous action
despite running through twenty-one shots per second.
15
See Michaels 2007:243–4.
16
The quadratic scheme (e.g., m. Zeb. 1:4 et passim) of HQYRZ-HKLWH-HLBQ-HUYX$ ––
slaughter (of the zoeme), reception (of its blood in a vessel), conveying (the blood from the spot
where it was received to the spot where it is applied to the altar), and application (of the blood to
the altar)––is the starting point for many rabbinic discussions relating to the praxemics of
animal sacrifice. Yet concerning even this basic scheme, there is some disagreement among the
Tannaitic authorities, namely whether conveying is a praxeme or not (see R. Simeon’s opinion
in m Zeb 1:4, as well as the discussion b. Zeb. 14a–15b, which tries to reduce the controversy
in the Mishna to a controversy over whether conveying is an HDWBE or not). However, one
should be cautious in identifying our concept of “praxeme” with the amoraic conception of
HDWBE: note that according to one amoraic opinion (14b), slaughter is not considered an
HDWBE either, since it does not require priestly activity. Our conception of “praxeme” is wider
than ritual activity limited to the priesthood.
17
Goody 1977:12–13, 67, 71–3.
11:50:13:06:14 Page 134

Page 135
Praxemics 135
too scanty, too disparate, and too inchoate for the text to serve effectively as a
sacrificial manual in any strict sense.18 In the first place, a considerable
amount of gap-filling is required of the readers if they are to make sense of
the praxemic sequences referred to in the text.19 P also contains a number
of contradictions––some of which may result from the history of its
composition––most notably in the field of praxemics.20
Most significantly, P is inconsistent in the degree of detail with which it
describes sacrificial sequences. In some cases, the absence of a detail from the
text in P appears intentional, as it signifies the absence of a praxeme from
the ritual. In other cases, the absence of some of the legal details appears
to be the result of abbreviation. In most cases, unfortunately, one simply
cannot know. This renders argument from silence particularly hazardous in
the context of P’s sacrificial laws.
P may mention a detail once without noting that it is a general law, and yet
this detail may be assumed elsewhere in P. For example, Lev 1:7 refers to
placing coals and wood on the altar in the context of B웧wholeburnt offerings.
While this probably need not be done for every wholeburnt offering, it is
probably not limited to B웧wholeburnt offerings; it appears to be mentioned in
1:7 simply because this is the first in a series of sacrificial laws.21
A similar situation is found in the case of the praxemes described in P as
ODmx
d H-TAe OYNiH
qdoKx
H IRoqHA
x ½YN
aBd iWBYRiQ
v W (“And Aaron’s sons, the priests, shall
v iHv
offer the blood,” Lev 1:5) and HL m EoHm -TA
e UY$
i oP W (“And the wholeburnt
v iHv
offering should be flayed,” Lev 1:6), mentioned only in the context of the
웧wholeburnt and not of other wholeburnt offerings. It is hardly likely that P
saw a praxemic distinction between 웧 and 웧 in these matters––that is,
that bovines should differ from other quadrupeds as wholeburnt offerings in
that only the former must be flayed, and only their blood must be “brought
forth” before it is tossed on the altar (see also 7:8, which implies that the law
of flaying applies to all quadrupeds offered as wholeburnt offerings).22
18
The weighty questions of whether any of P’s sacrificial texts were ever designed or
intended as manuals, and if not, what narrative, polemic, ideological, or theoretical motivation
drove their composition, cannot be discussed in the framework of the present study, but see
n. 000.
19
This is repeatedly emphasized by Gilders 2004:10 et passim.
20
See nn. 00–00.
21
On P’s principle of economy, see Schwartz 2004:207, 225. Similarly, the qualification
OTmA o H$ oe q
ETx OYU
d e oS (Exod 29:2) clarifies that by semolina, P means wheat semolina (TL
i iX TL
d eS o
alone appears to have also a general sense including the refined flour of other grains as well).
Thereafter, P does not qualify TL o with OYU
eS i iX and resorts to the common use of TL
d eSo , which
is limited to wheat semolina (see also 2 Kgs 7:1, 16). Though the explicit qualification OYU i iX
d
is absent from all subsequent texts in P, in this particular instance it is safe to assume that it is
implied (see Num 5:15, which mentions OYRiEo$ ov XM Q as an exception). This situation is related
xe
to the fact that the laws in P are embedded in a narrative with a chronological sequence.
22
It is disputable just how far such generalizations can be carried. For example, it stands to
reason that other types of offerings in which only the suet is placed on the altar need not be
11:50:13:06:14 Page 135

Page 136
In other cases, however, the absence of a law from one passage and its
presence in another seem to imply a praxemic distinction between two rituals.
For example, the fact that the requirement HN mPom
C xXdaBvM
Zd H v
ix x (“on
VReeY LE
the north side of the altar,” literally “on the side of the altar, northward”)
appears only in the context of 웧wholeburnt (Lev 1:11) and not in the context
of 웧wholeburnt (1:5) has led some commentators to believe that 웧wholeburnt
offerings may in fact be slaughtered anywhere in the court (contrary to
Tannaitic law, as at m Zeb 5:1). Though drawing conclusions from silence is
problematic, this interpretation is supported by the fact that the law of the
웧wholeburnt follows––rather than precedes––the law of 웧wholeburnt, and by the
fact that the same pattern, namely, the absence of this detail concerning
bovines and the presence of the detail concerning sheep and goats, is repeated
in the law of purification offerings (Lev 4:24, 29, 33, in contrast to vv. 4, 15).23
Due to these inconsistencies, many passages in P concerning the praxemics
of sacrificial rites remain ambiguous. For example, the law pertaining to
hand-leaning appears in the context of 웧wholeburnt; it is then absent from the
laws pertaining to wholeburnt and /wholeburnt, but then mentioned again in the
context of wellbeing, wellbeing, and wellbeing (Lev 3), as well as in the context of all
types of purification offerings (Lev 4). It is completely absent from the laws
of the reparation offering in 5:14–26 and 7:1–7. In the case of the /wholeburnt,
this absence appears intentional––it stands to reason that birds would not
require hand-leaning.24 But it is hardly likely that wholeburnt offerings do
flayed or cut apart (to be precise, if they are flayed and cut apart, this is done for practical
reasons, e.g., for the convenience of the person eating the flesh, and does not constitute a
praxeme). However, rabbinic authorities require that other types of offerings be flayed and cut
apart as well, and they refer to an anomalous situation in which a (burnt purification) offering is
cut apart without being flayed. This is attested already in the Sifra (Weiss 18b). See also b. Zeb.
50a–b, 103a, as well as Maimonides, HMQ 5:18.
23
One should add to this the corroborating evidence from Ezek 40:35–43, if interpreted
correctly by Milgrom, as well as his interpretation of TS 34:1–8 (Milgrom, Leviticus 1.164–5).
However, his statement that “bovines cannot be slaughtered on tables” depends on the height
and surface area of the tables, dimensions which are not mentioned in Ezekiel. Note that the
four tables, with an upper surface of 1.5 square cubits and a height of 1 cubit (referred to in
40:42), are not those upon which the animals are slaughtered (vv. 39–41) but those upon which
the utensils for slaughtering and the flesh of wholeburnt offerings are placed. Furthermore,
the fact that the utensils are referred to as XB xd
mZx
HvW OB
d
m HL
m WoEH e iWUX
m -TA qv$oiY R$
eoA
q OYL
id
a H (42)
Kx
implies (if XB xd
mZx
HvW is not an interpolation, as its awkward position may suggest) that wellbeing
offerings as well were normally slaughtered on the north side of the court, despite the
fact that these are not mentioned in v. 39. Note that even if one were to take P’s formulation
TAU m
dxXxH UXam$doT
i HL
d m EoH
m UXam
$
doT
i R$
d eoA
q OWoQM
vdiB (Lev 6:18) as an indication that every purification
offering––and every wholeburnt offering––must be slaughtered at a specific spot within the
courtyard (namely, on the northern flank of the altar), one could not deduce from this that
the law applies to bovines: the author of 6:18 clearly has the eaten purification offerings in mind,
as is evident from its context, and these are never bovine.
24
This is a physical stipulation, since hand-leaning/hand-pressing requires that some force
be applied. Furthermore, if one accepts Wright’s interpretation of this rite, then hand-leaning
11:50:13:06:14 Page 136

Page 137
Praxemics 137
not require hand-leaning.25 Therefore, it is also difficult to determine whether

P intends to imply that, as opposed to wholeburnt, wellbeing, and purifica-
tion offerings, zoemes offered as reparation offerings do not require hand-
leaning; or whether the absence of this praxeme from the relevant passages is
merely a result of the textual peculiarities of this law.26
In addition to these complications, it appears that P assumes in many cases
that certain ritual laws are known without stating them explicitly. For
instance, one cannot gather from the laws in P that eaten purification
offerings must be consumed within a fixed period of time, and if so, what
its duration is––a single day, like the flesh of thanksgiving offerings; two
days, like the flesh of wellbeing offerings; or some other period of time.27
However, a careful reading of one of the narratives in P reveals that this
on birds is obviated, since the offerer is likely to be holding the bird in his hands: “hand
placement is the signature on a letter delivered to the god by means of a cultic postman. When
the god receives the letter, he recognizes that it is from the one who signed the letter (i.e., the one
who performed the gesture), not from the postman who delivered it” (Wright 1986:443).
Wright’s theory, including its important distinction between leaning one hand or two hands,
accords perfectly with the law, explicit in rabbinic literature (and probably implicit in P, 1:4 etc.),
that public offerings do not require hand-leaning. It even explains those public offerings that do
require hand-leaning (m. Men. 9:7), but not those private offerings that do not require hand-
leaning, namely in rabbinic law (the firstborn, the tithe, and the paschal f웧, ibid.). However,
Wright’s statement concerning the analogous Hittite hand-leaning rites (p. 441) is problematic,
since it undermines his own conclusion about the distinction between leaning one hand and
two hands: “One might wonder if the two forms of the gesture (one with contact and the other
without) carry two different meanings. This is doubtful. Because both descriptions are used in
the same ritual contexts (i.e., with offering materials) both certainly had the same meaning.”
Note that his claim that each Israelite pressed two hands on the blasphemer (p. 435) may be
true, but the logic in this case is circular. Furthermore, the fact that hand-leaning in Hittite
rituals may be performed from a distance (tuwaz) raises some questions about the physical
nature of hand-leaning in P and leaves open at least the possibility that a symbolic act is implied.
See also Paran 1985.
25
See Eberhart 2002:25–6.
26
See Milgrom, Leviticus 1.151. The law of the reparation offering is unusual in many ways,
such as its irregular distribution between the “didactic” section (1–5) and the “administrative”
section (6–7; on this terminology, see Rainey 1970). By analogy to the laws of the other types
of offerings, particularly the purification offering, one would have expected the praxemic
description of the reparation offering to be found in the first section (after 5:15a, 18 and 25), not
where it is located in the present text of Leviticus (7:2–5).
27
It is possible that in P, the cereal offerings offered along with the thanksgiving offering
are also to be eaten within a single day. Although the law in P (and H) places a time limit
only on the consumption of the flesh, and not on the consumption of its coordinate jugates
(Lev 7:17; 22:30) or of grain offerings in general, P’s narrative (8:32, in the context of an
ordination offering) suggests that the cereal offering was at least in some cases expected
to be consumed within the same period of time as the flesh. See TS 20:12–13, LKAT AWL
$M$H [HY]LE A[WB]T A[WLW] LKAT AWHH OWYB JMX (“it is not to be eaten leavened; it
is to be eaten on the same day [les]t the sun s[e]t up[on] it”), which appears to refer to all
eaten cereal offerings. On the consumption of cereal offerings in Sadducean, Qumranic,
and Tannaitic halacha, see Regev 1996 and Noam 2003:250–4; see also Qimron and Strugnell
1994:150–2.
11:50:13:06:14 Page 137

Page 138
information is assumed to be known in P: the flesh of a purification offering

is to be consumed within a single day.28
Another peculiarity of P is found on occasion, where P seems to refer to its
own laws as if they were actually performed in the chronological sequence
in which they are enjoined in the narrative. This is awkward, but recurrent
enough to be considered characteristic. For example, Lev 4:20–1 notes that
TAU m
dxXx
H RPxvL H$o m
E R$eoA
qxdK RP
m
dxL H$
om m
EvW and RP
m
dxH TAa FRx$om R$
eoA
qxdK WoTA
o FRx$
om v
W
m (“he shall do to this bull just as he has done to the hatta’t bull” and
IWo$oARiH
“he shall burn it just as he burned the first bull”), referring to the law (and the
B웧) that immediately precedes this law (and this B웧) only in the narrative
sequence in which they are enjoined, but not in reality. Similarly, the law
in Lev 3:5 assumes that there is already a wholeburnt offering on the
altar––most probably not because this must be so in reality (the morning
wholeburnt offering would have been consumed within a few hours, and
there appears to be no requirement to have a wholeburnt offering on the altar
at every given moment), but because it is enjoined in an earlier passage
(Ch. 1). This type of confusing narrative-chronological assumption is to be
distinguished from the simple technique of stating a law once and then
assuming that it is known (as at Lev 4:26, 35): here the assumption is not only
that the earlier law is known before the later law (which is true if the reader is
careful), but also that it is performed before it (which is false) at every given
point in P’s history.
Due to these methodological limitations, I focus on the praxemes that are
relatively clearly evidenced and on ritual laws that are generally uncontested.
5.4 THE PRAXEMIC DESCRIPTION
The first task in the treatment of praxemics is to define as clearly as possible

which activities are part of Σ and which are not––and on this basis, which
activities qualify as praxemes and which do not. Aside from the problem of
discretization, discussed above, there are three aspects to this definition.
28
Milgrom deduces this from the wording of Aaron’s reply to Moses in Lev 10:19
(OWodYx
H TAUm
dxX YT
i vLx
d KmAvW, “if [lit., ‘and’] I had eaten the purification offering of the day”). Since
several later sources, including Philo (DSL, 1.240), Josephus (Ant. Jud. 3.231), and the tannaim
(m. Zeb. 5:3), state this law explicitly without deducing it from the narrative, it is possible that
this was common practice in the late Second Temple period. Note that for the author of the
Temple Scroll, a limit of one day would mean the day of the offering, before sundown. See
Milgrom, Leviticus 1.418, referring to MMT B 9–13; TS 20:13 (see also Qimron’s reconstruction
in Qimron and Strugnell 1994:46, 149–52); and Regev 1996.
11:50:13:06:14 Page 138

Page 139
Praxemics 139
(a) Chronological delineation, that is, defining where the ritual begins and
where it ends.
Taking a simple example of a ritual consisting of
zoeme0 (type: wellbeing, genus: volitional),
one must determine whether the ritual begins with the designation of the
animal as destined for the altar (in some cases this can be done away from
the sanctuary, as in the case of the animal tithe in Lev 27:32). Additionally,
one must decide whether the slaughter of this animal is part of the sacrificial
ritual, as many scholars assume, or a mere prerequisite for obtaining materia
sacra––pure flesh and blood of domestic cattle.29
Similarly, one must determine whether the ritual ends with the application
of the animal’s blood to the altar, with the placing of its suet (or its flesh,
in the case of the wholeburnt offering) on the altar, with their complete
incineration several hours later, or with the consumption of its flesh many
more hours later (in some cases two days).30 This definition may become
quite complicated with connection to the DYM m and other calendric offerings,
iT
d
which in a sense have a beginning (Exod 29:37–8, immediately with the
consecration of the altar) but no end, since in the laws of P, including H, these
rituals are designed “for all time, throughout the ages.”31
(b) Delineation by ritual value––that is, decisions concerning which
activities within the time-frame of the ritual form part of the ritual activity
and which do not. Clearly, certain acts may take place during the ritual with-
out being ritual acts: for example, the priest may scratch his head or shoo a fly
off his shoulder, certainly not parts of the ritual.
Other acts may be essential to the ritual from a purely logistical point of
view, but they need not be part of the ritual. For example, the officiating
priest must breathe while performing the ritual, and he must transfer the
29
See Eberhart 2011:24, 29 for a balanced view. More radically, slaughter may be viewed
as a purely instrumental prerequisite. Note the phrase iWHA avd
Ux YxiWHU
XvW am Xv $d W in the context of a
oiYx
i
purification offering in Lev 9:15. The striking fact here is that WHU amXv
$d W is not part ofiWHA
oiYx avd
Ux Yx,
XvW
which, as a denominative from TAU m
dxX, here implies “perform everything involved in offering
a purification offering, including the blood application and the placement of the suet on the
altar.” See Gilders 2004:122. In ancient Israelite law (including but not limited to P), the only
licit way to obtain pure flesh of bovines, ovines, or caprines is by slaughtering the animal. With
birds, the problem is somewhat more complex, since P insists that this act be performed by a
priest, not by a layperson, which lends support to the view that pinching is a ritual act.
Furthermore, P knows that birds may be slaughtered rather than pinched in non-sacrificial
contexts (as at Lev 14:5–6, 50–1), and so it is likely (but unproven) that P requires that bird
offerings be pinched rather than slaughtered because pinching is a fossilized, ritualized form of
an ancient custom, preserved from times when birds were commonly pinched, not slaughtered,
for food.
30
See Eberhart 2011:29.
31
Though the admonitions in H explicitly refer to the possibility of discontinuing the
sacrificial cult (Lev 26:31), the rituals are often referred to as (OK
eYTaoRD o vL/OT moRDo vL) OL
m WoE TQd
xuX,
as, for example, Exod 27:21; 28:43; 29:9; Lev 16:29–34; 23:14; 24:3, et passim.
11:50:13:06:14 Page 139

Page 140
blood from the location of its collection to the altar. The former assuredly
cannot constitute a praxeme in Israelite rituals, and the latter may or may not
be one––there is dispute in rabbinic literature as to whether conveying the
blood from the location of the slaughter to the vicinity of the altar (HKLWH)
constitutes an integral part of the ritual (HDWBE) or not. Similarly, it has been
suggested often that the “pouring” and “squeezing” of blood under particular
circumstances are not ritual acts per se––that is, they are not praxemes.32
(c) Finally, there remains the question of the value of “negative praxemes,”
events and actions that must not take place during the ritual. For example,
there is some doubt whether the prohibitions LYD diB
vx o in Lev 1:17 and 5:8
Y AL
are praxemes of the /wholeburnt and /purification rituals, respectively. Stated
more precisely, there is some doubt whether “not severing” a /––the two
parts of the body in the case of a wholeburnt offering, the head from the body
in the case of the purification offering––constitutes its own action in the
sacrificial system. In some cases, such negative praxemes are likely to be
consciously avoided, as in the example above, since not tearing asunder a
small bird’s carcass while tearing it open requires some deftness and care.33
Therefore, in a grammar of ritual, such negative events might be included
among the praxemes.
It is often the case that the distinction between praxemes and non-
praxemes is one of degree, not of kind. This appears to be true in all of
the above-mentioned examples––chronological delineation, delineation
according to ritual value, and the identification of negative praxemes. In
other words, from the point of view of these three criteria, as in the case of
discretization, there is a degree of arbitrariness in every determination.
For the purpose of the present study, we have adopted the following
methods: (a) we focus on the heart of the ritual, commencing with the
appearance of the materia sacra at the sanctuary and usually concluding with
its physical annihilation by turning it into smoke on the altar (or, in the case
of blood, by applying it to the altar), by burning it outside the camp, by
consuming it, or by burning the leftovers;34 (b) activities that are undeniably
32
See p. 000 and n. 000.
33
See HMQ 7:8, concerning the purification offering: $DQMB$ TW$Q TWDWBEM WZW.
According to this rabbinic tradition (unlike P, as interpreted here) the former LYD diB
vx o is
Y AL
optional, the latter obligatory (see HMQ 6:22). See also Milgrom, Leviticus 1.387, concerning
the care that must be taken not to extinguish the fire while clearing the ashes.
34
Rabbinic tradition refers to one further alternative in rare and unexpected cases––burial––
on which see HPM 19:11. Blood applied to the altar is considered annihilated for all practical
purposes (see m. Yoma 5:6). One interesting exception is the hide of the wholeburnt, which
becomes the private property of the priest (explicit in P). P is silent concerning the hides of
other sacrificial types (but see b Zeb 103a–b). However, P clearly indicates that the hides of the
other sacrificial types are never turned into smoke on the altar. Since there is no requirement to
burn them, and since they are clearly inedible, it stands to reason that the hides of all other types
of offerings should also become the private property of the priest (eaten purification offering,
11:50:13:06:14 Page 140

Page 141
Praxemics 141
part of the ritual are considered praxemes (where there is disagreement

among scholars, a minimalist approach is taken, unless stated otherwise); and
in particular, (c) negative praxemes are not included.
5.5 PRAXEMIC COMPONENTS IN DETAIL
As noted above, six components may constitute a praxeme: (1) atomact, (2)
agent, (3) object manipulated, (4) target, (5) location, and (6) time-frame.
These six components are defined below.
5.5.1 Atomact
Atomacts are the elementary physical acts constituting a ritual.35 Since the
ritual systems discussed here are for the most part abstracted from texts,
a single verb often corresponds to a single atomact, even though a series
of several physical actions may be implied in one verb. For instance, RYU i vQiH
involves lifting, transporting, and setting down (or tossing) materia sacra,
followed by its subsequent incineration––yet the entire process is here
considered a single atomact, “turn to smoke.” In other cases, P may employ a
single term to denote a sequence of several atomacts. Thus, the scope of H$ om mE
(Lev 16:24) may be used to denote a protracted sequence that includes con-
v iH is sometimes
secutive “slaughter,” “daub,” and “incinerate.” Similarly, BYRiQ
used with the general force of “sacrifice” (Lev 7:38).36 These last two verbs,
however, do not reflect atomacts in the present grammar, which seeks to
present the highest-resolution subdivisions of ritual activity found in the text.
However, as in the relation between zoemics and language,37 the praxemics
of Σ and the language used to describe it correspond only roughly. At times,
several synonyms are used interchangeably to denote the same atomact; con-
versely, the same verb can be used in different contexts to denote different
atomacts. Presumably, many such instances escape our notice, since our
reparation offering) or of the offerer (wellbeing offerings). An exception to this is the hide of a
burnt purification offering, which is burnt outside the camp (Exod 29:14; Lev 4:11–12; 8:17;
16:27). Hubert and Mauss (1964:45) intuit that the desanctification (e.g., of the hide) and
physical annihilation (e.g., by turning to smoke on the altar) are––on an abstract level
––processes of the same type within sacrificial ritual.
35
Some of the atomacts referred to here can be split into even smaller units. As in other uses
of the term atom, the etymology of this word should not be applied literally here.
36
On the use of this verb, see Eberhart 2002:23 and Marx 2005:109–10.
37
See p. 000.
11:50:13:06:14 Page 141

Page 142
perspective is often limited by that of the text. However, despite this

limitation, a number of such cases are discernible.
On the one hand, both QC xmY and v Vx
Pm$o clearly denote a single atomact in P
(“pour,” see Exod 29:12 || Lev 8:15); in one case $oYG i iH and BYRiQ
d v iH denote
the same atomact (“bring forth,” Exod 29:10 RP m
dxH-TA
e T mv
d BRxvQiHvW is fulfilled
in Lev 8:14 TAU m
dxXx
H RP x
d TA a $oGdad W);38 and in one case the participles QRaZo
xYx
(“tossing”) and BYRiQ
vMx denote one and the same atomact (Lev 7:14 and 33
39
respectively). On the other hand, a single verb (such as IT N or BYRiQ
xm v iH) can
be used in different contexts to denote different atomacts, an atomact and a
sequence of atomacts, or an atomact and a hypernym for several atomacts (to
name a few possibilities).40 Even FRxm o$ (“burn”) and RYUi vQiH (“turn to smoke”)
are identical atomacts; they are, however, always praxemically distinct, due to
different “locations”: RYUi vQiH is always paired with “altar,” but FRx$ om never is.41
Thus, the former may have the function of making smoke as a pleasing aroma
to , whereas the latter is a means of disposal.
Furthermore, from a diachronic point of view, a single term may come to
refer to a different atomact, or the atomact may remain unchanged but
assume a new name.42
A phenomenon deserving special attention in this context is the use of the
terms for certain atomacts as generic terms as well. For example, IT xm
N
(“daub”), which in P is used in contrast to HZ miH (“flick”), came to serve as a
d
hypernym in rabbinic literature. Thus, it came to denote several different
38
Note that BYRiQ
v iH and $oYG
di iH are distinct in Lev 2:8. See also $YGH in Mal 1:11; 2:12; 3:3, in
the context of grain offerings. The verbs $oYG i iH and BYRiQ
d v iH may also be synonymous in the more
general sense of both of these verbs, denoting a cluster of atomacts, as at Exod 32:6 (wellbeing
offerings only) and Amos 5:25. Mal 1:8 clearly suggests that the terms are synonymous, but
i iH is not employed in this sense in P.
$oYG
d
39
Similarly, AYBH and BYRiQ
ia v iH are synonymous in some contexts (1:3b || 4:28, etc., where it is
difficult to imagine a praxemic distinction between the verbs), as are RYS iaH and OYRiH a (see
Rendtorff 1967:212), IT xmN and O$ o (Lev 2:1 || 15). A formal comparison of Exod 29:15 and Lev
m
8:18 might lead to the conclusion that XQ xm L and BYRiQv iH may also be synonymous in some
contexts, but it is more likely that the text in Exodus is simply elliptical (with BYRiQ v iH implied).
40
See Rendtorff 1967:90, 155. On IT xmN, see below. Note that BYRiQ v iH may denote, aside from
the atomact “bring forth” (e.g., a layperson bringing a zoeme forth to the courtyard, Lev 1:3; a
priest bringing a zoeme forth to the altar, Lev 1:15; etc.), a general cluster of atomacts (e.g., 1:2 et
passim); and in 7:33 it serves as a hypernym for QRxmZ and RYU i vQiH; that is, it includes tossing blood
(of a wellbeing offering) on the altar and “burning” suet on the altar (the priest does not
“bring forth” the suet––it is the offerer who does this, 7:30 and 3:3, 9, 14 [MT]).
41
See also n. 000.
42
For example, the qal form HC mMm (BH) was replaced in MH by the pi‘el form HC M, as at m.
dm i
Zeb. 6:5 XBZMH RYQ LE HMD TA HCMMW. See, however, Sifra, Weiss 9a OHYN$M HCWMW, and
cf. Finkelstein 1989 2.62 and 3.123 for the variants HCMMW, HCWMW, AYCWMW, etc., and for the
noun HYCM. When the term remains fixed, but the atomact to which it refers changes, one
might say that in a sense, a shift occurs in the grammar of ritual, not in the language used to
describe it; when the opposite happens, one might say that a shift occurs in the natural language,
not in the grammar of the ritual it describes. Of course, both the ritual and the language may
undergo diachronic shifts simultaneously, as when a change in language corresponds to a shift
in the grammar of the ritual (on which see immediately following).
11:50:13:06:14 Page 142

Page 143
Praxemics 143
atomacts pertaining to blood application, including “flick” (HZ miH), “toss”

d
(QRxmZ), “squeeze/drain” (HC i ), and even “pour” (v
dm M Vx
Pm$o)43 ––but to the
exclusion of “daub” itself, which in fact disappeared from the system.44 In
effect it came to mean more generally “apply.” Similarly, QRxmZ became a
hypernym for “toss” (its more natural sense), “flick” (HZ miH), and––somewhat
d
forcedly––“pour” (vVx $o).45
Pm
Strictly speaking, however, this phenomenon belongs not to the grammar
of sacrifice––that is, to the structure of sacrificial activity per se––but to the
“grammar of grammar” of the sacrificial system, a description of the way in
which the ancient ritual texts themselves classify sacrificial activity.46 The
history of such classification deserves a special study. Here it is mentioned
only as a tool for tracking the permutations in terminology, an essential
prerequisite for an accurate praxemic description.
Thus, we see that even in praxemics, where one is forced to draw the most
basic categories from the textual representation of a ritual, natural language
and ritual correspond only roughly. It should be stressed that the above-
mentioned uncertainties render a praxemic description of Σ difficult, but by
no means impossible. The Priestly texts are systematic enough to allow for
many generalizations to be made with a high degree of certainty. For example,
the polyvalence of BYRiQ v iH has hardly been a source of serious confusion
for commentators and translators. More difficult cases, such as IT xmN, will be
discussed in due course.
43
See b Zeb 53a and HMQ 5:7 (flick); m. Tam. 4:1, HMQ 5:6 (toss); b Zeb 53a (drain/
squeeze); b Zeb 37a (pour and toss: ACY HKYP$B INTN$ HQYRZB IYNTNL IYNM “whence [do we
know] concerning those that are to be applied by means of tossing, that he who applies them by
means of pouring has fulfilled his obligation?”) and see, for a striking example in a late text,
HMQ 5:17: RWKBK TXA HNTM IMD ITWN$ XSPBW R$EMB IYDH AWH$ WDML HEWM$H YPM
(“it is known from oral tradition that concerning the tithe [of quadrupeds] and the paschal
offering, one similarly applies their blood by means of a single application, like a firstborn [quad-
ruped offering]”), followed by TXA HNTM IWEU OHM DXA LK OD XSPHW R$EMHW RWKBH
DWSYH DGNK HKYP$B (“the firstborn, the tithe, and the paschal offering––the blood of each of
these requires a single application by means of pouring adjacent to the base [of the altar]”).
44
In the Tannaitic system, the blood of purification offerings is not daubed on the horns of
the altar as in P, but flicked onto them; if the priest misses (sideways) but hits a spot within a
cubit from the edge (IAKM HMAW IAKW HMA), the atomact is considered acceptable. Another
shift took place in this praxeme, namely the possibility of applying the blood to a spot lower
than the horns (but above ARQYSH UWX, which is non-existent in P––see the controversy in the
baraita in b Zeb 53a).
45
For “toss” and “flick” see b Zeb 1:1 et passim. See the striking formulation in Maim.,
HMQ 5:13: TWYZH EB$ OHM DXA LKM OD QRWZ (“he tosses blood from each of them, seven
sprinklings”). For the logical relation between the categories “pour” and “toss,” see b. Pes. 121a
and HMQ 5:17.
46
This phenomenon is not restricted to praxemics, as it is found in other branches of the
grammar, notoriously in the classification of sacrificial types. For instance, the thanksgiving
offering in P is a subtype of the wellbeing offering; whereas in H both are subtypes of XB xeZ (see
Milgrom, Leviticus 1.15, but cf. Schwartz 1999:297). Likewise, in P, the reparation offering
and the purification offering are depicted as two independent types, whereas Philo considers the
former to be a subtype of the latter (see Daniel-Nataf 2000:269 and n. 271).
11:50:13:06:14 Page 143

Page 144
As a basis for the  offered here, a complete list of atomacts is

supplied below, alongside their most common designations in BH. Atomacts
whose praxemic value is questionable in P and atomacts that are attested too
sporadically to be fully incorporated in the grammar are printed in a hollow
font. For the sake of clarity, the order of atomacts here approximates the
chronological order of their occurrence in many of the praxemic sequences
in P. The full table without footnotes is found in rule 2 of the 
chapter of the . As most atomacts have been successfully identified
in previous studies, only problematic cases will be discussed in the notes;
the atomacts pertaining to blood manipulation, which have been a source of
confusion since antiquity, will be discussed in more detail below.
1. Bringing forth
forth/inin47 (AYB
iaH ,AYCivMiH ,$oYG
d v iH)48
i iH ,BYRiQ
49
2. Hand-leaning
(a) Single hand (WoDYmvVMxmS, as at Lev 1:4)
(b) Two hands (WYD mYm YT
a
d$ e . . .v
vo TA VM S, as at Lev 16:21)
xm
3. Immolation
(a) Slaughter (UX xm$o, as at Lev 1:5)50
(b) Pinching (QL xMm, as at Lev 1:15)51
47
This term is more precise than “presenting,” as there is little indication that formal presen-
tation is involved (see the apparent synonymity of BYRiQ v iH and AYB
iaH in Lev 1:3b; 4:23, etc.). It is
thus difficult to determine whether it constitutes an atomact or is merely a physical necessity
(see Gilders 2004:88–9, but cf. 106). The precise manner of “bringing forth” differs according to
the size and nature of the object brought forth: a live zoeme (Lev 1:3b) may be goaded to the
courtyard, a bird may be brought to the altar in one’s hands (1:15), blood may be carried to a
spot by the altar (1:5bα), and a grain offering (2:8) may be transported in a vessel or a bowl to
the officiating priest.
48
No clear system is evident in P concerning the use of these verbs. P may use BYRiQ v iH and
AYBiaH to denote bringing to the court (e.g., 1:3; 2:8a; 4:28) and $oYG i iH, BYRiQ
d v iH, and AYB iaH to
denote bringing materia sacra from one spot in the courtyard to a spot nearer the altar or to the
priest (6:7; 7:30, respectively). In 2:8, $oYG
di iH probably implies placement on the altar, following
“bringing” the grain offering to the priest. AYC iMv iH and BYRiQ
v iH are used in the context of
bringing blood from one spot in the courtyard to the priest standing by the altar (Lev 9:9, 12,
18). If “bringing forth,” is merely a physical necessity and not an atomact, then it is clear why
this act is not referred to systematically in P. Note too that in Lev 9, Aaron is commanded to
officiate by himself (9:7), and yet it is his sons who “bring forth” (iWBRiQ vd W,iWAC
xYx vd
iM W, vv. 9, 12, 18)
xYx
the blood to him, as if this were not an integral part of his duty––in other words, not part of the
praxemic sequence.
49
See Eberhart 2002:25–6 and n. 000 above. Note that LXX is consistent in referring to
(dual) hands in the case of wellbeing offerings (throughout Lev 3) but to a single hand in the
context of wholeburnt and purification offerings (Lev 16:21 is of course not counted here, nor
are instances where hand-placement is performed by several persons, as at Lev 8:14). This can
hardly be coincidental, since ch. 3 is systematic in this respect and contradicts chs 1 and 4. It may
be one of the few cases where LXX consciously reflects actual practice in the Jerusalem Temple.
Tannaitic tradition requires two hands wherever hand-leaning occurs (m. Men. 9:8).
50
See Eberhart 2002:26–7 and the bibliography cited there.
51
See Rendtorff 1985:75. This atomact pertains to birds only (in contrast to slaughtering,
which pertains to quadrupeds only). Note that this atomact achieves both immolation and
beheading. A quadruped’s head is probably severed only after immolation. In the case of
11:50:13:06:14 Page 144

Page 145
Praxemics 145
„
4. Acquisition, or “taking 52
(a) Of liquids
(i) By collecting (supposedly in a vessel)53
(ii) By dipping (a finger) (LB xmU, XQL, as at Lev 4:6, 25)54
xm
(b) Of solids
(i) Extraction (OYRiH a ,RYS iaH)55
56
(ii) Separation (OYRiH a)
5. Dismemberment
(a) Tearing (a bird) (ES d
x$ i o, as at Lev 1:17)
(b) Cutting apart (XT xiN, as at Lev 1:6)57
d
6. Washing (JX x Rm, as at Lev 1:9)58
7. Raising (FYNa , as at Lev 7:30)59
iH
8. Placement
(a) Of liquids
(i) Tossing (QRxmZ, as at Lev 1:5)
(ii) Pouring (vVx $o, QC
Pm xm Y, as at Exod 29:12; Lev 8:15)
(iii) Squeezing/draining (HC mMm, as at Lev 1:15 [in niph‘al ])
(iv) Daubing (IT xmN, as at Lev 4:18)
(v) Flicking (HZ miH, as at Lev 4:17)
d
(b) Of solids
(i) Sprinkling60
quadrupeds offered as wholeburnt offerings, beheading is apparently performed when the body
is cut apart (e.g., Lev 1:12).
52
See Rendtorff 1985:52.
53
See Rendtorff 1985:52. This may be a physical prerequisite, as at Lev 16:15. Since it is
unnamed in P (compare Tannaitic HLBQ and see 2 Chr 29:22), it is possible that the authors did
not consider it an atomact.
54
Rendtorff 1985:157.
55
For the synonymity of these verbs, compare 4:8–10 with vv. 31, 35. On the atomact
HMWRT, and OYRiH a as “set aside, give” rather than “lift up” (reserved for HPWNT), see Nihan
2007:121 n. 54; Milgrom, Leviticus 1.473–81.
56
See Milgrom, Leviticus 1.473–81. In Lev 2:9, OYRiH a may denote extraction and separation
simultaneously. This atomact might pertain also to skinning, inasmuch as it consists of the
“removal” of the hide prior to dismemberment (Eberhart 2002:31–2).
57
See Eberhart 2002:31–2, 61.
58
In Ezekiel, Chronicles, and Tannaitic texts, x
XYD a is used as well (Ezek 40:38; 2 Chr 4:6, m.
iH
Tam. 4:2). On this atomact see Eberhart 2002:37– 8 and below, n. 000.
59
See Nihan 2007:154 n. 6; and Milgrom, Leviticus 1.461–73.
60
The atomact “sprinkling” is apparently distinct from “placing,” see for example Marx
2005:66 (in the context of vegetable offerings). There is no verb in P denoting the atomact alone,
though it is implied in Lev 2:13. P employs XL xMm, with the materia sacra as direct object (as in
English, “to salt the flesh”). In Ezekiel’s formulation (43:24, XL xMe OH e YL
aq E OYN
iH H iWKYL
qdoKx iv W),
$oiHv
“salt” is the direct object and the materia sacra is the indirect object. As noted in , salt
holds a special place in the Israelite sacrificial system, namely somewhere between a jugate and
an additional praxemic requirement. In the praxemic description offered here, salt (like blood)
is the object manipulated and the materia sacra is the target.
11:50:13:06:14 Page 145

Page 146
(ii) Flinging (vVYL iv$oiH, as at Lev 1:16)61

v
(iii) Setting (VRxm E, ITxmN, O$om , as at Lev 1:8; 8:27 (compare Exod
29:24); 9:20; Num 6:19)62
9. Cooking
(a) Boiling (L$adBi )63
od
(b) Roasting (HL mm C, as at Exod 12:9)64
10. Ingestion (LKxmA, as at Lev 6:19)65
11. Incineration (RYU om , as at Lev 1:9, 4:12)66
i vQiH, FRx$
61
See Rendtorff 1985:78 on vVYL iv$oiH, and below on RYU i vQiH.
62
Concerning the difference between vVRxm E and RYUi vQiH, see below. Note that setting wood on
the altar (mentioned in Lev 1:7 as vVRxm E) is not considered an atomact in this study: it is not
required for every offering and appears in 1:7 primarily for literary reasons. Even in 6:1–5, it is
not an integral part of the praxemic sequence of a quadruped wholeburnt offering, but a
procedure required for regular maintenance of the hearth.
63
On this and the following atomact, see Marx 2005:115. Though there is ample evidence
that cooking took place in the courts of temples in ancient Israel (e.g., Lev 6:21; Deut 16:7;
1 Sam 2:13; Ezek 46:20, 24; Zech 14:21; 2 Chr 35:13), this in itself does not justify the inclusion
of “cooking” as an atomact, since it appears to be performed for the convenience of the persons
consuming the flesh and does not appear to be conceived of as mandatory (as it is never
enjoined). It is only due to the injunctions in Exod 29:31; Lev 8:31, and (implicitly) Num 6:19
that “cooking” is here considered an atomact.
64
Concerning baking, only a negative praxeme is found: JM am X HPeAmTa AL o (Lev 6:10); it is
unclear whether LK aA ma T x in the previous verse implies that the semolina must be baked
d TWodCM
(without leavening) or whether the phrase is noted only in order to eliminate the possibility of
leavening. The cooking referred to in Lev 2 takes place before the grain offering is brought to the
courtyard (see v. 8a) and is not treated here.
65
Several studies highlight the alimentary element in sacrificial contexts in general
(Robertson Smith 1894; Detienne and Vernant 1989; see Scheid 2003:94). However, ingestion
usually does not function as an atomact in Σ, as it is often optional (e.g., for wellbeing offerings)
rather than obligatory. However, the fact that in some cases eating is required, not merely
optional (see Lev 10:16–18) implies that it is an atomact in Σ; this is true regardless of the
specific explanation offered for the ritual significance of such consumption (see Rendtorff
1985:318–19).
66
Milgrom, Leviticus 1.161; Eberhart 2011:30; 2002:38–9; Marx 2005:109 See p. 000. The
verb RYU i vQiH is unusual in BH, as it can be followed by a directive (and usually is in P), such as
HXmda
BvM
Zi
dxH RYU i vQiH (or x Xda
BvM
Zi
dxH in Lev 6:8, which either means HX md
aBvM
Zi
dxH or is a graphic corruption
of it; see GKC §90d). Furthermore, even where x Xd
a
BvM
Zd
ix H LE x RYUi vQiH is used (e.g., Lev 4:10), LE x
probably denotes direction (as in x Xda
BvM
Zd
ixH LEx HLmmE*) rather than position (as in LE x DM xm E*
x
Xda
BvM
Zd
ixH). Therefore, one is tempted to translate it as “place,” “toss,” or with another verb
implying direction (depending on the precise manner in which the materia sacra reaches the
altar). However, its etymology (“turn to smoke”) and its usage elsewhere (e.g., Lev 17:6, and the
passive in Lev 6:15) clearly imply that the verb itself connotes not only the placement but also
the subsequent incineration of the materia sacra on the altar. See also Milgrom, Leviticus 1.161;
and Eberhart 2011:30. The choice of language in Lev 6:5 may imply that vVRxm E is used to describe
setting down (large portions of flesh), whereas RYU i vQiH denotes tossing or flinging (smaller
quantities of materia sacra, such as the head of a bird (Lev 1:15), its body (1:17), the token of a
grain offering (2:9), and the suet (3:5, etc.)); see also 1:8–9 (where LK d H could theoretically be
ox
limited to WYE m Rmv i ). However, in Exod 29:18 one finds HX
KiW WodBRvQ md a
BvM
Zd
ixH LYiA
xHm -LK
m -TA
d e Tm RvU
d xvQiHv
W
(see also Lev 8:20–21; 9:13; 2 Chr 13:11).
11:50:13:06:14 Page 146

Page 147
Praxemics 147
5.5.1.1 The Manipulation of Blood

From a grammatical point of view, blood manipulation is by far the most
elaborate ritual activity performed in Σ. In P, five distinct atomacts pertain
to the application of blood. These are denoted by six distinct verbs, two of
which appear to be synonymous: QRxmZ, IT xmN, v
Vx $o and QC
Pm xmY (synonymous), HZ miH,
d
67
and HC mMviN. None of these verbs is rare in BH–– IT N is extremely common––
xm
and all are found outside the ritual texts, with particular reference to the
manipulation of liquids. This fact, along with the fact that the verbs usually
appear within a particular context in the ritual texts, enables one to define the
nature of these acts with a high degree of certainty.68 For example, context
often allows one to infer whether large quantities or a few drops of blood are
used, and the relative locations of agent and target enable one to infer in what
manner the blood is applied.
These five atomacts differ from one another in the quantity of blood
involved and the physical forces applied to the blood. The actions of IT xmN
and HZmiH involve no more than a few drops, whereas QC
d xmY/v
Vx $o and QRxmZ always
Pm
imply a larger quantity. In QC Y/v
xm Vx $o only gravitational force applies, whereas
Pm
miH and QRxmZ entail the priest imparting momentum to the blood.69 The
HZ
d
word HC mMviN, like its qal counterpart HC
mMm,70 may imply squeezing by applying
external force to the body of the zoeme. Only IT xmN, “daub,” implies direct
contact between the agent and the target to which the blood is applied.
A few general comments are required concerning each of these five atom-
acts, particularly in relation to some aspects that have not been noticed or
have been misinterpreted in the past.
1. Daub [small quantity, direct application] (ITxmN)
This atomact is designated by IT x.71 Though this verb has
N, always with LE
xm
countless denotations and shades of meaning, its technical usage for blood
applications in P can be deduced from the contexts in which it appears. It is
always used to designate the application of blood to a very specific spot––the
horns of an altar, or the right earlobe, thumb, and big toe of a person, which
67
This last verb is the niph‘al of HC mMm and takes the blood as grammatical subject.
68
Note that in Exod 12:22 E i iH is used to describe an atomact pertaining to the application
xYG
d
of the blood of the paschal XB xeZ, performed with a bunch of hyssop (BWoZA a ). Furthermore,
according to vv. 24–7, this ritual activity was to be performed not only once (in Egypt) but in
future generations as well. It is noteworthy that the verb E i iH (but not its atomact) recurs in the
xYG
d
Passover Haggadah (referring to both paschal and other XB xeZ-offerings), where it is paired with
the wall of the altar: IWCRL VXBZM RYQ LE OMD EYGY R$A (Goldschmidt 1960:56–7, 126; the
form in the Geniza manuscript in the appendix, p. v, 10a, is OMD EYGY$ etc.).
69
Perhaps this may help account for the interchangeability of QRxmZ and HZ dmiH in Num 19:13, 18,
19, 20.
70
See p. 000.
71
In the text of Ezekiel, LA e is sometimes employed. See Ezek 43:20; 45:19, where no dis-
cernible difference exists between LA e and LE x. Ezekiel is notorious for its lack of distinction
between these two prepositions. See Gilders 2004:226 n. 19.
11:50:13:06:14 Page 147

Page 148
can hardly be aimed at from a distance. It is also distinct from HZmiH and v
d Vx $o,
Pm
often combining with them sequentially (for instance, in Lev 4:7, 18, 34;
16:14–19).
Where this verb occurs in sacrificial contexts in P, it is usually explicitly
stated that the act designated by IT xmN is performed with the agent’s finger
(Exod 29:12; Lev 4:25, 30, 34; 8:15; 9:9), and elsewhere it is practically
essential that there be direct contact between the agent’s finger and the target
(Exod 29:20; Lev 8:23; 14:14, 25), since the area is too small to be aimed at
from a distance. Also, in Lev 4:7, 18, in two cases where the anointed priest
offers a bull (B웧) as a purification offering, it is almost certain that IT xmN
denotes direct application with a finger, since in both cases the priest’s finger
is already stained with blood (see vv. 6, 17) immediately prior to the act
designated by IT xmN.72
It is noteworthy that elsewhere in P, even in contexts that relate to the
application of liquids within a ritual––but never in the context of blood-
application–– ITxmN may imply pouring, either from a cupped hand (Lev 14:18)
or, presumably, from a vessel.73 This is apparently one of several examples
where P takes up a word common in BH and employs it as a technical term
with a narrow, precise denotation in readily identifiable contexts, without
refraining from employing it elsewhere in its general sense.74
It is also noteworthy that, as mentioned (p. 000), this definition of IT xmN
does not hold for rabbinic literature. In rabbinic literature, IT xmN becomes a
hypernym, under which at least HZ dmiH, QRxmZ, v
Vx $o, and even MH HC
Pm i are
dm M
subsumed; hence OYMD ITM is used in the general sense of “blood applica-
tions.” Ironically, this generalized usage of IT N in rabbinic literature is
xm
perfectly in line with the wide semantic field covered by this verb before P
converted it into the technical term “daub.”75
72
Note the absence of LB xmUvW/XQxm W here, in contrast to vv. 25, 30, 34. Of the many passages in
Lv
P where IT xmN is used in the context of blood application, very few lack evidence that direct
application with a finger is implied. In Lev 16:18, direct application with a finger may be
intended but is certainly not stated; in Exod 12:7, the blood is applied with a bunch of hyssop.
Ezekiel is silent concerning the nature of the atomact IT xmN, providing no clear evidence that it
is conceived of as performed with a single finger (as in P), or even that it means “daub” (43:20;
45:19).
73
See Lev 2:1, 15 (where IT xmN appears to be synonymous with QC xmY), and Num 19:17.
74
See nn. 000 and 000 .
75
It is difficult to sustain the claim that such a general sense is already found in P. See
Schwartz 1999:111 n. 49. Note that in Lev 16:14, 18–19, there is no reason to assume that the
same atomacts that apply to the TReod Pd
xK (more precisely, to the two distinct spots that are referred
to as TReod
Pdx
KxH [YN d] LE
aP
v x and TReodPd
xKxH YNviL, see also 15b) also apply to the altar (18–19). The
aP
difference between the application of the blood to the altar (daubing and sprinkling) and the
application of blood to the TReod PdK (only sprinkling, from a distance) may be related to the fact
x
that one cannot imaginably touch the TReod K, whereas both daubing and sprinkling can be done
Pd
x
where the target is the altar. There is, admittedly, one Priestly text where blood-application to
the altar is referred to using IT
xmN as a hypernym, but this is within a unique context, with 
as subject (17:11), as Schwartz himself argues (1999).
11:50:13:06:14 Page 148

Page 149
Praxemics 149
2. Pour [large quantity, through the air, only gravity] (vVx Pm$o, QCxmY)
This atomact is designated by vVx $o (rarely QC
Pm xmY), commonly translated as
“pour”: the translocation of a large quantity of blood (more than a few drops)
from a high spot to a low spot using gravity. The verb QC xm Y is used in this
context only twice in P (Lev 8:15; 9:9). A comparison between Exod 29:12 and
Lev 8:15 may imply that QC xmY is a subcategory of vVx
Pm$o, or––more likely––that
the two are synonymous, at least in this context.76 In the sacrificial system
of P, the object upon which blood is poured is always the base of the
altar (x
Xd
aBvM
Zd
ix H DWoSYv). “Pouring” applies only to the rest77 of the blood of a
quadruped offered as a purification offering after some of its blood has been
applied elsewhere.78
This last fact has given rise to the suggestion that pouring is not essential to
the ritual, being merely a form of disposing leftover materia sacra.79 This
interpretation is perhaps supported by the overtones of the verb vVx $o,80 but
Pm
its synonym QC xmY does not carry such overtones. It is noteworthy that in D
(Deut 12:24, MT vVa$ PmdoiY), “pouring” appears to be an essential part of the
ritual of XBxeZ-type offerings, but only a means of disposal in non-sacrificial
contexts (OYiMmd
d K, “like water,” v. 24).81
x
3. Toss [large quantity, through the air, agent’s momentum + gravity] (QRxmZ)
This atomact is designated by QRxmZ, almost always paired with xXd
a
BvM
Zd
ixH LEx,
and usually accompanied by the adverb BYB S. This implies the application of
im
a relatively large quantity of blood––more than a few drops––through the air
and onto the altar.82
76
Contra Zevit 2001:292, who implies that QC xmY, as “decant,” is distinct from vVx
Pm$o, “pour.”
A discussion of the reasons for this difference in terminology is found in Gilders 2004:218 n. 47.
See Milgrom, Leviticus 1.545–9, concerning the chronological relation between Exod 29 and
Lev 8.
77
“All” (LK
m ) in the formulae d
d HMmdm
D LKm and OD
d mx
d H LKm (Exod 29:12; Lev 4:7, 18, 30, 34)
d
clearly means “all the rest.” See Milgrom, Leviticus 1.238, following Saadiah.
78
See Gilders 2004:116, whose gap-filling leads him to assume (concerning Lev 4:7, 18) that
in these cases all of the blood is to be taken into the shrine, a few drops of it used there, and the
rest taken back out. The formulation in Lev 16:15 suggests that this may in fact be assumed.
79
Contra Zevit 2001:294. Unexpected support for this view is found in the story of Gideon
(Judg 6:19–21), where the hero is required to place the flesh and the bread on a rock but to
“pour out” (vVxPm$o) the liquid broth he has offered to the angel. Accordingly, only the meat and
the bread are consumed by the miraculous fire (v. 21), but no mention is made of the broth. See
Gane 2005:281 n. 49 and Milgrom, Leviticus 1.305 concerning draining as well.
80
Note that LXX consistently reflects this by translating QRxmZ as προσχ#ω and vVx $o as κχ#ω
Pm
in the context of blood application in Exodus and Leviticus.
81
On the phrase “like water,” see Weinfeld 1972:214. LXX neatly reflects this distinction by
employing κχ#ω in the translation of vVx $o in v. 24 (π τ4ν γ1ν κχεετε α.τ A Bδωρ ), but
Pm
προσχ#ω (the equivalent, for the translators of Exodus and Leviticus, for QRxmZ in the context of
blood application) in v. 27 (τ δ" αCµα τν θυσιν σου προσχεε πρ τ4ν βάσιν το!
θυσιαστηρου).
82
This tossing is supposedly achieved with the use of a vessel named QRmvM Zi . Though P does
not mention this vessel in the context of blood applications, it does mention that the chieftains
donated such vessels to the temple, and both a HRmm EvQ and a QRmvM
Zi are said to contain grain
11:50:13:06:14 Page 149

Page 150
Linguistically, the phrase BYBimSxXd

a
BvM
Zd
ixH LE x . . . OD
m QRxmZ could denote either
d
tossing blood onto the upper surface of the altar on several places (perhaps
on its perimeter) or dashing blood against several of its sides.83 Ancient and
modern commentators are in agreement that the phrase denotes the
latter.84 However, two considerations suggest that the previous denotation is
intended.85 First, when P wishes to communicate the application of blood to
the walls of the altar, it states so explicitly, by means of the phrase RYQ i LEx
x
Xda
BvM
Zd
ixH (“upon the wall of the altar,” Lev 1:15; 5:9). By contrast, the phrase
x
Xda
BvM
Zd
ixH LE x is P’s standard formula for denoting the upper surface upon
which materia sacra such as flesh and suet is placed (Lev 1:7; 4:10; 9:13, 17;
see also Ezek 43:18). Second, the realia in several Priestly and non-Priestly
texts suggest that this kind of blood application can be performed from atop
the altar,86 in which case tossing blood against the walls of the altar would be
physically impracticable.
Though the matter is not certain, it appears that in P x Xda
BvM
Zi
dxH LE x QRxmZ (as at
Lev 17:6 MT and Num 18:17) is merely shorthand for BYB imSx Xda
BvM
Zd
ixH LEx QRxmZ.87
offerings (Num 7:13, 19, etc.). In any case, some kind of vessel is clearly necessary for tossing
the blood of quadrupeds, for practical reasons. See Gilders 2004:113, 216 n. 17; and TS 34:7
TWQRZMB [ODH T]A OYSNWK. The precise quantity of blood would depend, among other
factors, on the aperture of this vessel.
83
Contra Rendtorff 1985:52; Gilders 2004:66 and 210 n. 26, the phrase BYB imS . . . LEx may
denote “on (top of)” an object all around, and not only “on the side” of the object all around––
see Ezek 27:11a, 37:2, and the discussion in Meshel 2013a.
84
For a representative sample, see Marx 2005:122–3; Rendtorff 1967:146; 1985:52; Hartley
1992:21; Kornfeld, 1983:14–15; Knierim, 1992:56–7; Milgrom, Leviticus 1.155–6; Staubli
1996:50; Noth 1965:22; Murphy 1874:40; Snaith 1967:30–1; Maier 1994:52; Elliger 1966:35;
Kiuchi 2007:93; and Wenham 1979:54. Gerstenberger 1993:29 notes that the blood is poured
“ringsum an den Altarsockel” (perhaps under the influence of LXX on Lev 7:2, on which see
below). Dohmen 1981 expresses some doubt concerning Deut 12 (where vVx Pm$o is used), but not
concerning x Xd
a
BvM
Zd
ixH LEx QRxmZ; cf. Rütersworden 2011:64, 70–1, who simply considers v VxPm$o and
QRxmZ stylistic variants.
85
See in detail Meshel 2013a.
86
See Milgrom, Leviticus 1.587–8; and Staubli, 1996:83. See also 2 Kgs 16:12–13, where BYB im
S
is absent.
87
Many scholars, perhaps under the influence of the (much more common) x XdaBvM
Zd
ixH LEx
BYBim S, render the phrase x Xd
a
BvMZd
ixH LEx in these two verses as “against the altar.” See Gray 1903:231,
and cf. LXX on Lev 17:6. Note too that the phrases x Xd
aBvM
Zd
ixH-ToNRvQ x (Lev 4:25) and TWoNRvQ
x LE x -LE
x
BYBim Sx Xd
aBvM
Zd
ixH (16:18), which are also found in P and are certainly related to the phrase under
discussion, are clearly interchangeable (see, in particular, Exod 29:12 || Lev 8:15). Since Lev 17:6
and Num 18:17 are attributed to H, whereas the other Priestly verses discussed here are ascribed
to P, it is possible that the difference between the phrase preferred by H and the phrase preferred
by P reflects (1) stylistic preferences of different authors; (2) abbreviation due to the nature of
the texts, as Lev 17:6 and Num 18:17 do not describe the sequential ordering of ritual activity
and are thus not comparable to the other attestations (see also the use of the abbreviated form
OYMim Lv
$dox
H ODx -TA
d e QRadoH adoKxL in P, Lev 7:14); or (3) a legal disagreement between P and H as
Zx IH
to the precise nature of blood-tossing, the former envisioning it as several acts of tossing, the
latter as one act.
11:50:13:06:14 Page 150

Page 151
Praxemics 151
4. Flick [small quantity, through the air, agent’s momentum + gravity] (HZ miH)
d
This atomact is denoted by HZ miH.88 There is no single English term that
d
corresponds to the variety of acts referred to by HZ miH in BH, and the different
d
shades of HZ miH in BH should probably be translated by different English terms
d
according to whether the action is performed with a finger (as at Lev 4:6
et passim), with a branch of hyssop (Num 19:18–19), or with a live bird and
several other materials (Lev 14:6–7, in a non-sacrificial context).89
Since in strictly sacrificial contexts in P, HZ dmiH appears to be performed with
a single finger,90 the term “flick” seems suitable to describe this action.91 It is
possible––but cannot be proven––that even where P does not state so
explicitly, the action is performed with a single finger rather than a few fingers
or a hand, as at Exod 29:21 and Lev 8:11, 30, where the verb is preceded by
XQxmL, “take.”
In one case (in the context of a bird purification offering, Lev 5:9), both
EBd
xvCAe (“finger”) and XQ xmL are absent, but one need not imagine that the
blood application is performed directly from the neck of the bird.92 Since the
priest has just wrung the head of the bird, some of its blood must remain on
at least one of his fingers, perhaps his thumb. This is precisely why XQ xm
LvW is
absent from the verse (it would be incorrect) and why WoEB d
mvCA
ev
dB (“by means of
his finger”) is also absent (it would be superfluous).93
In one Priestly text (Num 19:13, 20), QRxZo ––the passive qal of ZR‰Q ––is
used to describe this physical act, which is also referred to as HZ dmiH elsewhere
(vv. 19, 21). This may imply that already in P, QRxmZ may be used as a hypernym
for “flick” (HZ miH) as well as “toss” (QRxmZ), as already noted concerning rabbinic
d
literature.94
88
Note that in qal, this root denotes “splash” (intransitive), as at Lev 6:20 and 2 Kgs 9:33,
where a small quantity of blood is imagined.
89
See also Rendtorff 1985:157, 161.
90
See Lev 4:6, ODmx
d H-IMi HZ
d
miHv
W OD
md
d x
B WoEB
d
mv e -TA
CA e IHadoKx
H LB xm W; Lev 4:17, -IM
Uv i WoEB
d
mvCAe IHadoKx
H LB
xmUv W
HZ
d
miHv W ODmx
d H; Lev 16:14, WoEB
d
mvCAev
B HZd
miHvW RP
m
dxH OD
xi
d M XQ xm W (in 16:15 this is implicit: H$
Lv momE R$
eoA
qd x
K
RP
mdx
H OD v ). In non-sacrificial contexts in P (Lev 14:6–7, 51; Num 19:18), several other materials
xL
may be dipped in blood before the action of HZ miH is performed.
d
91
With the sense of “throw (off, etc.) with a jerk” (OED, “flick” v. 2). With blood as a direct
object, English “flick” is slightly awkward, but grammatical. The precise physiology of this
movement is unclear. For several rabbinic interpretations, see t. Zeb. 6:10 (DRWYW FUWX) and
HMQ 5:7. In P, it could either refer to a swift movement of the wrist with the index finger
pointed (the motion used when scolding a child), or a fillip––a gesture made by the sudden
forcible straightening of a finger curled up against the thumb.
92
Gilders 2004:120 posits this scenario, perhaps following Tannaitic tradition (t. Zeb. 1:11).
93
miH denotes filliping (see n. 000), then the blood would spread from the thumb to the
If HZ
d
index or middle finger during the act. The use of HC aMmiY in the same verse (rather than v
d Poi iY)
Va$
implies that no vessel is used here. The argument that OD md
d x
B RAmv$od
iNx
HvW implies that the first
atomact of blood application is performed directly from the bird rather than from a finger is
refuted by the fact that similar expressions (though admittedly not OD md
d xB RA mv
$od
iNx
HvW) are found
even where the first atomact is clearly performed with a finger (e.g., Lev 4:17–18).
94
Note, however, that this text is not sacrificial, strictly speaking. It has been argued that both
attestations of QRxZo appear in interpolations of H (Knohl 1995:105). If this is the case, then these
11:50:13:06:14 Page 151

Page 152
5. Squeeze/Drain [any amount of blood; application of physical pressure to

carcass] (HC
mMviN)
This atomact is expressed by HC mM viN. The niph‘al in this case appears to
express the passive of the qal HC mMm, “drain/squeeze.”95 In P it is attested only
twice (Lev 1:15, WoMD
m HC mM Nv; 5:9, HC
viW aMmiY OD
d md
d x
B RA
mv$od
iNx
HvW), in the context of birds
(/). In both cases, the grammatical subject must be the blood, though the
logical subject, and the agent in this grammar, is the officiating priest.96
The presumed active counterpart HC mMm usually means “drain” and not
“squeeze” in BH (as at Isa 51:17; Ezek 23:34; Ps 75:9),97 as in the context of
drinking a liquid to the dregs; but the root MC‰Y is also attested in the
context of squeezing (Judg 6:38),98 and in any case squeezing is probably
involved in the context of / for practical purposes. In MH, the pi‘el HC dm M
i
is presumably used to express the same physical activity.99
This atomact pertains only to bird offerings: in the context of /wholeburnt
it is paired with xXd
a
BvM
Zi
dxH RYQi (Lev 1:15),100 and in the context of /purification
verses may be a case where H obfuscates P’s precise legal terminology by using the same terms
in their common, non-technical sense (see Milgrom 1992). If these are not interpolations of H,
then it appears that P, having coined a technical term, still uses the same term in its general usage
in BH (for a similar phenomenon, see Meshel 2008:227).
95
For this use of niph‘al, see GKC §51f. Compare, for example, the passive qal in XQ d
xTu
dxW
HEoRvP
x
d TYB
d
a H$od i
m AHm (Gen 12:15) with the late BH niph‘al form in RT
a
dvSA
e XQ
xd
m
LTix
d W (Esth 2:8, 16).
96
Contra LXX, in which στραγγιε and καταστραγγιε (respectively) are transitive verbs
whose subject is the priest. Arguably, HC aMmiY in 5:9 may be vocalized as HC
d MiY (contra MT), with
ev
the priest as the grammatical subject. However, LXX is an unreliable witness here, since it also
translates HC mMviN in 1:15 as the transitive verb στραγγιε with the priest as subject, though here
it is clearly a niph‘al, with the blood as grammatical subject. It is impossible to determine
whether LXX had the later form HC xvY (attested in rabbinic literature) or the ancient form
de M
HCeMviY in mind when translating 5:9, and as the evidence from 1:15 implies, the translator may
have had neither: faced with a passive or medio-passive sense in the Hebrew, the translator
simply chose to form an active sentence in Greek.
97
Ps 73:10 is difficult and can hardly be adduced as proof for either view. Perhaps read WCM o Y
with BHS.
98
Note that JM ed W here may be qal passive (like QC
iYx ed W (1 Kgs 22:35) and O$
iYx oe YY
d
ixW (Gen 50:26),
but see GKC §73f) with the dew as grammatical subject, or a transitive qal (at least according to
MT) with Gideon as subject (following RZ xY
dmxW in the same verse).
99
It is possible that according to one Tannaitic tradition, a praxemic distinction is made
between draining the blood of T/ywholeburnt (which is the only atomact pertaining to the blood of
T/ywholeburnt) and draining that of T/y purification. In m. Zeb. 6:4–5, the former is expressed by the
pi‘el HC M (with the priest as subject), whereas the latter is expressed by the hithpa‘el HC
dm i dm x
MTv iH
(with blood as subject) and appears to occur of itself, after the blood is flicked (HZ miH) on the
d
wall of the altar. This implies that after the HZ miH atomact is completed, no further squeezing is
d
necessary––the blood simply trickles down the wall to the base of the altar––so that in effect, the
praxeme squeeze+base has been eliminated from the system. See also b Zeb 64b, and compare
the language in HMQ 6:20 with 7:5. However, Mishnaic literature often refers to draining
the blood of the T/ypurification as YWC
d M as well, and refers to it as involving physical activity on the
part of the priest (e.g., m. Zeb. 6:7).
100
But LXX: κα στραγγιε τ αCµα πρ τ4ν βάσιν το! θυσιαστηρου, reflecting
squeeze+base. It is hardly likely that LXX is preferable to MT here, but the tendency in LXX
(or its Vorlage) to read the word DWSY into the MT even where it is absent is significant (see also
LXX on Lev 7:2; Deut 12:17). See Dion 1987.
11:50:13:06:14 Page 152

Page 153
Praxemics 153
with xXd
aBvM
Zd
ixH DWoSYv (5:9), where it refers to the remains of a purification
offering’s blood after it has been applied elsewhere. This is analogous to
v
Vx $o in the context of quadruped purification offerings (as at Exod 29:12;
Pm
Lev 4:30, etc.).
5.5.2 Agent
The agent is the performer of an atomact. In the grammars of some ritual

systems, non-human agents may be found: in some cultures, animals are
active participants in ritual procedures or are at least perceived as such.
For example, cattle may be expected to offer their consent to their own
immolation by certain gestures of the head.101 On the other hand, according
to Lawson and McCauley’s description of European marital rites, a super-
human being (God) functions as an agent––at least in a roundabout
manner––in every marriage ceremony conducted by an ordained minister.
These “culturally postulated superhuman beings” play an important and
often indispensable role in Lawson and McCauley’s analysis.
In the ancient Israelite sacrificial system, zoemes are not expected to do
anything. The question of ’s participation is somewhat more complex
and theologically laden.102 However, at least according to P, ’s role in
the sacrificial system is hardly part of the sacrificial procedure in a strict sense
of the term, but rather a response to it.103 For this reason,  is not
considered an agent in the present study.
The agent in Σ is one of three types of person: (1) offerer, (2) officiant, or
(3) unspecified third party, including lay assistants––that is, the offerer, the
officiant, or a third party may perform an atomact. Note that from a historical
101
See B.K. Smith 1994:269; Scheid 2003:83.
102
 is undeniably central to the Israelite sacrificial systems as described in P and
elsewhere (note, for instance, the ubiquitous HWHYL x xXXo YN
ixXYRa H$
aoi in P). Furthermore, the
dA
consumption of materia sacra by a divinely ignited fire according to Lev 9:24 implies ’s
active participation in the sacrificial ritual at least once in the mythic past, and most probably
from that point on as well (presumably, this divine fire was preserved, though new coals were
added periodically (see Lev 6:2, 5–6)). However, according to the ritual texts, ’s role
from that point on is limited to accepting or rejecting the offering (this is the meaning of
the technical term HC m Rm; compare HC i , HC
m RvN e Rma o ): the divine fire, which is continually
Y AL
nourished with new wood and perhaps also with new coals (1:6), does not refuse to consume
materials placed on it. In this regard, the interplay between practitioner, fire, and 
somewhat resembles a postal system: the envelope reaches its addressee, who is free to throw
it into the recycling bin, shut his or her ears so as not to listen to the messenger, or learn of the
letter’s contents and ignore them, but the mail delivery system continues to function
indiscriminately.
103
In Tannaitic literature, divine signs are given at the final stage of a (partially) sacrificial
ritual to indicate its success, as at m. Yoma 6:8 and b. Yoma 76b.
11:50:13:06:14 Page 153

Page 154
point of view, some atomacts were eventually assigned to the Levites,104 but
P does not assign such sacrificial tasks to them. These three agents are listed
below, along with some subdivisions:
1. The offerer105
(a) A single offerer (the person may be referred to as WoNB d
m RvQ
m BYRiQ
vM
x H, as
dx
at Num 15:4)
(i) A lay Israelite106
(ii) A chieftain (AY$ N, as at Lev 4:22)107
oi m
(iii) A priest
(1) An ordinary priest (IH adoK, as at Lev 6:16)
(2) The High Priest (x XY$i oM
mx
d H IH
ad H, as at Lev 4:3)
oKx
(b) Joint offerers108
(c) The community (HD mE aHm )109
104
See 2 Chr 30:16; Ezek 44:11 (in 43:19, the offerers are priests, so it comes as no surprise
that they also perform the slaughtering; see Gilders 2004:146). Concerning the lay offerer’s
gradual loss of cultic duties, see Olyan 2000 138–9 n. 81, in Gilders 2004:226 n. 30.
105
The offerer is the person on whose behalf an offering is made. Hubert and Mauss
(1964:ix) use the term sacrifiant, translated into English by Halls as “sacrifier.” Surprisingly, no
specific noun is repeatedly used in BH to denote “offerer.” Instead, the offerer is referred to as
BYRiQvx Y-YK
i OD
d mA m (Lev 1:2), BYRiQvTx -YK
i $oP
d e (2:1), and similar formulas. Compare Tannaitic OYL
eN imEvd
B
(pl. tant.), the yajamāna of the Vedic sacrificial literature, and the Sumerograms EN.SISKUR
(or EN.SISKÚR, EN.SISKUR.SISKUR) used in cuneiform literature.
106
See, for instance, Lev 1:2. P contains no specific term for “lay Israelite” (in contrast to
priests). See Lev 6:16 for a praxemic distinction between a priest’s offering and that of a lay
Israelite. Information in the Priestly literature concerning praxemic distinctions between the
offerings of Israelites and those of non-Israelites is sparse (see Lev 22:25, which is notoriously
ambiguous, and Knohl 1978 concerning post-biblical traditions). The indirect evidence from
the narrative context in which the laws are embedded should be used cautiously: P’s sacrificial
laws are addressed to Moses (and Aaron), to be relayed to the Israelites (sometimes to
the priests); some apply only to the Israelites (as Lev 1:2), while others seem to be more general
(Lev 13:2). However, even in instances of the last case, it is possible that the law applies only
to Israelites, since they alone are addressed (e.g., Lev 12:1–8, particularly v. 2: ½YN d -LA
aB
v e RBd
adx
D
RMo ALa LA ov iY). Characteristic of H is its demand that the offerings of the RG
a Rm$ a (“sojourner”) be
d
identical to those of other Israelites if he chooses to make an offering (Num 15:14). Note,
however, that this passage refers only to voluntary wholeburnt and wellbeing offerings, only to
the RG a (not to other non-Israelites), and only to the fact that the same subordinate jugation
d
applies to the RG a as to the Israelites.
d
107
There is little reason to imagine that in P’s narrative world there should be only twelve
chieftains at any given point in history, as P nowhere implies that every chieftain must be a
“Chief of Tribe.” Ezekiel imagines a single AY$ N, with sacrificial laws that pertain only to him
oi m
(e.g., 45:22; 46:12).
108
For instance, Exodus 29:15, 20–1, 24, where the offerers are priests (on which see below).
109
When the offerer is the Israelite community, the atomacts performed by the offerer may
be omitted or performed by representatives (such as the elders). In Lev 4:15, the elders are
explicitly singled out as representatives of the community to perform at least one atomact
(hand-leaning) on their behalf. This is also implied in OE mmH IB
d m in Lev 9:7, 15 (but see v. 5,
x RvQ
which suggests that the whole community is present). It is reasonable to assume that the elders,
according to the logic of this chapter, consume the flesh of the wellbeing offerings: the flesh does
not go to Moses, who receives only the breast, and the priests do not receive it because their
portion––the right thigh––is turned to smoke on the altar. Contrast this with a later Tannaitic
11:50:13:06:14 Page 154

Page 155
Praxemics 155
2. The officiant (officiating priest)110

(a) An ordinary priest111
(i) Single priest (IH H, as at Lev 1:15)
adoKx
(ii) Several priests ([OYN iH
qd H] IR
oKx oqHAx ½YN
aBd , as at Lev 1:5)
v
(b) The High Priest (xXY$i oMmx
d H IHad H, as at Lev 4:5)
oKx
3. Unspecified party
In BH, anonymous agency is designated either by an active verb in the
third-person masculine singular (such as UX xm W, FRx$
$ov W, Num 19:3, 5)112 or
om v
(more rarely) by a passive verb (such as UX am
$ i Lev 6:18), the grammatical
doT
d
subject being the object manipulated. In this case, the law does not require
that any particular person perform the act, just that the act be performed.113
Even more rarely, it may be expressed by an active verb in the third person
masculine plural (such asiWUX
qv$oiY, Lev 7:2).
Context often implies who the agent is even where grammatical ambiguity
exists, but at times more than one understanding is possible. This has led
to much disagreement and misunderstanding from the earliest translations
down to modern scholarship.114
5.5.3 Object Manipulated
The objects manipulated in Σ are almost invariably animal and non-animal

sacrificial materials. Most zoemes consist of three main categories of materia
sacra:
m , as at Lev 1:5)115
1. Blood (any zoeme, OD
d
halacha, which assigns the flesh of public wellbeing offerings to the priests (m. Zeb. 5:5). Note
that though this mishna is formulated as a general law, its author probably had only one case in
mind (namely, the 2s웧wellbeing referred to in Lev 23:19b).
110
The case where the offerer himself is a priest is treated as a special case in the ,
Praxemics, 7.d).
111
Again, P employs no distinct term for this category. However, in P’s narrative, where the
prototype of the High Priest is Aaron, his sons may serve as the prototypes of lay priesthood. As
a general category, it is created artificially by subtracting 2)b) from 2).
112
The text clearly states that the person who burns the animal and the priest who performs
the blood applications are not the same person (see v. 8, and note the use of WYNmYEavL in v. 5). The
person who slaughters the animal may be a third person (see also the use of WYN mvL in v. 3).
mP
113
This does not mean, of course, that every passive (or every niph‘al ) denotes an action that
can be performed by anyone. Many passive verbs refer to acts that a priest certainly must
perform (such as Lev 6:15b).
114
See, for example, Gilders 2004:62–5; Reventlow 2007:291–5.
115
It is impossible to drain the blood of an animal in its entirety. Therefore, it appears that
the zoemic blood referred to in P is the blood that flows from the wound for the first moments,
while blood is still pumped through the arteries (termed $PNH OD in rabbinic literature; see the
discussion in b. Pes. 65a–b). Only for birds is it required to squeeze blood out from the body,
since this is both practicable and necessary (very little blood is available).
11:50:13:06:14 Page 155

Page 156
2. Suet (quadrupeds only, BL eaX, OYBimLqX, as at Lev 8:25–6)116

(a) The suet covering the entrails (BRed e
QxH-TA e HS e
dxKMvxH BL eaXxH, as at
Lev 3:14)
(b) The suet around the entrails (BRed e
Qx H-LE x R$eoA
q BLea XxH, as at Lev 3:14)
(c) Both kidneys (TYomLdvKx
H YT
a
dv$o, as at Lev 3:15)
(d) The suet around the kidneys, on the sinews (IH eaLq
E R$ oe q
A BLeaXx
H
OYLim
Sd
v H-LE
Kx x R$ A, as at Lev 3:15)
oe q
(e) The caudate lobe (DB aK
mx
d H-LEx TReT Yx , as at Lev 3:15)
edoH
(f) The broad tail (ovines only, HYmL vAxH m , Lev 3:9)
3. Flesh and bones117
(a) Right hind leg (quadrupeds only, IYM id H QWo$o, as at Lev 7:32)118
mYx
(b) Breast (quadrupeds only, HZ m , as at Lev 7:30)
eX
(c) Foreleg (quadrupeds only, E xoRvZ)119
Other body parts are also designated in the ritual laws in the context of
particular ritual activities that pertain to them:
(d) Innards (quadrupeds only, RDeP d, as at Lev 1:8)120
e
(i) Entrails (quadrupeds only, BRee Q, as at Lev 1:9)
(1) Stomach (quadrupeds only, HB Q, as at Deut 18:3)121
ma
(e) Head (any zoeme, $oAR o , as at Lev 1:8, 15)
mvL, as at Deut 18:3)122
(i) Jaws (quadrupeds only, OYiYxX
(f) Shins (quadrupeds only, OYiE K, as at Lev 1:9)123
xRmd
v
116
Already in P, BL eaX is used collectively to include the suet and the other body parts listed
here (e.g., Lev 4:19). See also below on RDeP d. For a discussion of the precise physiological
e
identification of these categories (with illustrations), see Milgrom, Leviticus 1.205–8.
117
All of the portions referred to here, including the breast (on which see Milgrom, Leviticus
1.430) but not the stomach, consist of both muscle and bone.
118
Concerning the physiology of this limb, see Yadin 1983 1.157–9, 169–76 and Milgrom,
Leviticus 1.432 (who notes an apparent error made by Philo, DSL 1.145).
119
In P the foreleg is mentioned only in connection with the S웧 offered by a Nazirite
(Num 6:19). In D (Deut 18:3), the E xoRvZ is mentioned along with the lower jaws (OYiYxX
mvL) and the
stomach (HB maQ) in a context that is apparently sacrificial. Later traditions either dissociated this
law from the sacrificial system, interpreting it as pertaining to profane slaughter (b Hul 132a),
or combined the priestly prebends of P with those of D (TS 20:15–16).
120
The term RDeP d is probably a generic term that overlaps with BL
e eaX but is not identical with
it: BLeaX (“suet”) is a common word in BH that P adopts as a technical term for some of the suet
and several other organs (2.a.–f. in the present list). The word RDeP d simply denotes “innards”
e
(2.a.–e. in this list, as well as other internal organs easily separable from the musculature, such as
the intestines). For a different view, see Lipiński 2008.
121
See n. 000.
122
See n. 000.
123
See Milgrom, Leviticus 1.160 (cf. TS 33:14–15, OYLGR). The reason for singling shins
out is purely instrumental: normally they are not flayed (this is difficult for practical reasons)
and therefore must be washed, since the wool/hair in this area of the body is usually soiled.
Note that the skin on the head is not usually flayed either, and yet there is no requirement to
wash it.
11:50:13:06:14 Page 156

Page 157
Praxemics 157
(g) Crissum (with adjacent feathers; birds only, d

HTmm
CNov
d
B oWTA u,
m RvM
Lev 1:16)124
4. Hide (quadrupeds only, RWoE)125
In two cases, the object manipulated is neither a zoeme nor a jugate, but
another object that functions as the agent of a previous praxeme: the Levites
in Num 8:11, and the officiating priests in Exod 30:19–21 and Lev 16:4. In this
second case, which pertains to a rite preceding the sacrificial procedure,
the praxeme is reflexive, as the agent washes parts of his own body.
5.5.4 Target
The “target” of a praxeme is the place (or person) to or upon which an object
is applied. Whereas all praxemes consist of an atomact and an agent (at least
an unspecified agent) and most contain an object, only a limited number
of praxemes involve a target. The most outstanding and elaborate of these are
the blood-manipulation praxemes. In these, the same object (blood) may
be applied (in different atomacts) to several distinct targets: a number of
spots on the bronze altar and in the sanctuary and, occasionally, a number
of spots on the body of the offerer.
The following is a list of targets in P (BH terms are supplied where
applicable):
1. Inanimate objects
(a) The tent of meeting (DEaoWM LH
eAo , as at Exod 28:43)
(i) The inner sanctum (OY$ i oD
mQ
nx
d H $oD o , Exod 26:33; also called
eQ
$oDeQ
d
oxH, Lev 16:3)
124
Concerning this phrase and its various interpretations, see Rendtorff 1985:76–8; Milgrom,
Leviticus 1.169–70.
125
The hide is listed here separately, since it is not always considered materia sacra in P: it is
not offered with the HL m Eo (Lev 7:8), and since it is not R$ B, it need not be consumed or burned
omd
m
within any time limit (Lev 7:8, 17–18). See also TS 47:13, concerning the possible usage of the
hides of offerings. The hide is burnt (outside the camp) only in the case of burnt purification
offerings, where even the dung is incinerated. It is unclear whether the Priestly law required the
hide of the paschal f웧 to be burnt along with the leftovers of the roasted body (Exod 12:10).
Mishnaic law does not require this, but at least in modern times, Samaritans burn the hides
of their paschal lambs. On Passover night, 2009, I noted that the Samaritan community
(priests and laymen) burned the hides of its paschal lambs. This was done not in the fire of the
altar, where everything else is incinerated––suet sprinkled with salt as well as the leftovers of
the paschal meal––but by laying the hides on the rims of the large pits where the flesh is
to be roasted later. An informed participant gave a purely instrumental reason for this: the
hides would choke the altar fire. Concerning the status of the bones, horns, and hooves of all
sacrificial types aside from the wholeburnt offerings in rabbinic tradition, see Maimonides HQP
10:2 and HPM 19:9. See also HMQ 5:19, concerning the distribution of the hide among a cadre
of priests.
11:50:13:06:14 Page 157

Page 158
(1) The cover of the ark (TReod Pd

xKxH LE x, as at Lev 16:15)
(2) The area before this cover (TReod Pd
x
KxH YNaPviL, as at Lev 16:15)
(ii) The outer shrine ($oDeQ doxH, as at Exod 26:33)126
(1) The veil (TK eRomd
PxH, as at Lev 4:6)127
(2) The incense (“golden”) altar128
(a) The horns (OYM id
d xSxH TReU o vQ XBd
xvMZi TWoNRvQx, Lev 4:7)
(b) The upper surface (where the incense is placed,
BHmd
mZx
H XBd
xvM
Zi LE x = [WYLE], as at Exod 40:27)
(3) Other areas in the shrine, or the tent in general (LH eAo
DEaWoM, as at Lev 16:16)
(b) The outer (“bronze”) altar and its vicinity (T$ eoX
odvNx
H XBd Zi , XB
xvM d
xvM
Zi
HL m , as at Exod 38:30; Lev 4:10)129
m EoH
(i) The horns (TWoNRmvQ, as at Lev 4:30)
(ii) The upper surface (x XdaBvM
Zd
ixH LE x, as at Lev 1:6–7)
(iii) The walls
(1) A single wall (x Xd
aBvM
Zd
ix H RYQ i , as at Lev 1:15)
(2) Several walls (presumably x Xda
BvM
Zd
ixH TWoRYQ i *)130
(iv) The base (x Xda
BvM
Zd
ixH DWoSYv, as at Lev 4:7)
(v) By the altar (x Xda
BvM
Zd
ixH LC eA a , as at 1:16)
2. The Offerer (see Exod 29:20–21 for the following examples)
(a) The body in general (for example, IR oqHA x LEx)
(i) The right earlobe (TYN iM mvH
Yx WoNZ
vAmv ViWNTv)
d
(ii) The right thumb (TYN iMmvH
Yx WoDYm IHedoB)
(iii) The big toe of the right foot (TYN iM mvH
Yx WoLG edoB)
vRx IH
(b) The offerer’s clothing (WYD mGmBd LE
v xv W)
The structure of the altar envisioned in P is simpler and much smaller than
those of Ezekiel, the Temple Scroll, and the later amoraic literature: all of these
are imagined to have had at least one ledge more than P’s altar. P’s bronze
altar was also simpler and smaller than the altar envisioned in Tannaitic
literature, and perhaps the altar in the Herodian temple.131 Therefore, in these
one would expect to find––and often does––several more “targets.” These are
noted below.
126
On the terms $DQH and OY$DQH $DQ, see Milgrom, Leviticus 1:1061, 1065.
127
On the area before this veil, see Gane 2005:72–80 and below, n. 000.
128
Variously called, among other names, TReU o vQ RU
xvQi
MxXda Zi , R$
BvM eoA
q HWHY YN
aPviL R$
oe qAx Xd
a
BvMZd
ixH
DEaoWM LH
eAovB, BH
d md
mZx H XB
dxvMZi , and OYM
id
d x
SxH TReU
o vQ XB
d
xvMZi (see Exod 30:1–10; 40:26; Lev 4:7, 18).
129
Concerning the various parts of this altar, see Zevit 2001:276–314.
130
The form WYT i is found explicitly in Exod 37:26, with reference to the incense altar.
moRYQ
131
See Haran 1962:772. The term DWoSYv may imply the lower extremity of the altar, not
necessarily a ledge as it does in rabbinic literature (m. Mid. 3:1). Therefore, it is possible that
P’s altar was imagined to have had no ledge at the bottom. Note that Exod 27:1–8 does not
state BYB
imS DWoSYv WodL T
m Y$
oi m
EvW* (“make its base all around it”), though it does say WYT m NoRvQ
xTm Y$
oi m
EvW
(“make its horns”).
11:50:13:06:14 Page 158

Page 159
Praxemics 159
5.5.5 Location
The majority of praxemes considered in this study take place within the
one sanctuary described in Priestly law132 ––more precisely, within the area
encompassed by the “hangings” of the courtyard (RC amXeH YEavLxQ) of this
sanctuary. Certain acts must take place at a particular spot or within a par-
ticular area, for instance, in the courtyard (as opposed to the sanctuary, as
at Lev 6:19, DE aoWM LH eA
o RCxqXdB) or on the northern flanks of the bronze altar
x
(HNmP
omCxXdaBvM
Zd
ixHv x, Lev 1:11).133
VReeY LE
There are, however, several acts pertaining to the sacrificial rites that take
place outside these boundaries, such as disposing the ashes in the dumping
grounds outside the camp (Lev 6:4), dispatching the goat into the wilder-
ness (Lev 16), and preparing the ash of a “red heifer” (Num 19:1–10a).
Conceivably, these could be considered sufficiently integral to P’s sacrificial
system to be included in an expanded grammar. Note that the ashes removed
from the courtyard must be disposed of not just anywhere, but in a
“pure place,” designated elsewhere as the dumping grounds of the ashes
(I$
eoD
eH
d xvVe$Peo, Lev 4:12; 6:4), and that both the goat and the red heifer, though
not a IB
d m, are in fact termed TAU
m RvQ m
dx X (Lev 16:5; Num 19:9, 17).134
However, I will not include these instances, since they are not sacrificial
offerings in the strict sense of the term. As mentioned (p. 000), this study
treats sacrificial ritual as beginning at the appearance of materia sacra and
ending at its elimination, and so it refers only to acts that pertain to the
incineration or ingestion of materia sacra.
132
There is much uncertainty as to whether P or H wished to assert that all of the sacrificial
acts deemed applicable “throughout your generations” (OKYTRDL) must take place in a single
location, namely, a unique sanctuary, or in multiple locations (see n. 000). This uncertainty is a
central feature of the debate concerning the relationship between P and H and the centralization
of the cult in Judea. See, for example, Haran 2009:94–101, Milgrom 2000.
133
From a purely grammatical perspective, this could refer to a certain area on the northern
side of the bronze altar’s upper surface, but from context it probably refers to an area to the
north of this altar.
134
See n. 000. The reason the “red cow” is referred to as a TAU m
dxX in Num 19 is probably that
like TAU m
dxX-zoemes, the “red cow” is the source of a sanguine detergent: in the former, the
detergent is the animal’s blood; in the latter, it is the condensed powder subsequently mixed
with water. One might argue that TAU m
dxX in v. 9 refers to the ashes alone, meaning “detergent,”
rather than to the animal, and hence that P did not consider this animal to be a TAU m
dxX; however,
cf. v. 17, TAU m
dxXx
H TP ov , which echoes HRmm
x Ra$ d
PxH TP o in v. 6. Note that here, as elsewhere in
x Ra$
v
the formulae “[sacrificial type] + AW i H,” the grammatical gender of AW i H (at least according to
the Masoretes) agrees not with that of the noun denoting the sacrificial type (in which one
would expect to find consistently AWH i HL m oWE, AW
i H O$
m , and AWH
moA i TAU m
dxX), but with that of
the noun denoting the material offered: AWH i TAU m
dxX when it is TL eSo HPmAaHm TRiY$ E (Lev 5:11,
oi q
i H TAU
AWiH in 12) but AW m
dxX when the material offered is a RP d (Exod 29:14; Lev 4:21) or a
m
i H HL
RYEi$ om (4:24); see also AW m Eo (never AWH i , although HL m Eo is feminine, Lev 1:9, 13, 17),
i H O$
AW m (5:19, in agreement with LYiA
moA x ), and OH a OYAid i (8:28, not because OYA
uLM id i is
uLM
grammatically plural but because OYB im X is grammatically plural). See also Rendtorff 1985:61.
Lq
11:50:13:06:14 Page 159

Page 160
P’s relative reticence with regard to location is striking, particularly in

contrast to later rabbinic literature:135 it appears that most of the sacrificial
praxemes pertaining to the bronze altar may be performed anywhere nearby
or even on the altar.136 Therefore, the “location” component is only rarely
explicitly present, usually in the following forms (BH terms are supplied
where applicable):
1. The sacred precincts
(a) The tent of meeting (DE aoWM LHeAo)
(i) The inner sanctum (OY$ i oD
mQ
nx
d H $oD o , as at Exod 26:33; also
eQ
called $oDeQ
d
oxH, as at Lev 16:2)
(ii) The outer shrine ($oDeQ d
oxH in Exod 26:33)
(b) The outer court (DE aoWM LH eAo XTxeP, as at Lev 1:3)137
d
(i) The altar and its vicinity (for example, the area implied in the
phrase TRam $ovL x
XdaBvM
Zd
ixH-LAe OT m$
d B, Exod 28:43)
voiGd
v
(1) On the altar (for example, the unnamed area implied in
Lev 9:22)138
(2) By the altar (for example, the unnamed area where the
priest is located in Lev 1:15)
(3) The northern flank of the altar (HN mP
omCxXd
aBvM
Zd
ixHv x, as
VReeY LE
at Lev 1:11)
(ii) Temporary dumping grounds for ashes (I$ eoD
deHx OWoQMv , as at
Lev 1:16)
(iii) (Presumably) any other area in the outer court (RWoHU m OWoQMm
within DEaoWM LH eA o RC X, as at Lev 6:4–9)
xq
2. The camp (HN eX
qxM
dxH)
(a) In general (for example, Lev 17:3)
(b) Within a house (DX mAe TYiB B, as at Exod 12:46)
xv
d
i i)
3. Outside the camp (HN eX
qxM
dxL JWXM
(a) In general (for example, Lev 17:3)
(b) Dumping grounds for the ashes (I$ eoD
eH
d xv Peo, as at Lev 4:12)
Ve$
It is clear that the grammar could be simplified by a fusion of the “location”
and “target,” since these components are often interdependent. For instance,
if the target is the horns of the bronze altar, then the location must be
somewhere near (or on) the altar, and if the location is outside the camp, then
the target cannot be any part of the altar. However, these components are
135
Tannaitic literature contains much detailed information on the officiating priest’s
location at every given moment of a sacrificial rite, as well as on which routes he must take to
proceed from one location to another. See, for example, m. Zeb. 5:3; 6:3, 5; Tamid 4:1, and much
of tractate Yoma.
136
See the general references in the Priestly literature to “approaching” the altar, e.g., Exod
28:43; 30:20; Lev 21:23; see also the atomact “toss” p. 000.
137
For the precise denotation of this phrase, see Milgrom, Leviticus 1.147–9.
138
See Meshel 2013a.
11:50:13:06:14 Page 160

Page 161
Praxemics 161
clearly distinct, as in the formulation HM mDvQax Xd

a
BvM
Zd
ixH LCeA a d
HTm oA v
VYL
iv$oiHv
W
I$
eoD
mx
d H OWoQMv -LA
e (“And he shall cast it into the place of the ashes, at the east
side of the altar,” Lev 1:16). In this case, the priest may stand on any side of
the altar (a location), at the spot where he has just immolated the bird, but
he must toss the entrails to a spot designated I$ eoD
eH
d x OWoQMv (“the place of
the ashes”) to the east of the altar (a target).139 As a rule of thumb, it is useful
to think of “location” as the spot where the agent is located and of “target” as
the spot where the object acted upon is situated when the atomact is
completed.
5.5.6 Time-Frame
In certain systems, most notably natural languages, it would be impossible to

imagine a systematic study that does not take into consideration the relative
chronology of units within the system. This is for the simple reason that in all
forms of natural language––even in languages where word order is relatively
“free”––there are instances where the same units, placed in different orders,
form different sequences. Consider, for example, the distinction between tip
and pit on the phonemic level, or between the lions are eating the bears and the
bears are eating the lions on the syntactic level.
In Σ, on the other hand, chronological order plays a much simpler role:
deviation from a required sequential order may (or may not) render a rite
invalid, but there is not a single example in the systems under examination
where two distinct, valid rites are identical in all but their chronological
order.140 A hypothetical example would be that if the blood is tossed and then
139
Note that in this instance, HM mD a (here “on the eastern side thereof,” not “eastwards”)
vQ
does not imply the direction in which the crissum and feathers (d HT
mmCNov
d
B oWTA u ) are to be
m RvM
tossed––in which case the order would have been LA e (x
Xda
BvM
Zd
ixH LC
eAa ) HMmD vQ
adHT mAo vVYL
iv$oiHv
W
I$
eoD
mx
d H OWoQMv ––but rather the side of the altar where the ashes are placed. It is immaterial in P
whether the priest stands to the west of the altar (in which case he would be tossing the entrails
eastwards), or to the northeast of the altar (in which case he would be tossing the entrails
southwards). See m. Zeb. 6:5 and Sifra 9:3, according to which the priest tosses the crissum and
feathers southwestwards from the southeastern corner of the altar (since I$ eoD
eH
d x OWoQMv was
located, according to Tannaitic tradition, to the south of the altar but to the east of its ramp),
and compare also m. Zeb. 6:2. For this use of HM mD a, see Lev 16:14, according to which the priest
vQ
is to flick the blood HM mDvQ
a TReod
Pd
xKxH YN
aPd-LE
v x ––that is, on the ark cover (more precisely, on the
eastern side of this cover). In this case, the priest is most likely standing where he has just
entered, with his back to the veil. For a different view, see Gane 2005:280–4.
140
There is some evidence of special concepts of time that are prevalent in sacrificial
contexts, concepts that appear to be unique to ritual, such as the possibility of retroactive
annulment, for which see possibly Lev 7:18. In rabbinic literature this particular case of
retroactive annulment was eliminated, but other unique aspects of sacrificial time-frame
are found in rabbinic literature as well. For example, according to b Zeb 53a, it is only in the
context of sacrifice that the twenty-four-hour unit called OWoY lasts from sunrise to sunrise rather
than from sunset to sunset. See n. 000.
11:50:13:06:14 Page 161

Page 162
the suet is burned, the offering is a purification offering, whereas if the order
is reversed, the offering becomes a reparation offering. Such a “minimal pair”
of rituals, differing only in the order of events, is not to be found in the
priestly literature.141
5.5.7 General Remarks Concerning Praxemics
Upon examining the combinations of these six praxemic components and

their relation to the other operative categories within Σ, it becomes clear
that there is a high degree of interdependency in the system. We have
already noted the interdependency between “target” and “location,” which is
determined to a great extent by physical necessity. However, the inter-
dependency is often much greater than physical necessity would require. As
an example of the high degree of determination existing within praxemic
sequences, consider the following case: if the target is known to be the horns
of the outer altar and the agent is known to be an ordinary priest,142 one may
deduce several other components of this praxeme:
1. atomact = “daub” (later “flick”),
2. object = blood (OD
dm ),
3. location = “by [or on] the outer altar,” and
4. time frame = preceded by (“dip finger” + “blood”), followed by (“pour”
+ “base”).
Furthermore, in this case one may deduce that the ritual sequence includes
“placement” + “suet,” and that the sacrificial type is a purification offering––
more specifically, a quadruped offered as an eaten purification offering. From
here one may proceed to the zoemic restrictions pertaining to quadrupeds
141
Gane (2005:283) has argued for an elegant example of chronological inversion, where he
portrays the sequence of blood-applications in the outer sanctum on the Day of Atonement as a
mirror-image of the outer sanctum TAUX blood-applications throughout the year. The former
includes a single daubing on the horns of the incense altar, followed by sevenfold sprinkling in
front (east) of the incense altar (i.e., the movement is away from the ark); the latter includes
sevenfold sprinkling “in front of the veil,” which Gane takes to be in front (to the east) of the
incense altar, followed by a single daubing on the horns of the incense altar (i.e., the movement
is towards the ark). According to Gane, this “syntactic” interpretation (his terminology) con-
firms the theory of the two-staged TAUX: throughout the year, the priests thrust, as it were, the
human evils into the maw of the Holy; yet once a year, the High Priest reverses the direction of
his ritual activity and thrusts these same evils back out of the tent. For an evaluation of this
theory, see Meshel 2007. At any rate, the difference between the year-round inner TAUX and
that described in Lev 16 is much greater than simple chronological inversion, as many more
atomacts are found in the latter case (e.g., the use of two zoemes).
142
“Ordinary priest” excludes high priests, and in this case it also excludes the special cases of
prophet-priests, such as Moses (Exod 29:12) and Ezekiel (Ezek 43:20–1, following MT), who
also perform this same atomact in other contexts.
11:50:13:06:14 Page 162

Page 163
Praxemics 163
offered as eaten purification offerings, and deduce that the zoeme must be a
female ovine or caprine, or a male caprine.
What is more, one may then determine what subordinate jugates were
offered along with the zoeme––nothing according to one tradition, 0.1 ephah
of semolina (with oil and salt) + 0.25 hin of wine according to another––as
well as what higher hierarchic level it may be subsumed under, if any (here
several options are open, depending on the agent and on the specific zoeme
involved).
Thus, it is possible to reconstruct complete ritual sequences from only a
few pieces of information if one is acquainted with the grammar of the
sacrificial system under examination.
5.6 GENERATIVITY IN THE GRAMMAR OF SACRIFICE
The generativity of Σ is most clearly evident from the interface of the “agent,”
“object,” and “target” components, as it is possible to create new praxemic
combinations that are not found in P. This process of generating new
combinations is not haphazard, in that it cannot be performed by indis-
criminately combining items from every list: there are restrictions deter-
mining which combinations are acceptable (or “grammatical”) and which are
unacceptable (“ungrammatical”).
This possibility of creating new combinations on the basis of a given inven-
tory and set of rules is referred to as the generativity of a system by analogy
to generativity in the grammars of natural languages. Let us examine one
example of the generativity of Σ, drawn from the manipulation of zoemic
blood.
As noted above, five distinct physical motions make up the basic inventory
of blood-application atomacts in P: “daub,” “toss,” “pour,” “flick,” and
“drain” (usually denoted in P as IT xmN, QRxmZ, v
Vx $o, HZ
Pm miH, and HC
d mMm, respectively).
Together with diverse targets (such as “the horns of the bronze altar” or “the
right earlobe of a person healed from scale disease”), three types of agents
(“offerer,” “officiating priest,” “unspecified party,” with subcategories),
and an adverb modifying the manner in which the action is carried out
(BYBS, literally “around”),143 these can combine to create a large number of
praxemes.
For the sake of simplification, we will limit ourselves to the blood
manipulation that takes place on the inner (“golden”) and outer (“bronze”)
altars. Thus, there are only six targets where materia sacra may be placed: the
143
In P (in contrast to later traditions) the occasional absence of the adverb is probably a
form of shorthand. However, already in P this adverb indicates that some atomacts, but not all,
include iterative activity. See n. 000 and n. 000.
11:50:13:06:14 Page 163

Page 164
horns of the inner altar, its upper surface (where incense is placed), the horns
of the outer altar, its upper surface, its wall(s), and its base.
Again for the sake of simplification, let us stipulate that the grammar
consists solely of these five atomacts and these six targets. We will further
suppose that the “agent” and “object manipulated” components are fixed in
P (for instance, that in each and every praxeme they are “single priest” and
“blood” respectively) and that the “location” and “time-frame” components
are irrelevant. Let us also imagine that P contains no zoemic, jugational, or
hierarchic rules, and that only one rule applies in the grammar of P: any
combination of a single atomact and a single target is grammatical.
As defined here, the sacrificial system in P may be of the form Σ[A, T, R]
(read: “a system Σ based on a set of atomacts A, a set of targets T, and a set of
rules R”). The grammar of this system would thus consist of five atomacts
{a1, a2, . . . , a5}, six targets {t1, t2, . . . , t6}, and a single rule {r1} determining
which combinations of atomacts and targets are grammatical and which are
ungrammatical. In particular, the grammar of Σ[A, T, R] would be of the
following form:
A = {daub, toss, pour, flick, drain}
T = {horns of inner altar, upper surface of inner altar, horns of outer altar, upper
surface of outer altar, wall(s) of outer altar, base of outer altar}
R = {a + t is grammatical }
Rule r1 reads: any member of set A may combine with any member t of set T.
Since r1 places no restrictions on the grammaticality of a+t combinations,
and since sets A and T consist of five atomacts and six targets respectively, this
grammar of Σ would allow for the generation of 30 grammatical praxemes
(5 atomacts times 6 targets). However, it so happens that in P, only the
following combinations are attested:
1. daub+horns of bronze altar (HL m EoH
m XBd
xvM
Zi TNoRvQx LEx + ITxmN, in the case of
a quadruped offered as an eaten purification offering, as at Lev 4:25)
2. daub+horns of incense altar (OYM id
d x
SxH TReUo vQ XB
d
xvM
Zi TWoNRvQx LE x + IT N, in
xm
the case of burnt purification offerings, as at Lev 4:7a)
3. drain+base of bronze altar (x Xd
a
BvM
Zd
ixH DWoSYv LAe + HCmMviN, in the case of /
purification, Lev 5:9aβ)
4. drain+wall of bronze altar (x Xd
a
BvM
Zd
ixH RYQi LE x + HC mMviN, in the case of /
wholeburnt, Lev 1:15)
5. flick+bronze altar 144 (x Xd
a
BvM
Zd
ixH LEx + HZ
miH, as in Lev 16:19, “WYL
d mmE”)
144
It is difficult to determine whether x Xd
a
BvM
Zd
ixH LE
x HZmiH denotes the flicking of blood on the
d
altar in general or onto its upper surface. As is well known, only one altar is mentioned in
Lev 16, and it is located in the court (but see m. Yoma 5:5). It is possible that Lev 16:1–28 is one
of the oldest strata in P, and that therefore it does not assume the existence of an incense altar
(see Milgrom, Leviticus 1.1062–3).
11:50:13:06:14 Page 164

Page 165
Praxemics 165
6. flick+wall of bronze altar (x

XdaBvM
Zd
ixH RYQ
i LEx + HZmiH, Lev 5:9aα)
d
7. pour+base (HL m EoH
m XBd
xvM e + Vx
Zi DWoSYv LA v $o, as at Lev 4:7b (see also Exod
Pm
29:12b))
8. toss a few times+upper surface of bronze altar 145 (x Xda
BvM
Zd
ixH LEx + QRxmZ
BYB
imS, in the cases of quadrupeds offered as wholeburnt, reparation, and
wellbeing offerings; perhaps also toss once+upper surface of bronze altar,
in P and non-P material)146
One more combination is found in D, namely:
9. pour+upper surface of altar (x
Xd
aBvM
Zd
ix x +v
H LE Vx $o, Deut 12:27, MT).147
Pm
Though the evidence is extremely sparse, it appears from this one example
that Deut 12:27 reflects a sacrificial grammar similar to that of P, inasmuch as
the single blood-manipulation praxeme documented in this verse introduces
neither a new atomact nor a new target. It is derivable from the hypothetical
grammar of Σ[A, T, R] proposed above, in which r1 is operative, despite the
fact that it is not documented in any of the texts from which that hypothetical
grammar was abstracted.
There are, however, twenty-one remaining combinations that ought to
be grammatical according to rule r1 but are nevertheless absent from penta-
teuchal law (thirty theoretically possible combinations minus eight in P and
one in Deut 12). A small number of these may be ruled out for practical
reasons (for instance, toss+horns of altar may be difficult to carry out),148 but
this is not true in most cases. There is no practical reason why one could not
daub blood on the base or walls of the altar, toss blood onto its base, flick
blood onto its horns, or drain a bird’s blood onto its upper surface, to name a
few possibilities. It thus emerges that a large number of possible combinations
have not been employed, either because they are deemed inadmissible or,
perhaps, because the needs of the sacrificial system envisaged in P are not so
varied as to require more combinations than those evidenced in the text.
However, if the system is generative, one would expect later texts composed
by authors who were exposed to Σ to include new combinations not attested
145
Contrary to scholarly consensus, this is the meaning of BYB imS ...xXda
BvM
Zd
ix x QRxmZ; see
H LE
p. 000.
146
For instance, Num 18:17; 2 Kgs 16:13. This combination would not differ from toss a few
times+upper surface of bronze altar if one considers the adverb BYB imS to be implicit in these
texts.
147
One cannot speak of a “bronze” altar in D, since D supplies no information about the
composition of the one and only legitimate altar referred to. However, the altar in question,
termed VYHLA HWHY XBZM, is obviously a large one on which flesh is offered, and it thus
corresponds to P’s bronze altar. Concerning LXX on this verse, see below.
148
The horns of the bronze altar may have been rather small (their measurements are not
given in P), so that it would be hard to aim at them from a distance. In rabbinic literature, by
contrast, the horns of the outer altar are envisioned to be large: they are perfect cubes one cubic
cubit in volume each (m. Mid. 3:1).
11:50:13:06:14 Page 165

Page 166
in the pentateuchal texts. In fact, several of the remaining twenty-one

combinations do appear in post-P material:
10. flick+upper surface of incense altar (x
Xd
aBvM
Zd
id
xL$
eo oWRH
mv
d x + HZ
U LE d i )149
mYH
This atomact is found in rabbinic tradition in the context of the blood-
application of burnt purification offerings offered on the Day of Atonement.
11. flick+horns of bronze altar (x
Xd
aBvM
Zd
ix x + HZ
H TWoNRvQ d i)
mYH
This praxeme is found in rabbinic literature, where it replaces P’s daub+
horns, applying to cases where P uses the term x
Xd
aBvM
Zd
ix x + IT
H TNoRvQ xmN.150
12. toss (several times)+base (BYBS XBZMH DWSY + QRZ)
This praxeme is reflected in the Septuagint in Lev 7:2 in the context of
S웧reparation: κα τ αCµα προσχεε π τ4ν βάσιν το! θυσιαστηρου κκλ7
(“and toss the blood on the base of the altar, round about”, contrast MT).
It is probably reflected several times in the Temple Scroll, as at 34:7–8. In
this passage, the praxeme is mentioned in the context of a B웧wholeburnt offering
but is formulated as a general regulation––that is, it does not refer to a
specific occasion.151 The same praxeme recurs in the blood application
of the eaten purification offering in TS 23:13–14,152 along with other
149
m. Yoma 5:6 (vocalization following MS Kaufmann, including dageš in the U), where WRHU
denotes “its upper surface.” Admittedly, this praxeme came into being partially as the result
of the rabbinic identification of HWHY-YN aPviL R$
eoAq xXd
a
BvM
Zd
ixH (the altar that is before ) in
Lev 16 with the inner altar. However, the following instruction, DWSY LE VPW$ HYH ODH YRY$W
IWCYXH XBZM L$ YBREM (“the rest of the blood he would pour out on the western ledge of
the outer altar”), clearly suggests that the origin of this atomact was not the result of automatic
transference of the rites in Lev 16 to the inner altar.
150
See p. 000. The following qualifications should be taken into consideration (see b Zeb
53a): (1) according to the prevailing rabbinic tradition (Rabbi Judah the Patriarch contra
R. Eleazar son of R. Simeon), this need not be performed on the horns themselves but may be
performed on the upper half of each edge on a spot below the horn; and (2) according to
one Tannaitic tradition, it is permissible to apply the blood within a cubit of the line defined in
(1)––note that some confusion can be avoided if one recognizes that in rabbinic literature
the term IRQ denotes both a horn of an altar and a vertical edge of the altar (sometimes also
TYWZ). For a clear case of the latter usage, see m. Zeb. 6:2.
151
The passage reads : XBZMH DWSY LE WTWA OYQRWZW TWQRZMB [ODH T]A OYSNWK WYHYW
BYBS. See also TS 22:5, which mentions WQRZW, apparently in the context of 14 s웧wellbeing and
14 S웧wellbeing on the Festival of Oil (RHCYH DEWM). In this case, it is difficult to ascertain where
the blood was tossed, since the text is not completely preserved. See also 52:21, which requires,
for every (private) wellbeing offering, that HLWEH XBZM DWSY LE WMD TA WQRZW (“they shall
toss its blood on the base of the altar of wholeburnt offerings”).
152
The text reads BYBS XBZMH TRZE [D]WSY LE WMD TA QRZW (literally, “he shall toss its
blood on the bas[e] of the ledge of the altar, round about”). It appears that XBZMH DWSY and
XBZMH TRZE DWSY are synonymous––perhaps P’s DWoSYv and Ezekiel’s HRmmE Zq (Ezek 43:13–17, 20)
were joined by the author of TS as a hendiadys in construct form. (This is logically the case if the
“ledge” in question is the only one, or the lowest.) Thus, if such sparse data is not misleading, it
appears that in TS, the difference between the blood manipulation of wholeburnt and well-
being offerings on the one hand and that of purification offerings on the other is as follows: the
11:50:13:06:14 Page 166

Page 167
Praxemics 167
praxemes.153 It is uncertain whether the evidence from TS reflects pour

(several times)+base or toss (several times)+base, since the author of TS did
not distinguish linguistically between QRxmZ and vVx $o, so that the two are clearly
Pm
synonymous in the Scroll.154 In either case, it is a new praxeme not found in P.
13. toss (single tossing)+base (XBZMH DWSY + QRZ)
This praxeme is evidenced in LXX on Deut 12:27.155 It is also found in TS
52:21 in the context of (private) wellbeing offerings: DWSY LE WMD TA WQRZW
HLWEH XBZM.156 It is uncertain whether one should actually posit a praxemic
distinction between this and the previous praxeme.157
former consists of the single praxeme, toss+base (with BYBS explicit only concerning whole-
burnt offerings). The latter consists of several praxemes, depending on whether it is a burnt
purification offering or an eaten purification offering. If it is burnt, the blood manipulation
consists of two praxemes, daub+4 horns # toss+4 corners of ledge (HRZE; see 16:14–18, LE QWRZY
XBZMH TRZE TWNP EBRA, in the context of B웧purification). If it is an eaten purification offering, its
blood manipulation consists of three praxemes, daub+4 horns # daub+4 corners of ledge # toss
(several times)+base.
153
Though the text of TS 23:10–14 is not fully preserved, context implies that the rules
pertain to the eaten purification offering. The immediate context refers to a public 웧purification,
which from a hierarchic point of view is part of the larger wholeburnt+1 complex HUM TLWE
HDWHY (“the wholeburnt+1 offering of the tribe of Judah”), offered on the first day of the Festival
of Wood (OYCEH DEWM).
154
The evidence for this is unequivocal: in 11QTb (11Q20) 1–2 (Qimron 1996:25), one finds
the very same praxeme as in the following verses preserved in 11QTa (11Q19) 16:17 (discussed
in the previous note, XBZMH TRZE TWNP EBRA LE QWRZY, “he shall toss upon the four corners of
the ledge of the altar”), but here it is referred to as TRZE TWNP EBRA LE BYBS WKWP$Y ODH IM
XBZMH (“from the blood they should pour ‘roundabout’, on the four corners of the ledge of
the altar”; see also Qimron’s first note on p. 26, according to which the letter $ and a few
other unintelligible letters in 16:1 in 11Q19 may be traces of WKWP$Y). It is almost certain that
the same atomact referred to here as WKWP$Y is the atomact referred to as QWRZY in ll. 16–17
(the difference in number notwithstanding) because the text says so explicitly in ll. 14–15
([I]W$YARH RPL H$E R$AK, “as he did with regard to the first bull”).
155
The text reads τ δ" αCµα τν θυσιν σου προσχεε πρ τ4ν βάσιν το! θυσι-
αστηρου, “but the blood of your XB xeZ-type offerings you shall toss [see below] to the base of the
altar.” MT reads m VYH
eoLfA HWHY XB d
xvM
Zi LE xv PmdodY m
Va$ VYX
em BvZ-OD xWv. Note that LXX differs from MT in
two respects, aside from the passive-to-active shift. (1) It uses προσχ#ω, LXX’s consistent
equivalent of QRxmZ (“toss”) in sacrificial contexts––not κχ#ω, its equivalent for vVx $o (“pour”).
Pm
(2) As in Lev 7:2, it inserts “the base of” before “the altar.” Both of these differences reflect the
insertion of concepts from P into the text of D, but the result does not accord with the law in P,
which uses x Xd
aBvM
Zd
ixH LEx + QRxmZ, not x
Xd
aBvM
Zd
ixH DWoSYv + QRxmZ, in the context of XB xeZ. See also Dion 1987.
156
In Gen. Apoc. 10:15 (Machiela 2009:52), Noah describes himself as having applied the
blood of what must be wholeburnt offerings to the base of the altar (HLWK(?) L IWHMDADWSY
TRUQA AXBDM LE IWHR$B LK[W T]D$AW AXBDM (“. . . all(?) their blood to the base of the
altar and poured [it], and all their flesh I turned to smoke on the altar”). While this verse
appears be another example of toss+base (with the use D$A in the sense of “toss”), it is more
likely that the praxeme itself is not new––it reflects the pentateuchal praxeme (7) pour+base (see
the author’s use of D$A in 6.19, AYLYPN WD$A YD AMD, “the blood spilled by the Nephilim”)––
only its association with wholeburnt offerings is innovative.
157
As noted in n. 000, it is unnecessary to assume a strict praxemic distinction between
praxemes (12) and (13) on the sole basis of the adverb BYB imS, since this may be the result of
abbreviation. Note, however, that rabbinic literature creates a clear distinction between single
tossing and several tossings+BYBS, on which see praxeme (14).
11:50:13:06:14 Page 167

Page 168
14. toss (single tossing)+wall (RYQ* + TXA HQYRZ QRZ)

This praxeme is found in several Tannaitic traditions. It is apparently implied
in the formula DWSYH DGNK TXA HQYRZ WQRWZ (m. Pes. 5:6; “he performs a
single tossing of [the blood on a spot that is] adjacent to the base”) in the
context of the paschal f웧. In this passage, DWSYH DGNK does not denote “on
the (ledge at the) base of the altar,” but rather “on a spot on the altar’s wall,
directly above the ledge that is at its base.”158 Thus, one finds in rabbinic
tradition a clear praxemic distinction between toss (single tossing) and toss
(several times).159 Accordingly, a praxemic differentiation was made between
manipulating the blood of first-born offerings, of animal tithe offerings, and
of the paschal f웧 (XSPW R$EM RWKB) on the one hand, and manipulating
that of wholeburnt, wellbeing, and reparation offerings on the other.
Tannaitic tradition requires that the blood of the former be tossed only once,
on one of the four walls of the altar, whereas the blood of the latter is to be
tossed twice, on two diagonally opposite edges of the altar (TWNTM YT$
EBRA IH$, m. Zeb 5–8 et passim).
15. pour+(single) wall (RYQ* + TXA HKYP$ VP$)
This praxeme is found in some Tannaitic traditions, concerning the paschal
f웧, where it is clearly distinct from toss (single tossing)+wall.160 It is most
probably a purely imaginary praxeme, not practiced in any historical period.
158
According to a Tannaitic tradition, the ledge at the foot of the altar, termed DWSY
(“foundation,” as in P), does not surround the entire altar, but only its northern and western
flanks and part of its western and southern flanks (one cubit each, or all but one cubit each,
depending on the interpretation of m. Mid. 3:1). The formula DWSYH DGNK must mean “opposite
the ledge” and not “on the ledge” for three reasons: (1) “on the ledge” would be referred to
as DWSYH LE, see m. Zeb. 5:1–2 et passim; (2) DGNK carries the same sense with reference
to the horns (“adjacent to the horns, at a spot on the same vertical line as the horns”) in a very
similar context (b Zeb 53a: IRQ DGNK in contradistinction from Tannaitic IRQ L$ HPWGB);
(3) otherwise it is almost impossible to understand the use of -$ DBLBW in m. Zeb. 5:8
(DWSYH DGNK ITY$ DBLBW TXA HNTM IWEU IMDW). Note also the anomalous TXA HQYRZ WQRZ
XBZMH YBG LE in m. Pes. 5:8, which must mean “on the wall,” not “on the upper surface”
(despite the misleading YBG LE). Unexpected support of this interpretation is found in the
Passover Haggadah (see n. 000), where one finds IWCRL VXBZM RYQ LE OMD EYGY R$A with
reference to the paschal f웧 offerings as well as the wellbeing offerings offered along with it.
Note that the manner of tossing the blood of the paschal f웧 discussed here is according to the
Tannaitic opinion of R. Jose the Galilean, b Zeb 37a–b.
159
This distinction is based in part on an ultra-literal reading of the biblical text: where the
adverb BYB imS is lacking (as at Num 18:17), rabbinic tradition deduces that only a single tossing
is required; where BYB S is found (e.g., Lev 3:2 et passim), more than one tossing is required. See
im
the midrash halacha referred to in b Zeb 56b.
160
See b. Pes. 121a and b Zeb 37 a–b, as well as the commentaries ad loc. It appears that
R. Akiba and R. Ishmael agree that the blood of the paschal f웧 is to be poured on the wall of the
altar, and that the blood of other wellbeing offerings is to be tossed on the walls of the altar.
The difference is that, according to R. Akiba, if one pours (rather than tosses) the blood of a
wellbeing offering, the ritual is invalidated, whereas R. Ishmael opines that it is not invalidated.
Both agree that if the blood of the paschal f웧 is tossed, the ritual is invalidated.
11:50:13:06:14 Page 168

Page 169
Praxemics 169
Note that from a practical point of view this praxeme is rather awkward
(though not impossible): because “pour” (VP$), unlike “toss” (QRZ), in this
context implies the application of blood using only the force of gravity, the
priest would essentially have to place the vessel against the wall and then tilt it,
in order to pour the blood onto the wall.
16. toss (several tossings)+walls (XBZMH TWRYQ* + QRZ)
This praxeme is found in the Aramaic Document of Levi 8:1, AR$T IYDAB
HXBDM YLTWK LE AMD QRZML (“then you may begin to toss the blood on
the walls of the altar”). According to the immediate context, the author
appears to refer to at least all quadruped wholeburnt offerings.161 It is also
found in rabbinic traditions, where it appears in the form of two tossings
onto two diagonally opposite edges of the altar (EBRA IH$ TWNTM YT$,
e.g., m. Zeb. 5:4 et passim).162
Several of these praxemes appear to have been rooted in Israelite practice,
such as toss (several tossings)+walls, which is evidenced in the Aramaic
Document of Levi as well as in rabbinic literature. Others, like pour+wall, may
be purely imaginary, as noted previously. It should be stressed, however, that
praxemes (9)–(16) are all innovations of later traditions and are not found in
P (praxeme (9) is treated here alongside the later praxemes, though its dating
relative to praxemes (1)–(8) is debated). And yet all are based on the same
limited inventory of five atomacts, six targets, and one rule, which has been
abstracted from P.
On the other hand, certain constraints appear to be active in Σ, limiting the
degree of innovation. These constraints become most clearly apparent when
one considers the large number of atomacts that do not appear anywhere
in post-biblical literature, though they are perfectly practicable. A sense of the
possibilities can be gathered by comparing the rituals described here with
161
The development of this praxeme entails an interpretation of P’s x Xd
aBvM
Zd
ixH LEx as “on the
walls of the altar.” Therefore, one might conjecture that the author of ADL believed that the
same praxeme should be found in the application of the blood of reparation and wellbeing
offerings, which P similarly refers to as BYB im
Sx Xda
BvM
Zd
ix x + QRxmZ. Note that the Greek version of
H LE
ADL reads ττε α*ρξ: κατασπ#νδειν τ αCµα π τν τοχον το! θυσιαστηρου (“then you
shall begin to pour down the blood upon the wall of the altar”). Thus, it differs from the original
Aramaic in two respects. First, the blood is depicted as a libation (this is the connotation of
σπ#νδω) offered in a downwards motion (the force of κατα-). Second, it is apparently depicted
as being applied to a single wall (τοχον)––though one must entertain the possibility that in
this case the singular “wall” may still imply all four sides. This is not the result of simple mis-
translation––it probably reflects certain ritual presuppositions of the translator and may be
another example of praxeme (15).
162
See Meshel 2013a. Tannaitic literature consistently interprets P’s BYBimSxXd
a
BvM
Zd
ix x + QRxmZ as
H LE
tossing blood on the walls of the altar, not on its upper surface. This type of blood manipulation
pertains to wholeburnt, wellbeing, and reparation offerings (m. Zeb. 5:4–7 et passim).
11:50:13:06:14 Page 169

Page 170
blood rituals found in non-Israelite contexts, in non-Priestly Israelite texts,

and even P itself, in rituals that are not strictly sacrificial.163
Thus, both historical and imaginary forms can be viewed as realizations
of the same grammar of Σ, abstracted and internalized by the authors of
these later texts, who were exposed to Σ through their acquaintance with
pentateuchal law. Note that, whereas it is possible that some of these authors
were exploiting the grammatical potential of Σ consciously, this process need
not be conscious.
Admittedly, the generativity demonstrated here is rather rudimentary,
particularly because thirty praxemes (= five atomacts × six targets) is not
very many.164 Furthermore, some of these praxemes––but certainly not all––
evolved as byproducts of biblical interpretation, replacing previous
praxemes that fell into disuse. For example, praxeme (11) resulted from an
interpretation of ITxmN in this context as “flick,” and praxeme (16) evolved as a
reinterpretation of xXda
BvM
Zi
dxH LEx QRxmZ.
However, if one takes into consideration the fact that often––already in P––
two or more praxemes pertaining to blood-manipulation are combined to
create protracted ritual sequences (such as daub+horns of inner altar followed
by pour+base of outer altar in the eaten purification offerings, and flick+wall
followed by drain+base in the case of /purification), the number of potential
sequences is considerably higher.165 Since repetition of praxemes (such as
flick#daub#flick, Lev 16:14–19) is allowed, the number of grammatical
sequences is potentially raised indefinitely.
5.7 THE METHOD OF COMPOSING

GRAMMATICAL RULES
By now it may have become apparent that the most difficult part of the
praxemic description is the identification of the set of rules R that applies
in Σ. As we shall see presently, it is more difficult, in every chapter of the
163
See for example Wilke 2010:240 (atomact: mix [with rice]); Saraogi 73 n. 29 (atomact:
suck); Butler 2010:509 (atomact: smear); Georgourdi 1989:190 (atomact: trace; atomact: finger-
print); see also Exod 12:22 (atomact: asperse, by means of hyssop; non-P); Lev 14:4–7, 50–1
(dipping animal and non-animal materials in blood, aspersion); see also Num 19:18, where the
material is not zoemic blood.
164
The number may actually be higher, if one considers the distinction “single tossing”/
“several tossings” to be present already in P (see n. 000). One could thus argue that a similar
distinction inheres in P with regard to other atomacts as well.
165
See, for example, b Zeb 37a: DWSYL OYMD ITM IYNWEU$ OYMDH LK LE DMYL (“this
teaches that all blood must be applied [also] to the base”), rendering the sequence toss+
wall #pour+base grammatical (as at m. Tam. 4:1). See, however, Rashi ad loc., who argues that
this only applies if some blood remains in the vessel after the tossing.
11:50:13:06:14 Page 170

Page 171
Praxemics 171
, to abstract such rules from the texts than to identify the
inventories.
In order to demonstrate why this is so, let us return for a moment to the
state of our knowledge before we began to consider praxemes (9)–(16).
Based on the data culled from P, we were led to abstract five atomacts and six
targets. However, it was discovered that only nine of their thirty possible
combinations are attested in biblical law, praxemes (1)–(8) in P and praxeme
(9) in Deuteronomy 12. This may lead one to any of the following hypotheses
concerning R:
1. The other twenty-one combinations are physically impossible or
improbable. For example, toss+horns may be difficult to perform,166
since it may be hard to apply the blood to a precise spot from a distance.
This explanation is of course insufficient, since the vast majority of
combinations are physically possible.
2. Some of the praxemes predicted by r1 do not occur because certain
restrictions apply, such as
r2: a+wall is grammatical iff (= if and only if) a = drain
Read: the combination of an atomact (“a”) with the target “wall” is
grammatical if and only if the atomact is “drain”
r3: toss+t is grammatical iff t = upper surface
Read: the combination of the atomact “toss” with a target (“t”) is
grammatical if and only if the target is “upper surface”
r4: pour+t is ungrammatical if t = horns
Read: the combination of the atomact “pour” with the target “horns” is
ungrammatical
3. The set R has been identified correctly as consisting entirely of r1, but
the biblical texts either fail to preserve evidence of other ritual sequences
or Israelite society for some reason did not produce certain ritual
sequences that would have been considered grammatically licit.
We have just noted that option (1) cannot be weighed seriously as the
basis for a comprehensive theory. However, based on praxemes (1–8), it is
still impossible to choose between options (2) and (3). In fact, both are
conceivably correct explanations, inasmuch as each may apply only some of
the time: certain combinations may be unattested because they are ungram-
matical, whereas other combinations may be unattested although they are
grammatical. This leads to a fourth hypothesis:
166
See p. 000.
11:50:13:06:14 Page 171

Page 172
4. Certain restrictions on R are at play, determining which unattested

praxemes are grammatical and which are not, but these restrictions
cannot be determined with certainty given the textual evidence available
to us.
Let us assume, for the sake of argument, that the truth lies in (4), and
attempt to refine R accordingly. The textual data may lend itself to two
disparate approaches to the composition of a praxemic rules. One may pro-
ceed from the most stringent rule––namely, only combinations (1)–(8) are
grammatical––and modify the grammar every time a new piece of evidence
is discovered. The product of this modus operandi could hardly be termed
a grammar: in the first place, it would consist of a simple list of praxemes
and would not be generative. Furthermore, the discovery of each and every
one of praxemes (9)–(16) would lead to the falsification and subsequent
reformulation of this so-called grammar.
The other approach, which I have followed, is to compose the most lenient
rule r1 and proceed by refining it with further stipulations. These restrictions
must be conjectured, unless explicit in the text, and r2–r4 are examples of such
restrictions, composed on the basis of praxemes (1)–(8). According to a
Popperian model, every restriction is to a certain extent educated guesswork
arrived at through intuition and in accordance with the attested data.
However, each restriction is easily testable and falsifiable.167
Note that exposure to the pentateuchal laws entails exposure to Σ.
Therefore, Hellenistic Jewish authors, the Aramaic Document of Levi,
Jubilees, the Temple Scroll, and the vast corpus of rabbinic literature were all
exposed to (almost) the same grammar.168 If the theory presented here is true
and a grammar of Σ exists, then one might expect at least several of the
remaining twenty-two praxemes to appear in these later texts. As we have
seen, examples (9)–(16) validate this general theory. However, they do not
validate the rules r2–r4 that have been postulated: praxemes (14–15) invalidate
r2, and praxemes (12), (13), (14), and (16) invalidate rule r3.
By contrast, r4, which indicates that pour+t is ungrammatical if t = horns,
can be said to have stood the test of time. At the very least, it has not
been refuted by the large corpus of post-biblical sacrificial texts that
have been discovered thus far, and it is here hypothesized that the praxeme
x
Xda
BvM
Zd
ix x+v
H TWoNRvQ Vx
Pm$o will not appear in further textual findings. Though
167
Popper 1969.
168
On some occasions, as we have seen, one must also take into account the possibility that
different texts lay before different authors––for instance, some Hellenistic authors may have
been exposed to LXX rather than MT (see above on Deut 12:27).
11:50:13:06:14 Page 172

Page 173
Praxemics 173
the logic of this restriction eludes us, within Σ there is apparently something
inherently ungrammatical about pouring blood on horns.169
While here the process of conjecturing general rules and subsequently
refining them by means of specific restrictions has been demonstrated from
Praxemics, since Praxemics as an operative category most clearly evidences
this process, the same process pertains to all categories of Σ. Note, how-
ever, that an important difference exists between Praxemics and the other
operative categories of Σ. The preceding chapters treat operative categories
that are of the following forms:
[Z, R] where Z is an inventory of zoemes and R a set of rules,
[J, R] where J is an inventory of jugates and R a set of rules, and
[H, R] where H is a set of hierarchic levels and R a set of rules.
These brief inventories and rules are amenable to particularly concise for-
mulation, and are found––alongside examples of the interrelation between
various categories––in the first three chapters of the enclosed   Σ.
Praxemics, on the other hand, is of the form [A, G, O, T, L, R] where A, G,
O, T, L, and R are sets of a relatively large number of atomacts, agents, objects,
targets, locations, and rules, respectively.
Many of the specific restrictions within Praxemics pertain to the inter-
relation between the various components of each praxeme, and thus
necessitate a more detailed grammatical description than the other operative
categories. Such a description is found in the final chapter of the enclosed
, alongside references to the dependence of praxemic rules upon
other operative categories within Σ.
169
Note that reference to meaning in such instances––claiming that such combinations are
not attested in Σ merely because they would mean something that is either superfluous or is out
of place in Israelite ritual––is problematic for two reasons. First, it assumes that the rituals in Σ
evolved through a more-or-less conscious act of invention, and that the driving force in the
formation of specific ritual details is meaning-related. Second, it depends on a problematic
equation of knowledge of a ritual’s origin with knowledge of its meaning. See Staal 1989:324.
11:50:13:06:14 Page 173

Page 174
Meaning
The complex issue of ritual meaning in general, and of meaning in Israelite

sacrifice in particular, has been treated extensively in the literature.1 In fact,
many previous studies of Israelite sacrificial ritual have centered primarily
on meaning, at times at the expense of praxemic details. Unfortunately, even
studies attentive to praxemic detail appear to envision the study of Σ’s formal
properties as a mere preparatory stage for an investigation of its rituals’
meanings, thus obscuring the value of a purely formal analysis of ritual. This
state of affairs has become widely acknowledged only since the publications
of Frits Staal on South Indian ritual.
While a systematic discussion of ritual meaning in general, or even of
Israelite sacrificial meaning, is beyond the scope of the present study, I shall
briefly discuss some of the primary schools of thought on the matter, as well
as their respective merits and shortcomings, before examining the meaning-
related evidence within Σ itself and drawing some preliminary conclusions
about meaning from a specific sacrificial example.
6.1 GRAMMARS WITHOUT MEANINGS
Elaborate theories of ritual are not found in the oldest sources of the ancient
Near East, and in this regard, P is no exception. In contrast to its elaborate
detail on praxemic minutiae, P says little about the meaning of its rituals.2 It
1
See Michaels 2006; Kreinath 2006; 429–70; Severi 583–93; Quack 2010. Note, too, the caveat
expressed in Milbank 1995. In his pioneering application of general systems theory to ritual and
of “ritual syntax” to Israelite and other ancient Near Eastern rituals (Gane 2004), Roy Gane
presents a partially formalized conception of meaning in biblical rituals within the framework
of “cognitive task analysis.” See also Gilders 2011; Klawans 2008, 2011.
2
This absence is even more striking in light of the wealth of praxemically detailed texts
in ancient Near Eastern literature from the first millennia . See Watts 2007:181, and below,
n. 000.
11:50:13:06:14 Page 174

Page 175
Meaning 175
is tempting to consider P’s silence on meaning as intentional, a reflection of

its authors’ main concern with orthopraxy rather than an underlying theory
of the meanings of required praxis. In some cases, P seems deliberately to
avoid reference to meaning altogether, limiting the text to a strict praxemic
description. Consider Lev 6:2a–bβ:
RQ H-DE
edoBx x HL
mvL
Yd
xxH-LK
mx
d Xd
a
BvM
Zd
ixH-LE
x HDv WoM-LE
mQ x HL
m EoH
m AWH
i HL
m EoH
m TRxoWT
d TAZo
This is the instruction for the wholeburnt offering (literally, HL m , “that

m EoH
which ascends”)––that is, that which ascends upon its hearth upon the altar
all night until morning.3
Here, the author seems to avoid the immediate etymology of HL m (“the
m EoH
ascender”), which elsewhere in biblical literature is invariably associated
with this type of offering, calling to mind the ascension of the smoke up to
the deity.4 P nevertheless avoids this “meaning”-related etymology, offering
instead a purely praxemic etymology. Rather than ascend to please  in
His abode, the “ascender” ascends a mere five feet above the surface of the
ground, where it spends the night burning on its hearth. We shall return
briefly to P’s conception of the meaning of the rituals it describes. At present,
it should suffice to acknowledge that P’s interests lie elsewhere and that most
texts reflecting Σ have precious little to say about the meanings of rituals.
Even though we may conjecture that the texts’ authors did attribute some
meaning to the rituals they describe, we have hardly any access to their
theories on the matter. We are therefore confined to composing a grammar
for a system whose meaning––however defined––certainly eludes us with
regard to its particulars.
Throughout the construction of this grammar of Σ, we have endeavored to
avoid the issue of meaning as much as possible, instead centering primarily
on the formal aspects of Σ. This avoidance is in line with the Chomskian
tradition in linguistics,5 according to which the formal properties of
generative grammars are formulated without reference to semantics:
A great deal of effort has been expended in attempting to answer the question:
“How can you construct a grammar with no appeal to meaning?” The question
itself, however, is wrongly put, since the implication that obviously one can
construct a grammar with appeal to meaning is totally unsupported.6
3
Admittedly, the grammar of the entire verse is problematic. For the general syntactic struc-
ture of this verse (and for the reading d Dv oWM) see Milgrom, Leviticus 1.383–4. For alternative
HmQ
readings, see Rendtorff 1985:225, 234 and Bar-Asher 2013:9–10.
4
E.g., Judg 13:16–23 and particularly v. 20. On the two denotations of the hiph‘il HL mfEeH in
sacrificial contexts, see Marx 2005:20. In the present context, ascension is clearly implied, since
the qal is not used in BH in the sense of “offer as a wholeburnt offering” or “be brought forth.”
5
In this respect, the tradition goes back to Bloomfield; see Houben 2010:34 n. 23.
6
Chomsky 1957:93.
11:50:13:06:14 Page 175

Page 176
From the immediate context of this statement,7 it is evident that Chomsky

here equates meaning with semantics. As a demonstration of his point, it is
possible to recognize that a sentence is grammatical, even to know how to
pluralize or passivize it, without having recourse to its semantics:
Grobulous rumpalos besquipple trupulent skrannies. A trupulent skranny is
besquippled by a grobulous rumpalo.
In order to effect two transformations for the above sentence (plural to
singular; active to passive), it is not necessary to know the meaning of
“rumpalo” or of any of the other unfamiliar words. All one needs are
language-specific morphosyntactic rules about word order, subject–verb
agreement, and the like, properties that are independent of semantics.
Note that the identification of categories like noun, verb, and present tense,
essential for the application of the above transformations, does not require
semantic interpretation. However, the identification of these categories
does require some recourse to the range of meanings of nouns like
“rumpalo” and verbs like “besquipple,” and cannot usually be done on
purely phonetic, morphological, or syntactic grounds.8 Let us now examine a
sacrificial situation in Σ, one similar enough to the above sentence to assist in
delineating our territory in a discussion of ritual meaning.
Consider the act of slaughtering an S웧, tossing its blood on the upper
surface of the bronze altar, and burning its suet there. One need not know
anything about the meaning––or lack of meaning––of these acts in order to
deduce the subsequent and preceding sequences. The action must be part of a
reparation offering because its suet is burnt on the altar, and so its flesh is the
perquisite of the officiating priest. However, one cannot know all of this
without also knowing that the zoeme in question is offered as an S웧, not as
an 웧 or . Praxemic evidence alone could not provide this information:
from observing the aforementioned praxemic sequence acted upon a ram,
one could deduce that this animal is offered either as an S웧, in which case it is
a reparation offering, or as a , in which case it is a wellbeing offering
(another possibility is 웧, in which case it would be a wholeburnt offering,
but this must be ruled out in our case where only the suet is burnt on the
altar).9
Furthermore, if one were to determine that the ram in question is actually
offered in fulfillment of a requirement stipulating a ––in which case the
7
Chomsky 1957:93: “I am not acquainted with any detailed attempt to develop the theory of
grammatical structure in partially semantic terms”; see also p. 94.
8
In this example sentence, perhaps some telltale morphemes assist the reader in identifying
verbs and nouns, such as the plural “-s” affixed to “rumpalo,” or the prefix “be-” of a transitive
verb “besquipple.” However, such part-of-speech earmarks need not be present.
9
The assumption is that in P there is no distinction between 웧wholeburnt and wholeburnt, and
therefore that they constitute a single zoeme here. A wholeburnt, by contrast, need not be slaugh-
tered on the altar’s northern flank in P.
11:50:13:06:14 Page 176

Page 177
Meaning 177
only possible conclusion is that it is a wellbeing offering––other grammatical

properties, such as whether it must be consumed within a single day or two
days, and whether it entails coordinate grain offerings or not, are not yet
known. In P, this depends on the incentive for offering it: if it is a standard
votive or volitional offering, it can be consumed within two days and is
accompanied by subordinate jugates only; if it is a thanksgiving offering, it
must be consumed within a single day and entails (at least) coordinate grain
offerings. This information too cannot be deduced from praxemics alone.10
Is the ability to distinguish between “reparation offering,” “votive/
volitional offering,” and “thanksgiving offering” equivalent to knowledge
of meaning? And if so, is this meaning the “meaning of sacrifice”? A clear
definition of “meaning” in ritual systems is essential for understanding the
relationship between the formal properties of sacrifice and the non-sacrificial
world in which sacrifices take place.
6.2 THE MEANINGS OF “MEANING”
Though there is some disagreement concerning the term “meaning” in

linguistics, there does exist a tradition of scholarship that attempts to
define the word and to provide commonly accepted technical terminology,
such as “extension,” “reference,” “sense,” and “analytic” versus “synthetic”
propositions.11 By contrast, in non-linguistic systems, there is no such
tradition, and scholars from such diverse fields as semiology, anthropology,
ritual studies, and theology use the term rather freely.12 Thus, a vast body of
scholarship about the meaning(s) of various rituals exists, and yet there is no
agreed-upon definition of what might constitute meaning in a ritual system.
Although the investigation of ritual meaning cannot rightfully progress with-
out some conception of the import of “meaning,” many previous attempts
to pin down the precise denotation of this term in the field of ritual only
demonstrate the exasperation of theoreticians. It has been variously suggested
10
Several other examples are found where praxemic evidence alone is insufficient for
determining which sacrificial rite is taking place. For example, the offering of a lamb as an 웧
for an Israelite woman’s volitional wholeburnt offering and the daily public wholeburnt offering
of a lamb as an s웧 may be similar in their basic praxemics in the rabbinic system, since in both
cases there is no hand-leaning (m. Men. 5:7; 9:6–7). The rabbinic system is also replete with
anomalous situations where the praxemics are of one type but the offering is really of another
type: this is designated by formulas such as HLWE O$L HLWE H$EMK HA$E, where the
offering is really a TAUX (m. Zeb. 7:1, 2, et passim).
11
This tradition carries back at least to Gottlob Frege. See Frege 1975:116–28 (first published
in 1892).
12
See Sperber 1975a:7–14.
11:50:13:06:14 Page 177

Page 178
that (1) ritual systems are semiotic systems that signify;13 (2) ritual systems
are symbolic systems but, since symbolic systems are not semiological
systems, they do not signify;14 (3) ritual systems have semantics, but in a
broader sense of “semantics” than commonly meant in traditional lin-
guistics;15 and (4) ritual systems are semiological systems but lack semantics,
and “hence” they are meaningless.16 This rather confusing sample is indica-
tive of the wide spectrum of senses in which “meaning” is used in relation to
ritual systems.17
As noted above, a systematic discussion of ritual meaning in general or
Israelite sacrificial meaning in particular is well beyond the scope of the
present study. However, it is worthwhile to review a few of the alternatives
that have been suggested with particular regard to Σ. We have grouped several
of the most important approaches into three groups, according to criteria that
relate specifically to sacrificial ritual: (1) task-oriented approaches, which
view ritual as a system of “nuts and bolts,” and in which meaning is roughly
equivalent to desired effect; (2) a cluster of approaches referred to here as
“functionalist,” which attempt to explain how some rituals come into being
and what social functions they may fulfill, and in which meaning is used
approximately in the sense of raison d’être; and (3) a collection of approaches
that use meaning in the sense of representation. These schools of thought, it
should be noted, are not mutually exclusive. The presentation of these
approaches will be followed by an approach that views ritual as meaningless.
13
See the discussion of Lévi-Strauss’s view in Sperber 1975a:51–84, as well as Penner 1985,
who appears to agree with Sperber on this issue. It should be noted that de Saussure (1959:16,
quoted in Sperber 1975a:51), who coined the term “sémiologie,” intended it to include
“symbolic rites, polite formulas, military signs, etc.” Lévi-Strauss, true to the original meaning
assigned to the term, continually asserts that anthropology should be a branch of semiology.
14
This is Sperber’s main thesis in Rethinking Symbolism. Of all these theories, Sperber’s alone
approaches a systematic analysis of the problem of meaning in ritual. Although Sperber’s
treatise deals with symbolism in general, it is clear that ritual systems are subsumed under
“symbolic systems” throughout the first four chapters of Rethinking Symbolism (see Sperber
1975a:110).
15
Lawson and McCauley advocate this view. See particularly Lawson and McCauley
1990:141, 148–9. It also appears to be Penner’s own view, as formulated in the conclusion of
“Language, Ritual and Meaning” (Penner 1985).
16
Staal’s view as understood by Penner (1985:11). The double quotes on “hence” indicate
what Penner takes to be a false inference of Staal’s. Penner (pp. 2–3) claims that Staal’s view is
simply wrong, since (in Penner’s view) the presence of syntax alone is enough to constitute
meaning (though not reference).
17
Handelman (1990:62) has noted that “meaning [is] likely the greatest bugbear of scholars
who deal in symbols.” Additionally, like many bogies, as soon as it is exposed to the light of day,
it tends to vanish and can no longer be subjected to critical research. For example, when Sperber
(1975a:53, 85–113) attempts to shed light on the semiotics of symbolic systems (including ritual
systems), he discovers that symbolic systems are not semiotic at all. And when he turns to
elucidate the meaning of symbols, he finds that symbols simply do not mean.
11:50:13:06:14 Page 178

Page 179
Meaning 179
6.2.1 The Task-Oriented Approach
Let us begin our review of ritual meaning interpretations with the least
theory-laden approach, which we will call “task-oriented.” These approaches
conceive of a ritual’s performance as physically manipulating the nuts and
bolts of a mechanism in order to attain a desired effect.18 Accordingly, the
meaning of a ritual in this view is the purpose that the ritual fulfills, just as
mopping a floor “means” that the floor will be clean.
Milgrom’s description of the role of blood as purging the altar’s horns of
their miasma exemplifies the task-oriented approach to ritual meaning: “If
you have daubed this sanguine detergent on the horns of the altar, that means
that the altar is purged from the miasma that clung to it.”19 Here, the term
“mean” is appropriate inasmuch as daubing the blood effects the purgation.
The blood does not necessarily signify or denote anything (though it may also
do this); it simply is a detergent. Note that the accuracy of the task-oriented
interpretation does not depend on the validity of its ontology. That is,
crediting the task-oriented view does not necessitate believing that rituals
actually accomplished their practitioners’ goals, or even believing in the ritual
mechanism or cogs (for instance, that there is a deity named  and that
there exists miasma). Rather, it only depends on the assumption that some of
the ancient ritualists accepted this ontology and considered its mechanism
effective.
As noted above, the sacrificial texts of P focus on rituals’ practical aspects
(mostly on praxemics), often to the exclusion of other aspects.20 What little
evidence does surface in P suggests that its authors considered the offerings
efficacious in a task-oriented sense. This much is hinted at in such statements
as WYL
mmE RPa
dxKvL oWL HC
m RvN
iWv HL
m EoH
m $oAR
o LE mv
x WoDY VMxm
SvW (Lev 1:4, “he shall lay
his hand upon the head of the wholeburnt offering, that it may be acceptable
on his behalf, in expiation for him”),21 as well as in the recurrent reference
to wellbeing and wholeburnt offerings as H$ A (“a food gift”) and x
edoi XYRa
HWHYLxx i (Lev 1:13, 17, “a pleasing odor for ”).22
XoXYN
18
See the remarks in Lévi-Strauss 1981:670 and below, n. 000.
19
Milgrom 1976b. Note that the general argument here is not dependent on the specifics of
Milgrom’s analysis and is equally applicable to alternative reconstructions of the process
of purgation, e.g., Gane 2008.
20
See Eberhart 2011:119. Marx 2003 rightfully speaks of reconstructing the meaning of
sacrifice according to P from “scraps of theology.”
21
See also Lev 16:21, which implies physical transfer.
22
Note that the term H$ i is usually in the construct form (H$
edoA a
dAoi ), technically denoting “a
food-gift of pleasant odor.” The terms H$A and XXYN XYR are commonly used with reference to
wellbeing and wholeburnt offerings, but the term H$ i is used once with reference to the
edoA
reparation offering (Lev 7:5), and the formula HWHYL x xXXo YN
ixXYRa is used once with reference
to the purification offering (4:31). Despite the fact that P often uses “buffer language”––for
11:50:13:06:14 Page 179

Page 180
A rather simplistic approach to the question of meaning in P is the

argument that its authors believed that the system worked mechanistically,
and that if their ontological presuppositions were erroneous, then they were
simply in error over the mechanistic effectiveness of sacrifice. One might then
argue, with Pascal, that in circles that no longer believed P’s machine to work
in a literal sense, it was necessary to conjure symbolic interpretations of the
details of P in order to salvage its truth (now “Truth”).23 However, symbolic
interpretation is often found side by side with task-oriented interpretation:
profound spiritual Truth is sought even where the literal, carnal truth is not
denied.
As an interpretative framework, the task-oriented approach fails in a
fundamental way. The view that sacrifices are thought to have desired effects
such as purging miasma or supplying a deity with sweet-smelling substances
only asserts that there is a mechanism; it does not supply a theory at all in
terms of how this mechanism’s effect is attained.24 In this respect, the task-
oriented view is not really an explanation of ritual at all, and we must look
elsewhere in the quest for a clearly laid groundwork for a theory of meaning
in ritual.
6.2.2 Functionalist Approaches
Another group of approaches has gained substantial popularity in the

past few decades (“hundreds of studies,” according to Michaels),25 a group
which we call “functionalist” in a very wide sense, including not only the
Functionalism of Radcliffe-Brown and Malinowski but also the evolutionist
approach of Burkert (and at times J.Z. Smith), the historical-sociological
approach of Girard, and the cultural-materialistic approach of Marvin
Harris.26
example, it refers to the offerings as producing a sweet smell before  but does not depict
 as ingesting the flesh––it does not deny that the sweet smell has an effect on the deity.
Concerning the differences between P, H, and Ezekiel in their use of buffer language, see
Knohl 1995:128–37. On one rare occasion, H probably expresses a mechanistic conception of
blood-application praxemes: RP i H OD
a
dxKvY $oP
eN
d
ed
xB AW mx
d H-YK
i (Lev 17:11; see Schwartz 1999:108–16).
d
23
Pensées §687: “When the word of God, which is true, is literally false, it is spiritually true”
(“Quand la parole de Dieu, qui est véritable, est fausse littéralement, elle est vraie spirituelle-
ment”). Note that Pascal’s second example in this passage (xXXo YN
ixXYRa) pertains specifically to the
Israelite sacrificial system.
24
In the Greco-Roman world, the first attempt to account for the manner in which sacrificial
ritual is efficacious by means of a comprehensive theory is found in Iamblichus’ De mysteriis
(see Clarke et al. 2003, particularly Book V).
25
Michaels 2006:248: “Basically, these studies use biological, functionalist, or religionist
(confessional) arguments.” See also Gruenwald 2003:199.
26
See Malinowski 1948; Radcliffe-Brown 1952; Evans-Pritchard 1956; Burkert 1987; J.Z.
Smith 1987; Girard 1977, 1987; and Harris 1974. On the “Functionalists” (here capitalized) and
the wider category of functionalists (here lowercase) see Goody 1973.
11:50:13:06:14 Page 180

Page 181
Meaning 181
Despite the fundamental differences between them, these approaches share

two important properties that distinguish them from the task-oriented
approach: they are all etic approaches,27 and while they may be true or false,
illuminating or obfuscating, their explanatory power is limited to under-
standing only the most general contours of sacrifice. They cannot––and
they do not attempt to––explain fine distinctions between ritual sequences,
but rather refer often to sacrifice as a general phenomenon that requires
explanation.28 In other words, these theories may explain why humanoids
started sacrificing animals in the first place, why they persisted in doing so,
why Neolithic humans usually selected domestic animals as their victims, and
perhaps why certain functionaries received different portions of the flesh, but
they cannot enter the fine resolution, sought in this study, that is concerned
with such issues as “why are toss+base, toss+upper surface, toss+base, and
toss+wall grammatical, whereas pour+horns is ungrammatical?”
Furthermore, while there is much to be said in favor of this approach,
it is uncertain whether the results it supplies have anything to do with
“meaning.”29 In other words, ritual systems may fulfill social needs or
reflect an evolutionary development, and it is even possible that their form
is determined to some extent by environmental factors. However, these
observations lie in aspects external to rituals’ grammar, and it is questionable
whether the fulfillment of such needs constitutes the meaning of a ritual
system.30 For instance, it is only by stretching any sense of “meaning” to an
extreme that one may argue that a wellbeing offering among the Semites
“means” social cohesion with members of the clan (Durkheim); or that
killing a helpless domestic beast “means” a “transferal of aggression”
(Girard), “giving expression to a biologically determined urge” (Burkert),
or “playing out domestication” (J. Z. Smith). These interpretations resemble
explaining a language’s syntax by bespeaking the social benefits of fluency
or literacy––hardly anything to do with linguistic meaning. Many of these
explanations deal entirely in generalities; none pertains to the “meaning” of
sacrifice in any rigorous sense; and, most importantly, they fail to account for
the wide variety of praxemes in ritual systems. 31
27
Girard’s theory is an extreme example of this (1987:78): “In order to be genuine, in
order to exist as a social reality, as a stabilized viewpoint on some act of collective violence,
scapegoating must remain nonconscious.” In other words, as soon as members of the society
under investigation figure out that they are scapegoating, they are no longer doing it; only the
sociologist can discover this for them.
28
Evans-Pritchard’s analysis (1965) of the Nuer sacrificial system, however, is exceptionally
detailed.
29
See Goody 2010:27 (esp. n. 36), who demonstrates Radcliffe-Brown’s ambiguity on this
matter.
30
See also Staal 1989:324, who demonstrates convincingly that knowledge of a ritual’s origin,
even when available, is not equivalent to knowledge of its meaning.
31
See Girard 1977, 1987; Burkert 1979, 1987; J.Z. Smith 1987; and the summary in Naiden
2013:3–15––in particular the discussion of Meuli.
11:50:13:06:14 Page 181

Page 182
6.2.3 Representational Approaches
The task-oriented and functionalist approaches share a common

denominator: neither equates meaning with representation, let alone in the
high-resolution sense that specific ritual minutiae represent specific external
items or concepts. For this reason, they do not pertain to meaning in a way
that is analogous to natural languages or semiotic systems in general. By
contrast, a representational approach to ritual meaning resembles the study
of linguistic semantics insofar as it employs some relationship of signification
or representation between signifier and signified.32
6.2.3.1 The Extensionist Approach

The most rudimentary tool for distinguishing the meanings of different ritual
sequences is by differentiating one occasion for their performance from
another. For the sake of explanation, consider the following analogy to an
extensionist theory of meaning in linguistics. Stated simply, such a theory
asserts that the meaning of a proposition is the set of real-world situations
where this proposition is true. For example, the meaning of the sentence “A
cat is under a red chair” is the sum total of all situations where a cat is located
under a red chair.
This approach offers a convenient analogy to sacrificial ritual. Consider the
following statement (from Num 28–9):
(1) (2B웧, S웧, 7s웧)wholeburnt means [New Moon, or Feast of Unleavened Bread,
or Feast of First Fruits]
(2) (B웧, S웧, 7s웧)wholeburnt means [1st of the 7th month, or 10th of the 7th
month, or 22nd of the 7th month]
According to this scheme, the New Moons, the seven days of the Feast of
Unleavened Bread, and the Feast of First Fruits constitute one meaning,
whereas the first, tenth, and twenty-second days of the seventh month con-
stitute another. One may then compare the meaning of zoemic combinations
in P with those in Ezekiel (45:18–46:11) and find that the meanings (that is,
the sets of occasions) are different: here, the seven-day Feast of Unleavened
Bread is grouped together with the seven-day Autumnal Feast (the offerings
are 7B웧, 7S웧), whereas the New Moon and the Sabbath seem to be more
32
The term “representational” is intended to include the semiotic (or semiological)
approach and the symbolic approach, which is what many moderns and ancients have in mind
when considering the “meaning” of ritual. See Sperber 1975a:5: “a myth is a sign in a different
system from that of the language in which it is told. The difference between signs (i.e., signifieds+
signifiants) in language and in symbolism is that the former are created and exist only within
language, whereas the signs in symbolism exist also independently (a star in the physical world,
a word in a language).”
11:50:13:06:14 Page 182

Page 183
Meaning 183
closely related (the offerings on both these days are 6s웧, S웧; on the New
Moon, a B웧 is added).
The extensionist approach may assist in the interpretation of some
numeric combinations in TS 13–29 and in other post-biblical systems where
such combinations are found. Still, its utility is limited, since few occasions
share precisely the same zoemic combinations. Moreover, the approach has
a very weak explanatory power, since it fails to point to a different type of
representation likely to be found in these numeric combinations, such as the
fact that the combination 6s웧, S웧 (for the Sabbath in Ezek 46:4) corresponds
in some way to six days of the week (6s웧) and one Sabbath (S웧).33 This latter
type of representation is often referred to as “symbolic” and deserves our
special attention.
6.2.3.2 Symbolic Approaches

There exists an ancient exegetical tradition that associates the legal minutiae
of sacrificial systems with the non-ritual (real or imaginary) world in a
manner often referred to as “symbolic.”34 Despite this fact, and perhaps
because of it, there is some lack of clarity about what constitutes a symbolic
system.35 Some studies have even claimed that the study of symbolic systems
does not belong to the examination of meaning. According to Sperber, for
example, symbolic systems––under which he subsumes ritual systems––
are based on representation, as are semiotic systems, and in his view both
symbolism and semiotics are based on a relationship between signifier and
signified. However, he argues, symbolic systems lack semantics in the sense
of natural languages, since symbolic systems lack the properties of para-
phraseability and analyticity.36 In fact, according to Sperber, symbolic systems
33
In this particular instance, it is difficult to determine what precisely is achieved by symbol-
ically conveying the obvious (namely, “this is a Sabbath, and it is preceded by six weekdays”).
Alternately, it is possible that the 6 + 1 model was not chosen for this reason, but simply because
it is a symbol of completion, and therefore considered appropriate for the Sabbath.
34
This term (adverbially, συµβολικ) is already found in Philo with particular reference
to the ancient Israelite sacrificial system. See n. 000. Note that in the Vedic tradition, the corpus
of Brahmanic literature––ostensibly a symbolic/allegorical interpretation of the ritual laws in
the Śrautasū¯tras––almost invariably precedes the composition of the Śrautasū¯tras themselves:
the latter were often composed by subtracting the non-praxemic details from an existing text.
See Gonda 1977:496–7.
35
See for example Sperber 1975a. Sperber clearly wishes to include ritual systems under the
umbrella of symbolic systems, though most of his treatise, Rethinking Symbolism, is dedicated
to mythology rather than ritual. For a balanced account of the value of the symbolic approach,
see Gilders 2011, presented as a view contrary to Klawans (2006, 2008, though Klawans, too,
cautions against the attempts to find symbolism in every detail (2011:116–17). See also Asad
1993:55–62.
36
They lack the property of paraphraseability inasmuch as one set of symbols cannot be
paraphrased with another set of symbols, and they lack the property of analyticity since one
symbol cannot be defined with two or three others.
11:50:13:06:14 Page 183

Page 184
are not even semiotic systems, since they lack the possibility of fixed pairing
between code and message.37 Hence, he argues, symbolic systems, including
ritual systems, simply do not mean.38
Unfortunately, the types of representation that are potentially symbolic
appear to be too diverse to fit an economical theory. To this end, consider
a few examples of zoemes culled from Σ and other ancient Near Eastern
sacrificial systems:
(1) A correspondence may exist between a zoeme’s sex and that of its
offerer: a female animal is offered for a female offerer and a male
animal for a male offerer. This is common in Hittite nakušši
(“scapegoat,” literally “letting go, dispatchment”) and tarpalli (“sub-
stitution”) rituals.39 One might argue that the zoeme relates to the
offerer either metaphorically (as they have analogous body parts) or
metonymically (as the offerer is expected to be in contact with the
zoeme), or both. It is nonetheless unclear whether such representation
should be termed “symbolic,” since one might argue just as effectively
that the correspondence is necessitated on purely task-oriented
grounds: where substitution is involved, every part of the zoeme’s body
serves as a substitute for the corresponding part of the offerer: “head
for head, genitalia for genitalia.”40
(2) A zoeme’s sex may correspond to that of the deity to whom the
offering is made, in cases where a pantheon exists.41 It is questionable
whether the representation here operates in the same sense as the
previous cases, as the gods are not the ones sacrificed but the ones
receiving the offering.42
In the Israelite system, one does not find a direct correspondence between
the sex of the animal and that of the offerer (let alone that of the deity), but a
different correlation is found among the purification offerings of Leviticus 4.
37
What differentiates semiotic systems from other representational systems is that only in
the former is there the possibility of paraphrase and of the composition of an inventory (such
as a listing of basic units in the left-hand column of a table and their denotations in the right-
hand column). Symbolic systems are not semiotic systems, since even for the weakest form of
semiology, there has to be the possibility of fixed pairings between “code” units and “message”
units, and this cannot be done with symbols (Sperber 1975a:14–16).
38
Sperber further argues that since de Saussure, semiology has centered on the question of
“how do symbols mean?” rather than “what do symbols mean?,” implying that symbols, in fact,
do not mean (1975a:53).
39
See Gurney 1977:50–1.
40
Gurney 1977:54.
41
See, for example, the ritual of Puliša referred to in Gurney 1977:48: both male and female
animals are used since it is unknown whether a male or female deity is responsible for the
calamity. For the text see Kümmel 1967:112–115 (see in particular ll. 23, 40) with additional
examples of this phenomenon in the comments on p. 122. See also Scheid 2003:80.
42
Note that a task-oriented argument––for example, that male gods prefer male zoemes––is
insufficient here, since one would still have to account for such a preference.
11:50:13:06:14 Page 184

Page 185
Meaning 185
(3) A male zoeme is required for chieftains (Lev 4:22–6) and a female
zoeme for laypersons (vv. 27–35). The reasons for this cannot be
purely economic, since females are the less dispensable animals in a
household’s flock, and so some symbolic correlation may be at play
here. The metaphorical reference in Ezek 34:17 to the political elite as
i TE
“rams” (OYL a , S웧) and “billy-goats” (OYD
i YA iW d x, G웧) lends support to
this theory. Though it is unnecessary to assume that a RYE om (웧) in P
i$
is supposed to symbolize a chieftain (as rams and mature billy-goats
do in Ezekiel’s prophecy), it is reasonable that an animal representing
virility was chosen as the offering of powerful political leaders, whereas
an “effeminate” zoeme (웧) was selected for other Israelites.43
(4) A similar correspondence may exist between the economic value of a
zoeme and the social status of the mythic figure that it represents in
a sacrificial rite: for example, “A steer and seven sheep are thrown alive
into a pit: they are Kingu and his seven sons.”44 In this Mesopotamian
context, it appears that the larger and more expensive animal repre-
sents the father and chief-traitor Kingu, whereas his minor accomplices
are represented by sheep.
(5) Some Jewish texts identify a correlation between zoemes and mythical
events. For example, according to one tradition45 bulls are offered in
commemoration of Abraham (who offered a RQ ie to his guests
B-IB
md
m
in Gen 18:7), rams in commemoration of Isaac (who was substituted
by a ram on Mt Moriah in Gen 22:13), and goats in commemoration
of Jacob (who covered his skin with goat hides to deceive his father in
27:16).46
(6) Another example of symbolic interpretation is the selection of a heifer
as a purification offering which a rabbinic tradition justifies by the
statement “let the mother [i.e., the B웨] come and wipe off the feces
[of her son, b웧].”47 The son’s feces are the heinous deeds of the
Israelites in relation to the golden “calf” (LGeEa in Exod 32, which the
rabbis understand as b웧).
43
Marx (1994:161) notes that Saul, David, and Solomon are symbolized by rams in Enoch
89:42–8. Ezekiel’s Temple vision does not mention which zoemes are used for the private
offerings of the AY$ N. The zoemes mentioned in chs 45–6 are all male, but they are also all
oi m
public.
44
Heinsohn 1992:126. Other forms of symbolic representation are found in this context: a
dove cut in two represents Tiamat, who was cut in two by Marduk.
45
On this tradition in the Aramaic Targumim, see Klein 1980:92–3.
46
Note the implicit pun: according to this tradition, the zoeme RYE
imo$ (웧) commemorates
the occasion where Jacob pretended to be an RE im i (“hairy man,” Gen 27:11).
o$ $YoA
47
Tanh. Huqqat 8; note the variants in Rashi, in Chavel 1982:464. See, however, above,
nn. 000, 000.
11:50:13:06:14 Page 185

Page 186
In relation to ritual systems, there appears to be a range of symbolic

interpretations: the correspondences between “ritual correspondent” and
“non-ritual correspondent” for some may be historically sensitive, whereas
for others, such correspondences are patently artificial and post hoc. An
example of the latter type is Philo’s misinterpretation of the adverb BYB imS
(following LXX’s κκλ7) and consequent argument that blood is applied in a
circular motion because the circle is the most perfect of geometrical shapes
(DSL 1.205). This perfection, Philo explains, reflects the perfection of the
mind and its paths in its willingness to do God’s service.48 The apparent
artificiality of such correlations has led to their disregard in much of modern
scholarship, as they attribute meanings to a system in which they do not
inhere––much as astrology attributes meanings to the (inherently meaning-
less) relative positions of celestial bodies.49
It would be unjust to ignore such far-fetched interpretations entirely in a
systematic treatment of the meaning of ritual. Unlike the astronomical bodies
studied in astrology, sacrificial systems are a product of human societies.
Furthermore, some symbolic observations, such as 6s웧, S웧 being offered for
the seven days of the week (Ezek 46:4, as already noted), are potentially
useful in understanding the meaning behind aspects of Σ. Therefore, there is
a priori some likelihood that the rites carry symbolic meaning.50 However,
the inclusion of all post hoc exegesis results in an interpretation that is un-
systematic and does not admit of any substantial verification; for this reason,
we are forced to reject some of the more apparently forced interpretations.
6.2.4 Structuralism
It is within structural anthropology that the analogy between ritual and

language reaches its most elaborate formulations. Structuralist anthro-
pologists such as Lévi-Strauss, Leach, and Douglas pay particular attention
to the formal properties of ritual, especially to logical patterns in the
48
Philo specifically refers to this representation as operating symbolically (συµβολικ, l.5).
See also Gilders 2011.
49
See Staal 1989:452; less radically, Gane 2008:211. These superordinary attached meanings
are, in Staal’s view, in constant flux, in contrast to the ritual acts themselves, which remain
relatively stable over comparatively long periods of time (p. 125). The opposite view, however, is
espoused by Levine 1974:45–52, who implies that the basic meanings of some rituals can be
relatively stable, even though they undergo many praxemic shifts and changes in structure over
time.
50
On a rare occasion, P appears to endorse a symbolic interpretation, with regard to the
jugate salt (Lev 2:13), which is associated with an “eternal covenant.” Compare the figurative
usage in Num 18:19 OLWE XLM TYRB, and see Milgrom, Leviticus 1.191, but see also Haran
1978:143 n. 12 and Schwartz 2004:210. For another putative example, the “token” (HRKZA), see
Klawans 2011:114.
11:50:13:06:14 Page 186

Page 187
Meaning 187
arrangement of ritual elements.51 However, many structural studies of ritual,

inspired by structural linguistics and information theory,52 do not limit
themselves to the formal aspects of ritual but rather attempt to describe the
ultimate signification of various ritual systems, proposing that they are based
on signifier–signified relations. The structuralist approach differs from the
symbolic approach in suggesting that signification is not straightforward,
as there is no fixed meaning in every constituent unit. Rather, signification
arises––as in language––from the combination and juxtaposition of several
constituent units.
Ironically, the Structuralist approach bears significant resemblance to the
task-oriented approach, which does not view meaning in terms of significa-
tion. In the first place, both stress the centrality of grammatical minutiae in
ritual systems rather than offering an overview of their general character,
and yet both avoid resorting to symbolic representation.53 Additionally, in
some of Lévi-Strauss’s writings, it appears that if meaning is to be found in
a structural study of ritual, then this “meaning” is nothing more than the
rearrangement of the same basic units in a different––homologous or con-
flicting––rite.54 In fact, Sperber argues that Lévi-Strauss actually proves the
opposite of what he intends: he has attempted to demonstrate that several
non-linguistic systems such as myth and ritual are semiological systems, but
he has shown that in fact they signify nothing.55 Thus, Structuralism paves
the way back to a skeptical approach to meaning in ritual, and eventually to
the view that rituals are inherently meaningless, as in the writings of Frits
Staal.
6.2.5 Frits Staal and the Meaninglessness of Ritual

6.2.5.1 The Meanings of “Meaninglessness”
Theoretical approaches championing the meaninglessness of some or all
aspects of ritual are found in a number of classical traditions that interpret
Vedic and Israelite sacrifice.56 In modern anthropological circles, this
51
For example, Leach 1964 and Douglas 1966. For a summary and discussion, see Kunin
2004, as well as the application of his theory to a specific aspect of Israelite rituals on
pp. 98–103.
52
See Lévi-Strauss 1963c and Sperber 1975a:22.
53
See Kunin 2004:31, and his critique of Douglas, particularly pp. 48–52.
54
Sperber 1975:7–10.
55
See Sperber 1975a:51–84.
56
On the Vedic conception that rituals are meaningless, see Staal 1989:234. On the analogous
Jewish tradition, see, for example, the narrative about R. Yohanan ben Zakkai and his students
in Numbers Rabbah 19, Pesiqta d’Rav Kahana 4:7. One would be very hard-pressed to find
evidence of such a conception in P, or, for that matter, anywhere else in the biblical texts.
11:50:13:06:14 Page 187

Page 188
approach is most closely associated with the work of Frits Staal.57 There are
two senses in which Staal uses the term “meaningless” with reference to ritual
systems.58 The first is “useless,” in the sense that ritual action does not achieve
anything. As Penner notes, Staal here follows the Weberian equation of the
meaningless with the irrational (that is, that which is not goal-oriented), since
rationality is always a means–end calculation.59 Note that this conclusion of
Staal’s entirely rejects the task-oriented approach, which interprets sacrifice as
a tool for manipulating the universe through praxemic minutiae. In this
respect, the uselessness of ritual may run afoul of the views of ritualists
themselves: as Milgrom argues, for instance, it is likely that P considers that
the blood of purification offerings really does achieve its purpose of purging
miasma from the bronze altar.60
The second sense in which Staal uses the term “meaningless” is “devoid of
semantics,” a meaning that has more theoretical implications for the gram-
mar of ritual. Staal repeatedly argues that ritual is “pure syntax without
semantics”––that is, that it displays structural properties akin to syntactic
structures in natural languages but lacks the corresponding semantic proper-
ties.61 I find Staal’s hypothesis enticing, yet his argument is faulty. In the first
place, it identifies lack of semantics with lack of meaning, and yet it has been
demonstrated (see p. 000) that purely syntactic structures (devoid of seman-
tics) may nevertheless have a rudimentary form of meaning. This error may
have arisen from Staal’s identification of meaning with reference. As Penner
notes, Staal is mistaken in this––syntactic components such as “and” and
“not” lack reference but are certainly not devoid of meaning.62
6.2.5.2 English and n•“Buffalo”

Staal’s theory that rituals entirely lack semantics is faulty for another reason.
Let us examine one of Staal’s simplest syntactic rules––namely, the “trans-
formational rule” identified by Staal as
A AA
57
Michaels, in response to Staal, offers a solution to the problem of ritual meaning,
formulated in a koan-like fashion. “Rituale sind bedeutungslos, aber das ist nicht ohne
Bedeutung” (2007:255; see also 2006:201, 248): rituals are meaningless, in that they signify
nothing and achieve nothing, but the fact that people are engaged in such meaningless activity
is not meaningless. (This characterization, however, does not claim to explain the details of
rituals.) No longer should one speak of meanings of rituals, though these, Michaels admits, are
imposed by many practitioners; there exists only a meaning of ritual, one grand significance in
the very fact that humans worldwide engage in meaningless activity. Thus, in this case Michaels
avoids the problem of the relationship between ritual “syntax and semantics” (his terminology).
58
Staal 1979, 1980, 1989.
59
Penner 1985:2.
60
Milgrom 1976b; see, however, Gane 2008.
61 62
See Staal 1979, 1980, and compare 1989:234. Penner 1985:2–3.
11:50:13:06:14 Page 188

Page 189
Meaning 189
This rule states that ritual unit A may be placed after another identical ritual
unit A. This rule can be repeated indefinitely, with the unit A placed again
and again before an increasingly protracted ritual sequence, and hence can
produce the streams AAA, AAAA, AAAAA, and so on.63 Unit A may consist
of uttering a sāmidhenı̄ verse, as in Staal’s examples,64 but according to his
scheme, it may also consist of offering an s웧, or even salt.
Staal formulates this rule inductively, according to the observation of
rituals of the form A, AA, AAA, and so on. Since a long stream of “A”s does
not seem to be conducive to semantic analysis, Staal uses this example to
suggest that such rituals are devoid of semantics.
However, the assumption that a transformation A AA is purely syntactic
would be false if applied to natural languages, and there is no reason a priori
to assume that it should be devoid of semantics in ritual. Several languages in
fact contain the transformation A AA effecting a semantic shift. In BH,
for example, the indefinite “anyone” ($oYA i $oYAi ) is formed by reduplication
of the word for “man” ($oYA i ). A more active example is found in some
Sumerian plurals formed by reduplication as well: if KUR denotes “land,”
KUR-KUR denotes “lands.” Admittedly, these are limited to a single case of
reduplication, and the sequences resulting from these transformations do not
produce complete sentences. However, it is easy to imagine a language where
the transformation A AA can be applied more than once to produce longer
chains, such as $oWoDQ
m $oWoDQ m or “very, very, very. . .,” and even constitute
m $oWoDQ
a proposition.65
More importantly, the argument that the rule A AA is operative reflects
an unproven assumption that individual ritual elements are simply grouped
together cumulatively. Therefore, the statement that the forms A and AA
are products of a “purely syntactic” process is circularly dependent on the
assumption that their sequencing is achieved in a manner devoid of
semantics.
In natural languages, for example, it is possible to imagine a sentence where
the transformation A AA appears to take place but in fact is not at work at
all. Consider the following English utterance:66
Buffalo buffalo Buffalo buffalo buffalo buffalo Buffalo buffalo.
63
Staal 1989:113.
64
Staal 1983, 1.310–11 claims that A AA is applied twice to the first and last sāmidhenı̄
verses recited in the animal sacrifice for Vāyu. See also Staal 1989:139, where this transformation
is said to be applied several times in mantras uttered when a small golden image of a man is
buried under the fire altar of the agnicayana, creating sequences such as “phal, phal, phal, phal,
phal” and “bham, bham, . . . [eighteen times].”
65
See for example the triple repetition of the phrase HN dmM
e
diM ON
iQd
mxH YN
v$vo-TXxT
dx RTod v
Px W in
Kv
Exod 25:35.
66
See Pinker 1994:209–10.
11:50:13:06:14 Page 189

Page 190
Though rather awkward, and very unlikely to be uttered in an everyday

conversation, this sentence is perfectly grammatical, stating that “Buffalo
buffalo” (buffalo bulls from the city of Buffalo) that “Buffalo buffalo buffalo”
(whom other buffalo bulls from Buffalo bewilder) themselves “buffalo
Buffalo buffalo” (bewilder yet other buffalo bulls from Buffalo).67 The
sentence is analogous in structure to Red bandanas Oxford students wear anger
grumpy bulls, so several transformational rules are at work here, none of
which is the transformation A AA.
In fact, Σ supplies us with a counterexample to Staal’s abstracted trans-
formation. A long chain of salt offerings might appear to an observer to be a
sequence of saltA offerings generated by a transformational rule salt salt
salt. However, the jugational rule posited about salt (§ 3.9.2) was not of the
form salt salt salt, but rather A A salt. Moreover, the various salts in
the resulting chain are on different jugational levels (A, B, C, etc.). This
analysis undermines Staal’s rule in two ways. Most obviously, postulating a
transformational rule such as Staal’s would erroneously predict that only salt,
rather than any materia sacra, would be followed by salt. More importantly,
Staal’s rule fails to capture the underlying structure of such a transformation:
the appended salt jugates are not coordinates, since they are not on the same
jugational level, much as the various “buffalo”s form subordinate clauses on
different levels in n•“buffalo.”
In order to establish that a ritual sequence n•A is devoid of semantics, it is
insufficient to surmise that transformation A AA is probably at work.
In order to establish this, one must at least prove that it is at work here: it
is not at work, for instance, in the sentence n•“buffalo” (as we have seen, the
sentence involves multiple rules, none of which involves reduplication). So
too, drawing conclusions only from lengthy chains of salt jugates would
incorrectly suggest that the element salt can be adjoined to itself––while the
correct operative rule is that salt can be adjoined to any jugate. Both systems
produce examples that resemble A AA in form, but neither contains a rule
that is analogous in structure.
Thus, in order to demonstrate that ritual is pure syntax without semantics,
Staal is required to adopt certain semantic preconceptions concerning the
rituals under examination. He cannot have arrived at a syntactic analysis of
these rituals without a priori deciding that they lack semantics. If this is true
67
It can be proven by induction that any sequence with n ≥ 1 occurrences of the word
“buffalo” is grammatical in English: the sentence is clearly grammatical for n = 1 (“buffalo” as
an imperative), for n = 2 (the first “buffalo” is a noun, the second a verb), for n = 3 (the first
“buffalo” is a noun, the second a verb, the third a direct object), and for n = 4 (same as n = 3,
with the adjective “Buffalo” qualifying either noun). Beyond this, the subordinate clause
“buffalo buffalo,” meaning “whom buffalo bulls bewilder,” can be inserted after any noun to
qualify it, so that sentences of the word “buffalo” can be made arbitrarily long.
11:50:13:06:14 Page 190

Page 191
Meaning 191
in the case of the simple transformation A AA, which is only roughly

paralleled in natural languages, it is all the more valid with regard to his other
transformational rules that resemble syntactic structures in natural languages,
such as the hierarchic structures in the agnicayana ritual (§1.3.3). While
a purely formal analysis of ritual––“purely syntactic, devoid of semantics,” to
use Staal’s terms––is a fruitful line of investigation, Staal’s hypothesis that
rituals are meaningless has not been proven.
6.3 THE INTERFACE BETWEEN JUGATION,

HIERARCHICS, AND MEANING
However, it appears that Staal’s theory is faulty for a more fundamental

reason, as the following example from Σ demonstrates. In , we
noted that the Temple Scroll systematically requires every calendric purifi-
cation offering to be accompanied by a grain offering as well as a libation
(§3.7.2). Let us now examine how the author of TS arrived at such a con-
clusion, which does not accord with P.
Yadin argues that the author of the Scroll required that subordinate jugates
accompany calendric purification offerings by analogy to the offering of the
person purified from scale disease (Lev 14:10).68 However, this argument
cannot be sustained, as it raises more problems than it solves. First, the person
purified from scale disease offers only a cereal offering, and not a libation
of wine. Second, this cereal offering, as we have seen (§3.3.2), is apparently
coordinate with the offered zoemes and can hardly serve as a paradigm for
subordinate jugation in purification offerings. (In any case, it is by no means
subordinate to the purification offering.) Third, even if the purification
offering of a person purified from scale disease were to entail subordinate
jugates, it would still require a leap of imagination to extend this law to
include many other purification offerings.
There appear to be much more immediate considerations that led the
author of the Temple Scroll to make this requirement. Two suggestions are
made here.69 The first pertains to a development in the semantics of the
Hebrew language; the second, the essentials of which were noticed by
68
“It seems that in this matter. . . the prescription in the scroll was laid down in accordance
with the sin offering of the leper.” See Yadin 1983 1.145.
69
For a third solution, see n. 000; but it is unlikely that this was the cause of the polemic.
Similarly, it is difficult to accept Milgrom’s (1990:243) suggestion that TS was misled by the
final form of Num 28–9, since, as demonstrated here, in none of the ancient versions do the
texts of these chapters imply that every purification offering ought to entail both wine and grain
subordinate jugates.
11:50:13:06:14 Page 191

Page 192
Schiffman but not pursued,70 pertains to a development in the grammar of

sacrifice. The latter may also explain why the author was so adamant about
this seemingly minute detail.
(1) It is known that the term OYKimSvN, which in P denotes liquid libations,
eventually came to serve as a general term referring to both cereal
offerings and libations.71 In fact, though OYK imSvN sometimes denotes
“libations” in MH, its most common use in rabbinic literature is
roughly “subordinate jugates of a zoeme.”72 There is some inconclusive
evidence that this semantic shift occurred somewhat earlier, as
indicated in Chronicles and Daniel.73 It may have been known to the
author of the Temple Scroll, who uses the term WKSN TXNM in 26:7–8.74
Given, then, that OYK
im SvN had come to mean “libation + grain offering” in
the Hebrew of the late Second Temple, it is understandable that the majority
of calendric purification sacrifices depicted in Numbers 28–9 (as discussed,
§3.7.2) were interpreted as requiring both a libation and a cereal offering.
The application of this rule to the purification offerings assigned for the
17th–22nd days of the seventh month would simply be a harmonistic
70
See Schiffman 1995, as well as Knohl and Naeh 1993:26 and Werman 2006:95.
71
It is difficult to determine why this semantic development occurred, particularly con-
sidering that the etymology of vVe N relates to liquids. However, the mixture of semolina and oil
Se
according to the measurements in Num 15 yields a viscous liquid, hence the extension of the
term vVeSeN to include this mixture is not entirely misguided from the point of view of etymology.
72
See b Zeb 5a, 9a; b. Ber. 15b, et passim, and Maimonides’ definition (HMQ 2:1):
OYKSN IYARQNH OH IBRQH OE IYAYBM$ TLSHW IYYH (“the wine and semolina brought with
the sacrifice are what is termed OYKSN”), followed by OYKSN TXNM TARQN HDBL TLSHW
(“and the semolina alone is termed OYKSN TXNM”). Thus, the grain offering is usually termed
OYKSN TXNM (partitive genitive), rarely HMHB TXNM (possessive genitive, implying subordinate
jugation), in rabbinic literature.
73
See perhaps 2 Chr 29:35 HL m EoL
m OYK
im Sd BiW (“and with the libations and cereal offerings
vNx
for the wholeburnt offering[s]”) and 1 Chr 29:21 OH e YK
av
d SiW
Nv FL
eA
e OY$
oi m
Bd
vK (“a thousand lambs
and their libations and cereal offerings”). Since Chronicles often reflects P’s system, it is more
reasonable that the term OYK imSvN includes grain offerings than that the Chronicler imagined that
only libations accompanied the zoemes here. Note, however, that one cannot use these verses
as proof of a semantic shift without entering into a circular argument. Compare also Dan 2:46,
HaL HK
d mSmxL
d Nv RMxqA IYXiX
o YN
iWv HX NiiW, which, though not in a cultic context, is quite striking (the
mvM
HX Ni is literally “libated”).
mvM
74
See n. 000 concerning the partitive genitive; this is also how Yadin 1983 2.116–17 under-
stands the phrase. However, the possibility must be entertained that the author meant
WKSN TAW WTXNM TAW, since the lack of a wine libation would be disturbing here (compare 16:9,
18, also in the context of a B웧purification). See also the terms WNM$ TXNM (24:5) and WTLWS TXNM
(34:12), both denoting the subordinate grain offering including its oil. The term TXNM
TWNBRWQH (corrected from OYNBRWQH TXNM, TS 38:7 in fragment 4Q365a) probably does not
denote grainB offerings (and is therefore not analogous to the rabbinic OYKSN TXNM, TXNM
HMHB) but rather grainA offerings, since they are eaten; more specifically, it denotes raw grainA
offerings, since the author specifies HNWBL HYLE HABH. Recall that the author of the Temple
Scroll (20:9–11) implies that grainB offerings are not eaten and do not entail frankincense, and
that only raw grainA offerings entail frankincense.
11:50:13:06:14 Page 192

Page 193
Meaning 193
measure, applying the rule of the majority (the larger textual body and greater
number of sacred occasions) to the minority as well.75
(2) As we have seen (in ), the author of the Temple Scroll
posits a much more robust system of hierarchics than any of the
other texts examined here. Systematically, the author believes that every
purification offering in reality consists of two sacrifices, a purification
offering (the portions of flesh that are either burnt outside the camp
or eaten by priests, and the blood that is daubed) and a wholeburnt
offering (the suet, which is burnt on the altar).76
The requirement that every calendric purification offering should entail a
subordinate libation and cereal offering possibly then stems from the notion
that every purification offering entails a small wholeburnt offering of suet. As
a micro-wholeburnt offering, this suet would require subordinate libation
and cereal offerings. This would be a very odd––but formally correct––read-
ing of Num 15:3. Recall that the claim that TS understood that parts of each
purification (and ordination) offering constitute micro-wholeburnt offerings
is not based on conjecture: it is explicit in 15:3–9 and is all but explicit in
16:6–13.77
This may perhaps lead to a better understanding of the Temple Scroll ’s
sacrificial system, for if this theory does accurately reflect the Temple Scroll ’s
system, it would have the following ramifications. Recall from Num 15:3
that only those (quadruped) offerings that are HWHYL x H$ i (“a food-gift
edoA
to ”) and HWHYL x x XX
o YN
ixXYRa (“sweet-smelling to ”) require sub-
ordinate jugation. If the author of the Temple Scroll considers that the 웧
calendric purification offering requires subordinate jugation, then it follows
that this 웧 (or at least its suet) is considered a sweet-smelling gift to .
75
This solution would be more plausible if it could be demonstrated that the singular VSN,
not only OYKSN, denoted “libation + grain offering” in late Second Temple Hebrew. The use of
the singular WKSN TXNM in TS 26:8 may be an example of such usage.
76
The same may be true of ADL, though this is more difficult to prove. Note that from a
literal point of view, chs 8 and 9 in their entirety pertain to “wholeburnt offerings” only. For
example, 7:7 speaks of wood placed AXBDM LE ATLE TWXTL (“beneath the whole-
burnt offerings upon the altar”), and yet 9:1 indicates that wood is required for suet as well,
in cases where suet is offered alone. See nn. 000 and 000. This could imply (though it is
highly conjectural) that the suet is also considered a micro-wholeburnt offering. Concerning
2 Chr 29:35, see n. 000.
77
Note, however, that the specific phrasing of the Scroll may suggest that the subordinate
jugates are still subordinate to the zoeme as a whole, not merely to those parts of it that
constitute micro-wholeburnt offerings. This is evident for example from 15:9: the pronominal
suffixes in the phrase WKSNW WTXNMW (“and its [m. sg.] cereal offering and its [m. sg.] liba-
tion”)––though arguably referring to an abstract collective––most likely refer to the ram as a
whole (LYA, m. sg., l. 6), not to the phrase LYAH IM HLWE (f. sg., l. 6) or to the list of internal
organs (ll. 6–8).
11:50:13:06:14 Page 193

Page 194
This conclusion is in fact consistently found throughout the text of the Scroll
(for example, 28:5–6), lending credence to our interpretation.
The rabbis figure it the other way around: since only quadrupedal sacrifices
that are both HWHYL H$ A (“a food-gift to ”) and HWHYL
edoi x xXX
o YN
ixXYRa
(“sweet-smelling to ”) require subordinate jugation (Num 15:3), and
since the 웧 offered as a calendric purification offering is almost never called
HWHYL x xXXo YN
i x
XYRa H$a
dAoi in the Pentateuch,78 it should require neither a
subordinate grain offering (as P itself suggests) nor a subordinate libation
(contrary to what P explicitly states concerning all calendric purification
offerings except those of the 17th–22nd days of the seventh month).
If this reconstruction is correct, then it indicates an important difference
between the rabbinic tradition on the one hand and the Temple Scroll on the
other hand:79 are purification offerings sweet-smelling to ?
The Tannaim believed they were not. In this respect, they accord with many
biblical verses but must ignore, or offer their own interpretation of, a few
others (such as Lev 4:31; Num 28:24; 29:6).80 The Temple Scroll believed that
they are sweet-smelling to . In this the Temple Scroll betrays a biased
reading of P, since P almost always implies that they are not, though this
reading would not formally contradict any biblical verse.
This may explain why the author of the Temple Scroll seems so adamant
about these particular ritual minutiae: they reflect a serious theological
statement concerning sin, ’s olfactory preferences, and purification. If
this analysis is correct, it would not be the only instance where subordinate
jugation depends on whether or not a sacrifice is considered sweet-smelling.
Two analogies are in fact explicit in P:
(a) In Lev 5:11b, the absence of B-level jugates in a grainA offering
(HNmB
o vL m
HYLem diY-AL
E IT
a ov W IM
e$eo Hm YL
emE OY$ Y-AL
oi m o “One shall place no oil
upon it and put no frankincense upon it”) is justified with the state-
ment AWH i TAU m
dx i (“because it is a purification offering”).
X YK
d
(b) In Num 5:15b, the absence of subordinate jugates to a grain offering
(HNmB
o vL WYL
mm diY-AL
E IT
a ovW IMe$
eo WYL
mm E QCo iY-ALo , JPS: “One shall pour no oil
upon it and put no frankincense upon it”) is similarly justified with
i H TA
the statement IWomE TRedeKvM
Zx IWoRK
miZ TX
d xvM
Ni AW o mQ
Nv TX
xvM i (JPS: “for it
Ni YK
d
is a meal offering of jealousy, a meal offering of remembrance which
recalls wrong-doing”).
78
See n. 000.
79
See Knohl and Naeh (above, n. 000). ADL may share the view of TS, see n. 000.
80
Upon careful examination, the latter two cases, when read in their immediate context, may
actually support the rabbinic view: 28:19, 24 make a point of noting that only the wholeburnt
offerings are H$ i , and 29:2, 5–6 makes a point of noting that only the wholeburnt offerings are
edoA
x
XXo YN
ixXYRa.
11:50:13:06:14 Page 194

Page 195
Meaning 195
In these two cases, as in the case of the purification offering in Tannaitic

tradition, the fact that an offering is not considered a sweet-smelling gift
is reason enough to prevent it from being accompanied by subordinate
jugates.81
Thus, we have before us a case in which the interpretation of an offering’s
“meaning” relates to a shift in hierarchics, which in turn entails a change in
the offering’s jugational pattern. The author of the Temple Scroll interprets
the purification offerings mentioned in P as sweet-smelling food-gifts to
; this relates to a shift in the hierarchics of purification offerings, since
purification offerings came to be viewed as comprising a micro-wholeburnt
offering and a micro-purification offering; and this in turn leads to the
requirement that purification offerings entail subordinate jugates of grain
and wine, inasmuch as every purification offering contains a wholeburnt
offering. Finally, through a process of several transformations between the
various operative categories, the final ritual sequence is expressed in terms
of praxemics, inasmuch as the zoemes and their jugates are physically offered
at the altar.
As this example demonstrates, an authority interpreting the “meaning” of
a ritual sometimes also inserts formal changes in the ritual’s structure in
accordance with his interpretation. In such cases, it is inaccurate to detach
“syntax” from “meaning” (as Staal uses the terms), since the “syntax” of the
ritual is determined by the meaning attributed to it by the ancient ritualists.
It is therefore possible that in ritual, as in language, “syntax” can never be
fully detached from “semantics.”
In sum, Staal fails to demonstrate that rituals are meaningless: by utilitarian
standards, his argument is perhaps trivially true (from an etic point of view)
but unsatisfactory. And in terms of an alleged lack of semantics, although
Staal succeeds in enticing the reader to consider this possibility, we have seen
no rigorous proof of the meaninglessness of ritual in Σ.
6.4 CONCLUSION
It remains for the ritual grammarian to define clearly what constitutes

meaning in a ritual system. As Sperber, Severi, and Kreinath have argued,
the analogy to linguistic semantics is likely to be misleading.82 However, if it is
possible to construct a semantics for a non-linguistic system comparable
to that of natural languages, then according to the Bloomfieldian and
81
This link between “meaning” and “jugation” is explicit in rabbinic literature; see b. Sot.
15a.
82
Sperber 1975a; Severi 2006; Kreinath 2006:468.
11:50:13:06:14 Page 195

Page 196
Chomskian model, the formulation of a grammar should require minimal

reference to a rudimentary meaning. This rudimentary form of “meaning”
should include the identification of such operative categories as “zoeme,”
“jugate,” “hierarchic level,” and “praxeme,” and of the particular inventories
and rules that are operative in the system––but not of semantics. After
composing the grammar, one may inquire whether a ritual system’s, meaning,
symbolic or otherwise, correlates to its form in a manner comparable to
the syntax–semantics correlation of natural languages. The present state of
inquiry would seem to suggest that this is usually not the case, but the
question must be left for future research.
At any rate, our investigation of particular rituals within Σ has demon-
strated that a change in a ritual’s jugatio-hierarchic structure can correspond
to a change in its meaning. In particular, the authors of the Temple Scroll
do not consider purification offerings primarily as a sanguine detergent, but
rather as a sweet-smelling food-gift (HWHYL xx XX
o YN
ixXYRa H$
aoi ).83 This shift in
dA
the so-called “meaning” of the sacrifice––the specific significance assigned to
it by the texts and, possibly, the participants and officiants––is linked to a
“grammatical” shift expressed in this offering’s hierarchics and jugational
patterns. While it is impossible to ascertain which shift occurred first, the fact
that an alternative interpretation and alternative hierarchic and jugational
patterns are found within the same text suggests a certain interdependence of
form and meaning in sacrifice. However ritual meaning is defined, this paired
shift suggests that Staal’s assertion that rituals lack all meaning whatsoever is
not confirmed by our analysis of Σ.
On the other hand, in  we noted that a single zoeme may be
a purification offering on one level (purification0) but a reparation offering
or a wholeburnt offering on another level (reparation+1 or wholeburnt+1,
respectively). It was noted that in these cases the Level 0 designation of this
offering accords with its praxemics––the blood of a purification0 offering is
invariably daubed, not tossed, on the altar (or, if it is a bird, flicked and
drained)––while its Level +1 designation must denote something else.
What precisely this something else might be was referred to as “one of the
mysteries of sacrifice” (p. 000) and was provisionally associated with the
“desired effect” of the sacrifice, which many scholars would identify with its
meaning.84
It is essential that in these cases a single term has two distinct denotations,
the first pertaining to the praxemics of a rite, and the second pertaining to its
so-called “function” or perhaps “meaning.” The fact that a TAU m
dxX may be
offered as an HL m oWE or as an O$ m may be confusing to the modern reader, but
moA
the ancient ritualists apparently did not find this problematic––in fact, our
83 84
See §6.3. See n. 000.
11:50:13:06:14 Page 196

Page 197
Meaning 197
discussion revealed that such dual usage became more common with time.
The fact that within hierarchics, the same offering is a TAUm
dxX in its praxemics
but an HL m oWE in its “meaning” suggests that praxemics does not determine
meaning in a consistent manner. Thus, the identification of a grammatical
property, hierarchics, has a direct bearing on the interpretation of the mean-
ing of sacrifice.
11:50:13:06:14 Page 197

Page 198
The Grammar of Sacrifice and

the Sacrifice of Grammar
Thus far, we have demonstrated that it is possible to compose a grammar of a

ritual system, if “grammar” denotes a finite set of rules amenable to concise,
formulaic notation that, together with a finite inventory, may be used to
generate an infinite number of combinations. Such a grammar may predict
which combinations will be licit and which will not within Σ and is related to
the meanings of rituals in a complex manner.
Now that a grammar for Σ has been composed, it remains to (1) investigate
whether its rules are internalized unconsciously; (2) demonstrate that while
the grammar is complex, introducing a new set of operative categories and
novel notation, it is worthwhile in its explanatory power; and (3) indicate
some directions for future research. Finally, we must ask the question that has
lurked in the background throughout our discussion: In what sense is the
present analysis of Σ really a grammar, and should it be given a different
name?
7.1 UNCONSCIOUS INTERNALIZATION
Within linguistics, one of the most important arguments presented by

Chomsky in favor of generative grammar pertains to the manner in which
language is internalized by native speakers. Children acquire the rules
for a language in its entirety in the first few years of their lives through
exposure to a limited number of grammatical sequences and an even
smaller number of corrected ungrammaticalities. Furthermore, certain
grammatical errors are simply never made by children acquiring their
mother tongue and are therefore never corrected. Chomsky therefore
likens the mental faculty that inhibits the formation of such ungram-
maticalities to a switchboard with an initially restricted number of modes:
11:50:13:06:14 Page 198

Page 199
Grammar of Sacrifice and Sacrifice of Grammar 199
each switch is set to its fixed mode by exposure to a limited number of

grammatical forms.1
Ritual grammar might seem to be acquired in a completely different
manner, at least where complex rituals are concerned. Inasmuch as Σ is the
product of intellectual literary elites, the student is often consciously and
actively guided by the teacher, and both are aware of the existence of general
rules. However, it is necessary to distinguish between explicit and implicit
rules. Clearly, some rules are explicitly found in the earliest texts. The
sacrificial chapters of Leviticus are replete with examples of rules such as
HL m -TA
m EoH e UY$ i oP W and H
v iHv m YX
eTmvL
Ni d
HTmAo XT Nv “[the wholeburnt offering]
xiW
d
should be flayed and cut up into its parts” (Lev 1:6), which by all appearances
should belong to the grammar of Israelite ritual (see , Praxemics,
rule 15)a)iii)(3)).
Other rules, of a more general type, are often not explicated in the text.
Such are the generalizations listed by Maimonides in his Commentary to
the Mishna and in the Code. Let us return to these generalizations referred
to in the :
SYL IA D’GT TNAP AHRK’D ODQT YTLA TWNBRQLA EYM’G TYRQTSA A’DAW
DYXY TAUX LK IA D’GT VL’DKW . . . H’GWB HBQN RWBCH TWNBRQ EYM’G YP
IA’CLA EWN IM TAUX RWBCH TWNBRQ EYM’G YP SYL IA IYBY VL’DKW . . . HBQN
. . . ZEAMLA EWN IM HLWE ALW
And if you consider all of the abovementioned sacrifices, you will find that
(1) there is not a single female zoeme (x웨y) among all of the public offerings . . .
Similarly you will find that (2) every layperson’s2 purification offering (xpurification)
is a female (x웨purification). . . Similarly it will become evident to you that (3)
there is not among all of the public offerings an ovine3 purification offering
(!purification), (4) nor a caprine wholeburnt offering (!wholeburnt). . .
1
See in particular Chomsky 1988:16, 62–3; and Chomsky’s argument with Putnam in Block
1981a. See also the arguments that Lakoff (1969:126) raises in defense of Chomsky contra
Hockett.
2
This translation is based on a subtle but important intentional misreading of Maimonides
in order to simplify the scheme. Maimonides refers to individual versus public offerings.
However, since in his system there are three non-public (i.e., individual) offerings that do not
follow this rule, Maimonides states them as exceptions: (1) LK LE ABH RP=XY$M IHK RP
TWWCMH, (2) OYRWPKH OWYB LWDG IHK RP, and (3) AY$N RYE$. Our translation obviates this
cumbersome formulation and is more economical.
3
Here Maimonides uses the term IA’C to refer to ovines only, though the term is a cognate
of BH IAC o and Aramaic IAE, which mean (i.e., “small cattle,” comprising ovines and
caprines) in these languages. Similarly, he uses the term ONG (Arabic ‫ )ﻏﻨﻢ‬to designate the same
category () elsewhere in his commentary. I see no semantic difference between the two terms
IA’C and ONG, and it appears that in Maimonides’ Jewish Arabic dialect both can denote .
However, his choice to designate ovines by both terms, contrary to biblical and Mishnaic usage
(and contrary to Medieval Hebrew usage elsewhere), led to some confusion in the translations.
See Kafah. 1963:9 n. 6. Maimonides’ considerations might have been stylistic, since in one place
he wrote $ABK (cognate of o$e K, or s웧 in BH) which he later corrected to ONG, with no
Bd
e
apparent change in meaning.
11:50:13:06:14 Page 199

Page 200
These rules can now be schematized as follows:

Concerning quadrupeds:
(1) x웨y (public) is ungrammatical
(2) x웧purification (layperson) is ungrammatical 4
(3) !purification (public) is ungrammatical
(4) !wholeburnt (public) is ungrammatical
It must be acknowledged that these generalizations are rather straight-
forward, in that only a little pencil-work is required in their elucidation.
Especially in light of their simplicity, there is no reason to assume that these
rules needed to be taught explicitly.
Let us now examine a rule in Σ that by all appearances was never
consciously formulated in antiquity, yet which may assist in explaining
the emergence of several new sequences found in post-biblical texts. In
, we extrapolated a number of rules from individual attested
sacrificial structures to find that these were merely special cases of a more
general rule:
H6: zp+1 = [(xp)a , (xp)b ,. . .,(xp) m , (yp)a , (yp)b ,. . .,(yp)n] is grammatical
It is quite out of the question that rule H6 was consciously formulated in the
Second Temple period, either in words or in the form of a diagram (let alone
a formula). In any event, we have no evidence of formulations for such
abstract general rules.5 Yet several sacrificial combinations not found in P
appear to have been generated on the basis of this rule, both in rabbinic
4
These two first rules hold true for all documented strata of the Israelite sacrificial system,
with the ironic exception of P itself. Maimonides clearly considered rule (2) to be true in P as
well, since he followed the Tannaim in assuming that d HT
mm$
Nvo TBd a (g) is a female (g웨). A
x ZE
similar situation may exist with regard to rule (1), for P mentions a public wellbeing offering of
a bovine of any age or sex (RWo$o, Lev 9:4,18). Since Maimonides ignores this offering in his
Introduction (probably because it was performed uniquely in the Wilderness), explicitly limit-
ing public wellbeing offerings to the lambs mentioned in Lev 23:19, it is unclear whether he
considered the bovine in Lev 9:4,18 to be male. However, it is noteworthy that at least some
medieval scholars (e.g., Ibn Ezra, ad loc.) did consider this bovine to be male (this can be
deduced from the fact that Ibn Ezra limits the term to mature specimens, revealing that it
cannot denote B, and hence must denote B웧, the word’s other denotation in rabbinic Hebrew).
Such an interpretation is legitimate on the basis of rabbinic Hebrew, but probably not on the
basis of BH (see Zoemics, sections 2.4.2, 4.2.9). Thus it appears that we are faced with two cases
where the internalization of “grammatical” rules in post-P texts was linked, ironically, to a
misinterpretation of P. Moreover, in both cases the misinterpretation was made in the direction
of “grammatical hypercorrection”––forcing the originally exceptional cases (a layperson’s
purification offering with sex unspecified in Num 15; a public offering with sex unspecified in
Lev 9) to conform with the majority rule (exclusively female purification offerings for
laypersons; exclusively male public offerings).
5
As Goody has shown, complex systems represented graphically in one society––such as
the formulaic notation of hierarchic properties and jugational patterns in this study––are some-
times expressed orally in other societies (Goody 1977:13).
11:50:13:06:14 Page 200

Page 201
tradition and in the Temple Scroll. Recall that in P, H6 is attested only for p = 0
(that is, only where two Level 0 materials combine to create a Level +1
complex), but the author of the Scroll invents structures in which p = −1,
while the Scroll contains no indication of a conscious perception of the
rule being employed. It thus seems that at least some ritual rules may be
acquired unconsciously, by direct exposure, and thus in a manner similar to
the grammars of natural languages.
Future research may lead to the composition a typology of unconsciously
acquired generative knowledge systems. The acquisition of oral-formulaic
metrical systems, and the unconscious aesthetic patterning of music com-
posers, for example, may fall somewhere in between grammar and the
“grammar of sacrifice.” 6
7.2 APPLICATIONS OF THE PRESENT GRAMMAR
As noted in the , recent scholarship has contributed many

insights into the possible grammar of ritual systems but has not offered a
single coherent, more or less complete grammar of any specific system. Staal
asserts that such a product is the next requisite step in an examination of
ritual grammar, and Ritualdynamik scholarship has further claimed that such
an undertaking is impossible. The first contribution of this study thus lies in
the very fact that such a  has been composed.
7.2.1 The Interpretation of Sacrificial Texts
A grammar of Σ offers multiple paths for future work, both for ancient
Israelite studies and for ritual studies in general.
Despite the fact that animal sacrifice is a system central to the cultures that
produced the biblical text, this institution is still considered obscure by many
modern scholars. Yet, if one wishes to understand the thoughts of the ancients
who composed these formative texts, one must attempt to decipher the inner
workings of their sacrificial systems. Thus, in our continuing endeavor to
understand these authors, we have no choice but to delve into the minute
details that were clearly important to them. In this respect, there is theoretical
value in the identification of the operative categories of Σ’s grammar, just as
in the identification of the categories “phonetics” and “syntax” in linguistics.
Moreover, without the systematic analysis of the pentateuchal sacrificial
6
See Lerdahl and Jackendoff 1983; Gane 2005.
11:50:13:06:14 Page 201

Page 202
rituals in the framework of a grammar, a literal understanding of several

biblical passages is impossible. This grammatical analysis allows for the
interpretation of a number of biblical sacrificial texts that appear to have
been misinterpreted for over two millennia. For example, several words in P
have been misconstrued from the earliest interpreters down to modern
times. Without identifying the binary system in P’s classification of sacrificial
animals, and without noticing P’s careful practice of coining technical terms
out of existing words, one is likely to misconstrue the precise denotation
of nouns like HB d
m$ iK (“female sheep” of any age, not “female lamb”) and
ov d
HT
d mm$Nvo½TBd
x ZEa (“young goat” of either sex, not “female kid”). Similarly,
without identifying the system of hierarchics as defined here, one cannot
correctly interpret the term O$ m in Lev 5:6–7 (“reparation offering,” not
moA
“debt”) or the term HL m oWE in Ezra 8:35 (“wholeburnt+1,” not “wholeburnt0”).
Thus, the grammatical analysis of the Israelite sacrificial system offered here
is a tool for the literal interpretation of biblical texts.
7.2.2 Rituals and Meaning
This study also bears some conclusions about meaning, conclusions that
would have eluded us without the grammatical analysis. In the first place, it
calls into question Staal’s hypothesis concerning the meaninglessness of
ritual. We have seen that at least in some cases, a shift in the conception of the
meaning of sacrifice is interlinked with a shift in its jugational and hierarchic
properties.7 This entails only a partial refutation of Staal’s theory, since it is
still possible to hypothesize that rituals are ab initio meaningless, and that
only at some later historical stage did they come to be considered meaningful.
Furthermore, simultaneous transformations in the form and meaning of a
sacrifice are only rarely traceable.
The fact that this study only partially accords with the results of Staal’s
studies leaves open the question as to how useful grammars of this type might
be for the description of other sacrificial systems. For example, there exists a
fundamental difference between ancient India and the ancient Near East: in
the former, there exist high-resolution parallels between the formal properties
of ritual and language, whereas in the latter the parallels are much less
pronounced. This dissimilarity may be the result of the fact that in ancient
India grammar was a highly developed field of scholarship, whereas in the
ancient Near East it was not.
Nevertheless, we have seen that there is a disconnect of sorts between
the acts (the praxemic sequence) and the meaning of a ritual, regardless of
whether “meaning” is interpreted as “symbolism,” “function,” “desired
7
See §6.3.
11:50:13:06:14 Page 202

Page 203
effect,” or something else. The same praxemic sequence––in the most

extreme example––can be said to be a hatta’t in its praxemics, but an ‘ola
or an asham in its desired effect, in what it symbolizes, or in the function
it fulfills. Much beyond this, we have not solved the mysterious issue of
meaning in ritual in this study, and in fact one of the next requisite steps in
the investigation of ritual meaning is a more precise, unified, and verifiable
working definition of the term “meaning” in a ritual context.
7.3 BEYOND Σ
7.3.1 Further Afield within Σ
Although we have abstracted a multi-tier structure of rules for Σ, the gram-

matical analysis of this ritual system is far from complete. Future investigation
of Σ may reveal other rules or operative categories. For instance, it is possible
that the numeric combinations of zoemes are determined by a small number
of laws, or that the substitutions of one or more zoemes for others are thus
restricted.8
The grammar presented here can also be carried over to other, non-
sacrificial priestly systems, such as ritual purity and impurity, or sex and
food taboos (Lev 11 and 18, 20, respectively). While these systems are not
sacrificial, there are numerous similarities between them and Σ, and they
are believed to be the product of the same authors. The likelihood that the
grammar of Σ may be useful in describing P’s system of purity and impurity,
for example, is suggested by the fact that some of the distinctions cardinal to
Σ ––such as the distinction between age groups, species, and sexes in zoemics;
the concatenation of links to animal and vegetable materials in jugation;
and the physical ritual activity (often described with similar terminology) in
praxemics––are relevant within the system of purity and impurity as well.
If Kunin is correct in detecting a degree of abstraction applicable to a range
of different cultural areas, then there need not be a different model for food,
sacrifice, impurity, and the like; rather, these would be specific instantiations
of a single underlying structural model operative in the ritual of Israelite
society (or of its priestly elite).9
8
For a very general restriction concerning patterns of substitution in sacrifice, see
Lévi-Strauss 1966:223–8.
9
See Kunin 2004:5–28, 38 and in brief Kunin 1995:12–13. If one accepts this model, our Σ
would be equivalent to Kunin’s S3, as it is too specific to be relevant for other ritual systems such
as food taboos and impurity. However, it might be possible to build upon Σ in constructing level
S2; level S1 would pertain to non-Israelite systems as well.
11:50:13:06:14 Page 203

Page 204
7.3.2 A Family Tree of Ritual
Beyond the scope of Σ, it now remains to investigate whether other sacrificial

systems, and perhaps other ritual systems in general, have grammars. If so,
there is no reason to assume a priori that the grammars of two distant ritual
systems are any closer to one another than are, for example, the grammars
of two distantly related languages. It is possible that whole categories (such
as zoemics or hierarchics) may be found in one sacrificial system but not in
others.
Israelite sacrificial ritual also shares many common features with other
ancient Near Eastern sacrificial rituals.10 It is often difficult to determine
whether these similarities result from a genetic relation among the various
systems or from sporadic borrowing over time.
For example, texts composed in Hittite, an Indo-European language, in
the second millennium , offer several striking examples of rituals (par-
ticularly from Kizzuwatna, bordering northern Syria) that parallel sacrificial
procedures described in biblical texts from the first millennium .11
Among other features, an affinity is best evidenced in the manner of blood
treatment and the offering of identical combinations of animals on similar
occasions. Several hypotheses attempt to explain this resemblance. It has
been suggested that the two cultures are evolutionally related––that both
evolved from a common predecessor, perhaps located in northern Syria.
Countering this is the theory that the Hittite ritual system infiltrated the
Levant by a process of borrowing––perhaps through the mediation of first
millennium Neo-Hittite kingdoms in Syria, or by a demographically
insignificant migration of a literary elite as early as the second millennium.12
Since evolutional relationships cannot be determined on the basis of sporadic
similarities between particular rituals (just as evolutional relationships
between natural languages cannot be determined on the basis of particular
shared words), it is necessary to abstract and examine the grammars of
various ritual systems and then compare these systems to determine whether
they are in fact akin.
Eventually, the comparison of such grammars may enable scholars to
formulate a family tree of ritual, as in comparative linguistics. This in turn
could offer insight into the evolution and migration of cultural systems, as
well as into the scope and nature of interaction among several cultures in
antiquity. The present grammar is intended as a first step in this direction.
10
See for example Moyer 1983; Weinfeld 1983; Wright 1986, 1987; Beckman 1983; Schwemer
1995; Feder 2011 (Hittite); Klingbeil 1998; Fleming 2000 (Emar); Scurlock 2002; Scurlock 2006
(Mesopotamian); Pardee 2000; Pardee 2002 (Ugaritic).
11
See n. 000. The affinity is noted already in Sommer and Ehelolf 1924.
12
See Singer 2006 for several possible historical explanations of this connection.
11:50:13:06:14 Page 204

Page 205
7.3.3 Towards a Universal Grammar of Ritual
The other direction of inquiry on the basis of similar future studies in

other systems is the possible formulation of a Universal Grammar of Ritual
(UGR). As noted above, one of scholarship’s fallacies in the field of ritual
studies, appearing already in the work of Hubert and Mauss, is a premature
attempt to formulate a universal pattern underlying geographically disparate
sacrificial rituals. To help us imagine what such a UGR, if it exists, might look
like, let us investigate one possible abstract rule of these systems.
Recall rules 1–3 above in Section 3.8 (Jugation Induced by Jugation). We
claimed that it is possible to abstract a general rule (Rule 3):
if jugate a is located on Level A, subordinate jugational pattern b will be found
on Level B; if jugate a is located on Level B, jugational pattern b will be
modified to b*.
Rule 3 holds for the values a = semolina, b = oil and frankincense, b* = oil (but
not frankincense); as well as for a = wine, b = salt, b*= no salt.
Rule 3 is homologous to the transformational rule found in Staal’s analysis
of the agnicayana ritual wherein a darśapūrn.amāsa is modified (by the
addition of two sāmidhenı̄ verses) only if it is doubly embedded in other
ritual sequences. One could argue, then, that rule 3 is found to be correct for
the values: a = darśapūrn.amāsa, b = 15 sāmidhenı̄ verses, b* = 17 sāmidhenı̄
verses, only that here the relevant levels are not Levels A and B (as in rules
1–3) but Levels B and C.13 This would lead to the formulation of a fourth
rule:
Rule 4: If jugate a is located on level x, it may entail jugational pattern b; if
jugate a is located on the next jugational level (to the right of x), jugational
pattern b will be modified to b*.
Rule 4 is correct in the Israelite system for the values a = semolina, b = oil and
frankincense, b* = oil (but not frankincense); as well as for a = wine, b = salt,
b*= no salt. It is also found to be correct in at least one Vedic tradition for the
values a = darśapūrn.amāsa, b = 15 sāmidhenı̄ verses, b* = 17 sāmidhenı̄ verses.
If Rule 4 were to apply to several other sacrificial systems, it would suggest a
universal property of sacrificial ritual, perhaps one rule in a future Universal
Grammar of Ritual (UGR).
Already in Section 1.3.3, it was noted that the manner in which modifica-
tion takes place in Staal’s agnicayana is analogous to the situation in certain
Note that the analogy made here is between a tiered operative category in Σ (Jugation)
13
and a tiered operative category in the agnicayana ritual examined by Staal, where the relation
between the various elements could be considered somewhere between ancillary (“jugational”)
and constitutive (“hierarchic”).
11:50:13:06:14 Page 205

Page 206
languages where syntax affects morphology. The example offered there

pertains to certain languages in which the location of a verb in specific types
of subordinate clauses entails morphological changes in the verb. This was
demonstrated from the German “Er kommt” (“he is coming”) “Er sagt,
dass er komme” (“he says that he is coming”) and the Akkadian “awı̄lum
ikrub” (“the man is praying”) “awı̄lum ša ikrubu” (“the man who is
praying”).
In these examples, the morphology of the verbs changes because they
are placed within a subordinate clause. In this case at least, the analogy to
the grammar of natural languages is appropriate. These are probably the
transformational rules that Staal has in mind when analyzing the agnicayana
rituals.
Now we may begin to conceive of the form of a Universal Grammar of
Ritual and Language. It might contain highly abstract rules such as
UGRL 1 XA,B,C, . . . JB,C,D,. . . XA,B,C, . . . J*B,C,D,. . .
Read: Element J is modified to J* in the context of immediate subordination to an
element X.
Here, X represents a ritual/linguistic element; * reflects the modified
version; denotes “is modified to.” The index sigla A,B,C, . . . and B,C,D,
. . . (respectively) indicate that the J element is immediately subordinate to
the corresponding X element to its left.14 UGRL 1 happens to be applicable in
the agnicayana (as described by Staal), in Σ and in several natural languages,
but it is not necessarily ubiquitous.
By now it is evident that we are still quite far from reliably postulating a
UGR, not to mention a UGRL. Furthermore, the grammars of ritual systems
are probably similar to those of natural languages only in a very general sense,
as in rule UGRL 1. Further research is needed to determine whether there is
in fact a universal grammar of ritual that transcends sporadic similarities
between historically unrelated ritual systems. At present there is very little
indication of the existence of a UGR.
7.4 THE USE OF THE TERM “GRAMMAR”
Unlike natural languages, Σ clearly has no phonetics, nor is it governed by

morphological or syntactic rules analogous to those of natural languages.15
14
That is, JB is immediately subordinate to XA, JC to XB, JD to XC, etc. See As.t.ādhyāyı̄ 1.3.10
and above, n. 000.
15
Contra Eichinger Ferro-Luzzi 1977; Staal 1979, 1989; Lawson and McCauley 1990; and
Michaels 2007. Nor would one expect a diachronic study of ritual “grammars” to reveal any-
thing analogous to a “Grimm’s law for the equivalent of ritual dialects” (see Goody 1993:42).
11:50:13:06:14 Page 206

Page 207
On the contrary, in Σ one finds at least the four operative categories of

zoemics, jugation, hierarchics, and praxemics. Furthermore, if the Israelite
sacrificial rituals bear any meaning, it is unlikely that this meaning relates to
the formal structure of ritual in any way comparable to the linguistic relation-
ship between morphology, syntax, and semantics.
For this reason, little evidence has been found that would justify a high-
resolution analogy between language and ritual. Much less should one expect
to find a particular affinity between the grammars of languages spoken in
particular cultures and the ritual systems found in those same cultures.16
In fact, the use of the term “syntax” for non-linguistic systems now seems
inappropriate. The present study thus calls into question Staal’s central
thesis, according to which ritual systems display formal properties that
closely resemble the syntactic properties of natural languages. The grammar
of Σ only resembles linguistic grammars inasmuch as both are generative,
rigorous, amenable to concise formulation, partially unconsciously internal-
ized, and have some relation to meaning––and in this respect, the analogy,
first suggested in ancient times, between the formal structure of ritual and
the formal structure of language has proven fruitful. However, the operative
categories in the grammars of this ritual system are different from the
operative categories in the grammars of natural languages.
7.4.1 The Sacrifice of Grammar
Recalling Patañjali’s allegory of the threefold-bound bull (§1.1), we must

recognize the beast’s role as imperfect and structurally deformed. As the
commentators note, several aspects of language lack a corresponding limb in
the allegorical animal. Moreover, the allegory fails to reflect certain internal
relationships: whereas in reality, nominal case endings are appended to
nouns17 and bovine horns are attached to bovine heads, Patañjali’s allegory,
confined as it is within the limits of the hymn, fails to reflect such inner
structural correlations. In fact, if the seven “hands” representing declensions
were to stem from the “horns” representing nouns, the bull would be even
more severely deformed, as in Figure 37.
Furthermore, Patañjali does violence to the original hymn through a
certain backhanded domestication of the beast he describes. In its original
setting, the immediate context of the Vedic verse makes it clear that the
16
Contra Ferro-Luzzi 1977.
17
Note that in the grammar of Pān.ini, vibhakti does not pertain only to nouns, but this does
not affect the statement made here.
11:50:13:06:14 Page 207

Page 208
Fig. 37. Two representations of Patañjali’s allegory.
animal is an adult male (“bull”)18 gaur, a type of wild buffalo (gaura, RV

4.58.2; species Bos gaurus),19 not normally deemed fit for sacrifice.20 However,
shorn of its context and bound within Patañjali’s allegory, this wild buffalo is
reappropriated as a simple bull, domestic and fit for sacrifice. Since in its new
context it is only referred to as a vr.s.abha (nowhere as a gaura), and since this
chapter in Patañjali’s treatise is rich in its sacrificial terminology and imagery,
the natural reading of this verse in its new context is as a reference to the
sacrificial binding of a domestic bull.21
Read in this light, the allegory may serve as a paradigm for all attempts,
including Patañjali’s own, to describe sacrificial rituals in linguistic terms:22
the result is invariably bound to be a deformed monster, prepared for sacrifice
though originally unfit for it.
18
As in English, the term for bull (vr.s.abha) may refer to an adult male bovine, but also to
an adult male of other animals (Apte 531b, see Sarup 1967:165 Sanskrit section, p. 146 English
section).
19
Note that the bull (vr.s.abha) referred to in v. 3 is clearly identical to the four-horned buffalo
.
(cátuh.śr.ngo gaurá) mentioned in v. 2. This wild animal is sometimes described in the R.gveda as
quenching its thirst or fleeing from hunters (e.g., 1.16.5c; 4.21.8c; 8.45.24c; 10.51.6d), which
references seem to imply that the gaur referred to is the wild subspecies (see also 5.78.2, where
gaurs are paired with gazelles; cf. 4.12.6=10.126.8).
20
The gaur is not paśu in the narrow sense of the term, and is therefore unfit for sacrifice; see
B.K. Smith 1994:249–55, particularly 251–2 and 279 n. 29. The gaura (English gaur, Latin Bos
gaurus) is to be distinguished from the mahis.a, more commonly referred to in the RV.
21
See Joshi and Roodbergen 1986:52 n. 163. The bellowing of the bull may be an allusion to
a tradition according to which a sacrificial bull is expected to express its self-dedication for
sacrifice before it is killed. See B.K. Smith 1994:269.
22
It is unlikely that Patañjali had such a reflexive interpretation in mind. However, that he
had a bent for paradoxical self-referential statements is evident from his discussion of unused
words, where his argument for their existence is followed by a list of four examples of such
words (Mah. 1.87, Joshi and Roodbergen 1986:126–7; Chatterji 1964:76–7).
11:50:13:06:14 Page 208

Page 209
Even the term “grammar,” like this two-headed, four-horned, three-legged

monster, is never truly suitable for sacrificial systems. It is merely beneficial
to speak of a grammar of sacrifice in the sense of a finite set of generative
rules that are unconsciously internalized, applied rigorously, and amenable
to concise, formulaic notation. These rules determine what is acceptable
(“grammatical”) within a particular ritual system and place certain con-
straints on the system’s development over the course of time.
Having shown some of the ways in which the grammar of ritual and the
grammar of natural language are fundamentally dissimilar, we are forced to
wonder whether the term “grammar” is even appropriate in the case of ritual
at all. We have noted that following the analogy to language too rigorously
has misled scholarship in its study of sacrificial systems. Therefore, having
identified the categories of Σ and having coined new terms for them, it might
now be wiser to discard this misleading term, “grammar,” or replace it with
an alternative and more accurate term such as GRAMMUR––an acronym
for Generative, Rigorously Applied, Mathematically Modeled, Unconscious
Rules.
Nevertheless, aside from the term’s widespread use in non-linguistic
contexts,23 there is one substantial reason to retain the term “grammar.”
Few systems that are neither linguistic nor ritual are describable in terms of
generative rules. This does not mean that other systems do not have
grammars (in fact, it is highly likely that other systems like music, dance,
and certain games are based on grammars of their own), but very few systems
are as rigorously rule-governed, and at the same time as complex and well
documented over a long period of historical time, as ritual and linguistic
systems.
Thus, in order to speak of a ritual system’s grammar, it is necessary to
sacrifice the narrow denotation of “grammar.” In sacrifice, one gives up some
of one’s possessions in order to obtain a higher objective; here, one must give
up the narrow denotation of a term in order to be granted an entry into the
inner workings of an ancient culture, as well as a more accurate interpretation
of its formative texts.
23
See n. 000. OED s.v. grammar (6.a), for example, reflects the expanded use of “grammar”
to denote the “fundamental principles or rules of an art or science.”
11:50:13:06:14 Page 209

Page 210
Bibliography
Achenbach, Reinhard. Die Vollendung der Tora: Studien zur Redaktionsgeschichte des
Numeribuches im Kontext von Hexateuch und Pentateuch. Beihefte zur Zeitschrift für
altorientalische und biblische Rechtsgeschichte 3. Wiesbaden: Harrassowitz, 2003.
Albeck, Chanoch. Mavo la-Mishnah [Introduction to the Mishna]. Jerusalem: Mosad
Bialik, 1956.
Alford, John A. “The Grammatical Metaphor: A Survey of Its Use in the Middle Ages.”
Speculum 57 (1982): 728–60.
Allen, Leslie C., and Timothy S. Laniak. Ezra, Nehemiah, Esther: Based on the New
International Version. New International Biblical Commentary, Old Testament
Series 9. Peabody, MA: Hendrickson; Carlisle, Cumbria: Paternoster Press, 2003.
Anderson, Gary A. Sacrifices and Offerings in Ancient Israel: Studies in Their Social
and Political Importance. Harvard Semitic Monographs 41. Atlanta: Scholars Press,
1987.
Arnaud, Daniel. Recherches au pays d’Aštata: Emar. 4 vols. Orbis biblicus et orientalis,
Series Archaeologica 20. Paris : Editions Recherche sur les civilisations, 1985.
Asad, Talal. Genealogies of Religion: Discipline and Reasons of Power in Christianity and
Islam. Baltimore: Johns Hopkins University Press, 1993.
Ashley, Timothy R. The Book of Numbers. The New International Commentary on the
Old Testament. Grand Rapids: Eerdmans, 1993.
Austin, J.L. How to Do Things with Words. Edited by J.O. Urmson and Marina Sbisà.
2nd edn. Cambridge MA: Harvard University Press, 1975.
Avigad, Nahman, and Yigael Yadin. A Genesis Apocryphon: A Scroll from the Wilderness
of Judaea. Translated by Sulamith Schwartz Nardi. Jerusalem: Magnes, 1956.
Baden, Joel S. J, E, and the Redaction of the Pentateuch. FAT 68. Tübingen: Mohr
Siebeck, 2009.
Bar Asher, Moshe. “The Bible Interpreting Itself.” In Rewriting and Interpreting the
Hebrew Bible: The Biblical Patriarchs in Light of the Dead Sea Scrolls, edited by
Devorah Dimant and Reinheird G. Kratz, 1–18. Berlin: De Gruyter, 2013.
Barkay, G. et al. “The Amulets from Ketef Hinnom: A New Edition and Evaluation.”
BASOR 334 (2004): 41–71.
Batten, Loring W. A Critical and Exegetical Commentary on the Books of Ezra and
Nehemiah. ICC 12. Edinburgh: T. & T. Clark, 1913.
Becker, Joachim. Esra/Nehemia. Würzburg: Echter, 1990.
Beckman, Gary M. Hittite Birth Rituals. Wiesbaden: O. Harrassowitz, 1983.
Bell, Catherine. Ritual Theory, Ritual Practice. New York: Oxford University
1992.
Bertheau, Ernst. Die Bücher Esra, Nechemia und Ester. Leipzig: Hirzel, 1887.
Blenkinsopp, Joseph. Ezra-Nehemiah: A Commentary. OTL. Philadelphia: West-
minster Press, 1988.
—— “An Assessment of the Alleged Pre-Exilic Date of the Priestly Material in the
Pentateuch.” ZAW 108 (1996): 495–518.
11:50:13:06:14 Page 210

Page 211
Bibliography 211
Block, N., ed., Readings in Philosophy of Psychology. Vol. 2. Cambridge, MA: Harvard
University Press, 1981.
—— “Introduction: What is Innateness?” In Block 1981a: 279–81.
Boas, Franz. The Mind of Primitive Man. New York: MacMillan, 1911.
Bodewitz, H.W. The Daily Evening and Morning Offering (Agnihotra) according to the
Brahmanas. Orientalia Rheno-Traiectina 21. Leiden: Brill, 1976.
Boorer, Suzanne. “The Place of Numbers 13–14* and Numbers 20:2–12* in the
Priestly Narrative (Pg).” JBL 131 (2012): 45–63.
Bourdieu, Pierre. Outline of a Theory of Practice. Cambridge Studies in Social and
Cultural Anthropology 16. Cambridge: University Press, 1977.
Breneman, Mervin. Nehemiah, Esther. New American Commentary 10. Nashville,
TN: Broadman & Holman, 1993.
Brichto, H.C. “On Slaughter and Sacrifice, Blood and Atonement,” HUCA 47 (1976):
19–56.
Brock, S.P. Testamentum Iobi. Pseudepigrapha veteris testamenti graece 2. Leiden:
Brill, 1967.
Brockington, Leonard Herbert, ed. Nehemiah and Esther. Century Bible. New series.
London: Nelson, 1969.
Bronkhorst, Johannes, ed. Mahābhās.yadı̄pikā of Bhartr.hari: Fascicule IV Āhnika I.
Bhandarkar Oriental Research Institute Post-Graduate and Research Department
Series 28. Poona: Bhandarkar Oriental Research Institute, 1987.
Budd, Philip J. Leviticus. London: Pickering, 1996.
Burkert, Walter. Structure and History in Greek Mythology and Ritual. Sather Classical
Lectures 47. Berkeley: University of California Press, 1979.
—— “The Problem of Killing.” In Hamerton-Kelly 1987: 149–76.
Caland, W., trans. Das Śrautasūtra des Āpastamba. Göttingen: Vandenhoeck &
Ruprecht, 1921.
Cardona, George. “On Attitudes towards Language in Ancient India.” Sino-Platonic
Papers 15 (1990): 1–19.
Carr, David M. “Changes in Pentateuchal Scholarship.” In The History of Old
Testament Interpretation, volume 3.1, The Twentieth Century, edited by M. Saebø.
Göttingen: Vandenhoeck & Ruprecht, 2014 [forthcoming].
Chakravarti, Prabhatchandra. “Patañjali.” IHQ 2 (1926): 262–89.
Chatterji, Kshitish Chandra. Patanjali’s Mahabhashya: Paspasāhnika. 3rd edn. Usha
Memorial Series 7. Calcutta: Mukherjee, 1964.
Cholewiński, Alfred. Heiligkeitsgesetz und Deuteronomium: Eine vergleichende
Studie. Analecta Biblica 66. Rome: Biblical Institute, 1976.
Chomsky, Noam. Syntactic Structures. Janua linguarum 4. Leiden: Mouton, 1957.
—— “Discussion of Putnam’s Comments.” In Block 1981a: 349–58.
—— “On Cognitive Capacity.” In Block 1981a: 305–23.
—— “Reply to Putnam.” In Block 1981a: 300–4.
—— Language and Problems of Knowledge. Cambridge, MA: MIT Press, 1988.
—— and Morris Halle. The Sound Pattern of English. Cambridge, MA: MIT Press,
1991.
Clason, A.T., and H. Buitenhuis. “Patterns in Animal Food Resources in the Bronze
Age in the Orient.” In Archaeozoology of the Near East III: Proceedings of the
11:50:13:06:14 Page 211

Page 212
212 Bibliography
Third International Symposium on the Archaeozoology of Southwestern Asia and
Adjacent Areas, edited by H. Buitenhuis, L. Bartosiewicz, and A.N. Choyke, 233–42.
Groningen: Arc, 1988.
Clines, David J. A. Nehemiah and Esther: Based on the Revised Standard Version. New
Century Bible Commentary. Grand Rapids: Eerdmans; London: Morgan & Scott,
1984.
Cohn, Naftali S. The Memory of the Temple and the Making of the Rabbis. Philadelphia:
University of Pennsylvania Press, 2013.
Daniel-Nataf, Suzanne, ed. Ketavim: Filon ha-Aleksandroni. Vol. 2, Sefer ha-Mitzvot.
Jerusalem: Mosad Bialik, 1991.
Davidson, Donald, and Gilbert Harman. The Logic of Grammar. Encino, CA:
Dickenson, 1975.
Davis, S.J.M. “ ‘Thou shalt take of the ram . . . the right thigh; for it is a ram of
consecration . . .’: Some Zoo-Archaeological Examples of Body-Part Preferences.”
In Uomini, piante e animali nella dimensione del sacro: Seminario di studi di
bioarcheologia (28–29 giugno 2002), Convento dei Domenicani-Cavallino (Lecce),
edited by D’Andria, Francesco, Jacopo De Grossi Mazzorin, Girolamo Fiorentino,
63–70. Bari: Edipuglia, 2008.
de Moor, J.C. “The Peace-Offering in Ugarit and Israel.” In Schrift en uitleg, 112–17.
Gispen-Festschrift. Kampen: Kok, 1970.
de Saussure, Ferdinand. Course in General Linguistics. Edited by Charles Bally and
Albert Reidlinger. Translated by Wade Baskin. New York: Philosophical Library,
1959.
del Olmo Lete, G. “gdlt/dqt: ¿ganado o pan como materia sacrificial?” Aula Orientalis
25 (2007): 169–173.
Detienne, Marcel, and Jean-Pierre Vernant (eds). The Cuisine of Sacrifice among
the Greeks. Translated by Paula Wissing. Chicago: University of Chicago Press,
1989.
Dietrich, M., O. Loretz, and J. Sanmartín, “Bemerkungen zu RS 24.260 = UG 5., S. 586
NR. 11.” UF 7 (1975): 543–4.
—— “Zu šlm kll im Opfertext von Marseilles (CIS I 165).” UF 7 (1975): 561–2.
Dillmann, August. Numeri Deuteronomuim und Josua. Leipzig: Hirzel, 1886.
Dion, P.E. “Early Evidence for the Ritual Significance of the ‘Base of the Altar.’ ” JBL
106 (1987): 487–92.
Dohmen, C. “XBZMH LE: Zur Bedeutung und Verwendung von hebr. LE.” BN 16
(1981): 7–10.
Dorival, Gilles. Les Nombres. La Bible d’Alexandrie 4. Paris: CERF, 1994.
Douglas, Mary. Purity and Danger: An Analysis of Concepts of Pollution and Taboo.
London: Routledge & Kegan Paul, 1966.
—— Leviticus as Literature. Oxford: Oxford University Press 1999.
Dozeman, Thomas B., Konrad Schmid, and Baruch J. Schwartz, eds. The Pentateuch:
International Perspectives on Current Research. FAT 78. Tübingen: Mohr Siebeck,
2011.
Eberhart, Christian A. Studien zur Bedeutung der Opfer im Alten Testament: Die
Signifikanz von Blut- und Verbrennungsriten im kultischen Rahmen. WMANT 94.
Neukirchen–Vluyn: Neukirchener, 2002.
11:50:13:06:14 Page 212

Page 213
Bibliography 213
—— Ritual and Metaphor: Sacrifice in the Bible. Resources for Biblical Study 68.
Atlanta: SBL, 2011.
—— “Sacrifice? Holy Smokes! Reflections on Cultic Terminology for Understanding
Sacrifice in the Hebrew Bible.” In Eberhart 2011: 17–32.
Elliger, K. Leviticus. HAT 4. Tübingen: J.C.B. Mohr, 1966.
Evans-Pritchard, E.E. Nuer Religion. Oxford: Clarendon, 1956.
—— Theories of Primitive Religion. Oxford: Clarendon, 1965.
Feder, Yitzhak. Blood Expiation in Hittite and Biblical Ritual: Origins, Context, and
Meaning. Writings from the Ancient World Supplement Series 2. Atlanta: SBL,
2011.
Feeley-Harnik, Gillian. The Lord’s Table: Eucharist and Passover in Early Christianity.
Philadelphia: University of Pennsylvania Press, 1981.
Flavius Josephus, Judean Antiquities 1–4: Translation and Commentary by Louis Feld-
man. Flavius Josephus: Translation and Commentary 3. Leiden: Brill, 2000.
Fensham, F. Charles. The Books of Ezra and Nehemiah. NICOT. Grand Rapids: Eerd-
mans, 1982.
Ferro-Luzzi, Gabriella Eichinger. “Ritual as Language: The Case of South Indian Food
Offerings.” Current Anthropology 18 (1977): 507–14.
Finkelstein, Louis. Sifra de-ve Rav: Ve-hu Sefer Torat Kohanim 'a. p. ket. y. Romi.
New York. Jewish Theological Seminary of America, 1989.
Fitzmyer, Joseph A. The Genesis Apocryphon of Qumran Cave I: A Commentary.
2nd ed. Rome: Biblical Institute Press, 1971.
Fleming, Daniel E. Time at Emar: The Cultic Calendar and the Rituals from the
Diviner’s Archive. Winona Lake: Eisenbrauns, 2000.
Fodor, J.A., Th.G. Bever, and M.F. Garrett, “The Specificity of Language Skills.” In
Block 1981: 324–38.
Frege, Gottlob. “On Sense and Reference.” Translated by Max Black. In Davidson and
Harman 1975: 116–28.
Gadegaard, Norman. “On the So-called Burnt Offering Altar in the Old Testament.”
PEQ 110 (1978): 35–45.
Gane, Roy E. Ritual Dynamic Structure. Gorgias Dissertations 14.2. Piscataway:
Gorgias, 2004.
—— Cult and Character: Purification Offerings, Day of Atonement, and Theodicy.
Winona Lake: Eisenbrauns, 2005.
—— “Privative Preposition IM in Purification Offering Pericopes and the Changing
Face of ‘Dorian Gray’.” JBL 127 (2008): 209–22.
Geertz, Clifford. The Interpretation of Cultures. New York: Basic Books, 1973.
Geldner, Karl Friedrich. Der Rig-Veda. 4 vols. Harvard Oriental Series. Cambridge,
MA: Harvard University Press, 1951–57.
Georgourdi, Stella. “Sanctified Slaughter in Modern Greece.” in Detienne and
Vernant 1989: 183–203.
Gerstenberger, Erhard S. Das dritte Buch Mose: Leviticus. ATD 6. Göttingen: Vanden-
hoeck & Ruprecht, 1993.
Gertz, Jan Christian. “Source Criticism in the Primeval History of Genesis: An
Outdated Paradigm for the Study of the Pentateuch?” In Dozeman et al. 2011:
169–80.
11:50:13:06:14 Page 213

Page 214
214 Bibliography
Gilders, William K. Blood Ritual in the Hebrew Bible: Meaning and Power. London:
Johns Hopkins University Press, 2004.
—— “Jewish Sacrifice: Its Nature and Function (According to Philo).” In Knust and
Várhelyi 2011: 94–105.
Girard, René. Violence and the Sacred. Translated by Patrick Gregory. Baltimore: Johns
Hopkins University Press, 1977.
—— “Generative Scapegoating.” In Hamerton-Kelly 1987: 73–105.
Goldschmidt, E.D. The Passover Haggadah: Its Sources and History. Jerusalem: Bialik,
1960 [in Hebrew].
Gonda, J. “The Ritual Sutra.” In Veda and Upanishads: A History of Indian Literature.
2 vols. Wiesbaden: O. Harrassowitz, 1.465–488, 1975–77.
Goody, Jack Rankine. “British Functionalism.” In Main Currents in Cultural Anthro-
pology, edited by Raoul Naroll and Frada Naroll, 185–215, Englewood Cliffs, NJ:
Prentice-Hall, 1973.
—— The Domestication of the Savage Mind. Themes in the Social Sciences.
Cambridge and New York: Cambridge University Press, 1977.
—— “Knots in May: Continuities, Contradictions and Change in European Rituals.”
Journal of Mediterranean Studies 3 (1993): 30–45.
—— Myth, Ritual and the Oral. Cambridge: Cambridge University Press, 2010.
Gordis, Robert. Koheleth, The Man and His World. New York: Jewish Theological
Seminary of America, 1951.
Grabbe, Lester L., ed. Can a “History of Israel” Be Written? JSOTSup 245. Sheffield:
Sheffield Academic Press, 1997.
—— Ancient Israel: What Do We Know and How Do We Know It? London; New York:
T. & T. Clark, 2007.
Gray, George Buchanan. A Critical and Exegetical Commentary on Numbers. ICC.
New York: Scribner’s, 1903.
Gray, John. “Cultic Affinities between Israel and Ras Shamra.” ZAW 62 (1950):
207–20.
Greenberg, Moshe. “Idealism and Practicality in Numbers 35:4–5 and Ezekiel 48.”
JAOS 88 (1968): 59–63.
Greenfield, Jonas C., Michael E. Stone, and Ester Eshel, eds. The Aramaic Levi
Document: Edition, Translation, Commentary. Leiden: Brill, 2004.
Grimes, Ronald L. Ritual Criticism: Case Studies in Its Practice, Essays on Its Theory.
Columbia, SC: University of South Carolina Press, 1990.
Gruenwald, Ithamar. Rituals and Ritual Theory in Ancient Israel. Leiden: Brill,
2003.
Gurney, O.R. Some Aspects of Hittite Religion. Oxford: Oxford University Press,
1977.
Halbertal, Moshe. On Sacrifice. Princeton: Princeton University Press, 2012.
Hamerton-Kelly, Robert G., ed. Violent Origins: Walter Burkert, René Girard and
Jonathan Z. Smith on Ritual Killing and Cultural Formation. Stanford: Stanford
Handelman, Don. Models and Mirrors: Towards an Anthropology of Public Events.
Cambridge: Cambridge University Press, 1990.
Haran, Menah.em. “Mizbeah..” In Entsik.lopediya mik.rait, 8. 763–80. 1962.
11:50:13:06:14 Page 214

Page 215
Bibliography 215
—— Temples and Temple-Service in Ancient Israel: An Inquiry into the Character of
Cult Phenomena and the Historical Setting of the Priestly School. Oxford: Clarendon,
1978.
—— “Behind the Scenes of History: Determining the Date of the Priestly Source.”
JBL 100 (1981): 321–33.
—— “Ezekiel, P, and the Priestly School.” Vetus Testamentum 58 (2008): 211–18.
—— The Bible and Its World: Selected Literary and Historical Studies. Jerusalem:
Magnes, 2009.
Harlé, Paul, and Didier Pralon, Le Lévitique. La Bible d’Alexandrie 3. Paris: CERF,
1988.
Harris, Marvin. Cows, Pigs, Wars and Witches. New York: Vintage Books, 1974.
Hartley, John E. Leviticus. WBC. Dallas: Word, 1992.
Hastings, Adi. “From Ritual to Grammar: Sacrifice, Homology, Metalanguage.”
Language and Communication 23 (2003): 275–85.
Heesterman, J.C. The Broken World of Sacrifice. Chicago: University of Chicago Press,
1993.
Heinsohn, G. “The Rise of Blood Sacrifice and Priest-Kingship in Mesopotamia: A
‘Cosmic Decree’?” Religion 22 (1992): 110–34.
Himmelfarb, Martha. “Earthly Sacrifice and Heavenly Incense: The Law of the Priest-
hood in Aramaic Levi and Jubilees.” In Heavenly Realms and Earthly Realities in
Late Antique Religions, edited by Ra‘anan S. Boustan and Annette Yoshiko Reed,
103–22. Cambridge: Cambridge University Press, 2004.
Hockett, Charles F. “The Origin of Speech.” Scientific American 203 (1960): 89–96.
Houben, Jan E.M. “Formal Structure and Self-referential Loops in Vedic Ritual.” In
Michaels and Mishra 2010: 29–63.
Hubert, H., and M. Mauss. “Essai sur la nature et la fonction du sacrifice.” In Mélanges
d’histoire des religions: De quelques résultats de la sociologie religieuse, 1–130. 2nd
edn. Paris: F. Alcan, 1929. Originally published as Henri Hubert et Marcel Mauss,
“Essai sur la nature et la fonction du sacrifice.” L’Année Sociologique 2 (1898):
29–138.
—— and —— Sacrifice: Its Nature and Function. Translated by W.D. Halls. London:
Cohen and West, 1964.
Humphrey, Caroline, and James Laidlaw. The Archetypal Actions of Ritual: A Theory
of Ritual Illustrated by the Jain Rite of Worship. Oxford: Clarendon Press,
1994.
Hurvitz, Avi. “The Evidence of Language in Dating the Priestly Code.” RB 81 (1974):
24–57.
—— “Once Again: The Linguistic Profile of the Priestly Material in the Pentateuch
and Its Historical Age: A Response to J. Blenkinsopp.” ZAW 112 (2000): 180–91.
Iamblichus, On the Mysteries. Translated by Emma C. Clarke, John M. Dillon, and
P. Hershbell. Writings from the Greco-Roman World 4. Atlanta: SBL, 2003.
Japhet, Sara. I & II Chronicles. OTL. London: SCM, 1993.
Jastrow, Marcus. A Dictionary of the Targumim, the Talmud Babli and Yerushalmi, and
the Midrashic Literature. New York: Title, 1943.
Josephus, Jewish Antiquities: Books IX–XI. Translated by Ralph Marcus. Loeb Classical
Library 326. Cambridge, MA: Harvard University Press.
11:50:13:06:14 Page 215

Page 216
216 Bibliography
Joshi, S.D., and J.A.F. Roodbergen. Patañjali’s Vyākaran.a-Mahābhās.ya. Vol. 1,
Paspaśāhnika. Publications of the Centre of Advanced Study in Sanskrit 15. Pune:
University of Poona, 1986.
Kafah., Yosef David. Mishna ‘im perush Mosheh ben Maymon: Mak.or ve-targum.
Jerusalem: Mosad ha-Rav K.uk., 1963.
Kämmerer, Thomas Richard, ed. Studien zu Ritual und Sozialgeschichte im Alten
Orient: Tartuer Symposien 1998–2004. BZAW 374. Berlin: Walter de Gruyter, 2007.
Katz, J.J. “Innate Ideas.” In Block 1981: 282–91.
Keil, C.F. The Books of Ezra, Nehemiah, and Esther. Translated by Sophia Taylor.
Edinburgh: T. & T. Clark, 1888.
Keil, C.F., and F. Delitzsch. Commentary on the Old Testament. Vol. 1, The Pentateuch.
Translated by J. Martin. 10 vols. Grand Rapids: Eerdmans, 1986.
Kellerman, Diether. “Bemerkungen zum Sündopfergesetz in Num 15,22 ff.” In Wort
und Geschichte, 107–13. AOAT 18. Kevelaer: Butgzon & Bercker, 1973.
Kettlewell, Percy W.H. The Books of Ezra and Nehemiah: With Intro-
duction, Notes, Maps and Appendices. London: Rivingtons, 1906.
Kidner, Derek. Ezra and Nehemiah: An Introduction and Commentary. Downers
Grove, IL: Inter-Varsity Press, 1979.
Kiuchi, Nobuyoshi. Leviticus. Apollos Old Testament Commentary. Nottingham:
Apollos, 2007.
Klawans, Jonathan. Purity, Sacrifice, and the Temple: Symbolism and Supersessionism in
the Study of Ancient Judaism. Oxford and New York: Oxford University Press, 2006.
—— “Methodology and Ideology in the Study of Priestly Ritual.” In Schwartz et al.
2008: 84–95.
—— “Symbol, Function, Theology, and Morality in the Study of Priestly Ritual.” In
Knust and Várhelyi 2011: 106–22.
Klein, Michael. The Fragment-Targums of the Pentateuch according to their Extant
Sources. 2 vols. Analecta Biblica 76. Rome: Biblical Institute Press, 1980.
Klingbeil, Gerald A. A Comparative Study of the Ritual of Ordination as Found in
Leviticus 8 and Emar 369. Lewiston: Mellen, 1998.
—— Bridging the Gap: Ritual and Ritual Texts in the Bible. Bulletin for Biblical
Research Supplements 1. Winona Lake: Eisenbrauns, 2007.
Knierim, Rolf P. Text and Concept in Leviticus 1:1–9: A Case in Exegetical Method. FAT
2. Tübingen: J.C.B. Mohr, 1992.
Knobel, August. Numeri, Deuteronomium und Josua. Kurzgefasstes exegetisches
Handbuch zum Alten Testament. Dreizehnte Lieferung. Leipzig: Verlag von S. Hirzel,
1861.
Knohl, Israel. “The Acceptance of Sacrifices from Gentiles.” Tarbiz. 48 (1978): 341–5
[in Hebrew].
—— “The Priestly Torah versus the Holiness School: Sabbath and the Festivals.”
HUCA 58 (1987): 65–117.
—— “The Sin-Offering Law in the ‘Holiness School’.” Tarbiz. 59 (1990): 1–10 [in
Hebrew].
—— The Sanctuary of Silence: The Priestly Torah and the Holiness School.
Minneapolis: Fortress, 1995.
—— and Shlomo Naeh. “Milluim Ve-Kippurim.” Tarbiz. 62 (1992): 17–44.
11:50:13:06:14 Page 216

Page 217
Bibliography 217
Knust, Jennifer Wright, and Zsuzsanna Várhelyi, eds. Ancient Mediterranean Sacrifice.
New York: Oxford University Press, 2011.
Koch, Klaus. Die Priesterschrift von Exodus 25 bis Leviticus 16: Eine Überlieferungs-
geschichtliche und literarkritische Untersuchung. Göttingen: Vandenhoeck & Ruprecht,
1959.
Kornfeld, W. Levitikus. Die Neue Echter Bibel 6. Würzburg: Echter, 1983.
Kraft, Robert A. The Testament of Job: According to the SV Text: Greek Text and English
Translation. Pseudepigrapha Series 4. Missoula: SBL, 1975.
Kratz, Reinhard G. “The Pentateuch in Current Research: Consensus and Debate.” In
Dozeman et al. 2011: 31–62.
Kreinath, Jens. “Semiotics.” In Kreinath et al. 2006: 429–70.
—— Jan Snoek, and Michael Stausberg, eds. Theorizing Rituals: Issues, Topics,
Approaches, Concepts. Numen Book Series; Studies in the History of Religions 114.
Leiden: Brill, 2006.
Kunin, Seth Daniel. The Logic of Incest: A Structuralist Analysis of Hebrew Mythology.
JSOTSup 185. Sheffield: Sheffield Academic Press, 1995.
—— We Think What We Eat: Neo-Structuralist Analysis of Israelite Food Rules and
Other Cultural and Textual Practices. JSOTSup 412. London and New York: T. & T.
Clark, 2004.
—— “Neo-Structuralism and the Contestation of Sacred Place in Biblical Israel.”
Temenos: Nordic Journal of Comparative Religion 41, no. 2 (2005): 203–24.
Lakoff, George. “Empiricism without Facts (Review of Charles F. Hockett, The State of
the Art).” Foundations of Language 5 (1969): 118–27.
Langer, Susanne K. Philosophy in a New Key: A Study of the Symbolism of Reason, Rite
and Art. New York: The New American Library, 1948.
Lawson, E. Thomas. “Ritual as Language.” Religion 6 (1976): 39–123.
—— and Robert N. McCauley. Rethinking Religion: Connecting Cognition and
Culture. Cambridge: Cambridge University Press, 1990.
Leach, Edmund. “Anthropological Aspects of Language: Animal Categories and
Verbal Abuse.” In New Directions in the Study of Language, edited by E.H.
Lenneberg, 23–63. Cambridge: MIT Press, 1964.
—— Culture and Communication: The Logic by which Symbols Are Connected.
Cambridge: Cambridge University Press, 1976.
Lerdahl, Fred, and Ray S. Jackendoff. A Generative Theory of Tonal Music. Cambridge,
MA: MIT Press, 1983.
Levering, Matthew. Ezra and Nehemiah. Brazos Theological Commentary on the
Bible. Grand Rapids: Brazos, 2007.
Levine, Baruch A. In the Presence of the Lord: A Study of Cult and Some Cultic Terms in
Ancient Israel. Studies in Judaism in Late Antiquity. Leiden: Brill, 1974.
—— Leviticus. Philadelphia: JPS, 1989.
—— “Ritual as Symbol: Modes of Sacrifice in Israelite Religion.” In Sacred Time,
Sacred Place, edited by B.M. Gittlen, 125–35. Winona Lake: Eisenbrauns,
2002.
—— Numbers: A New Translation with Introduction and Commentary. 2 vols. 4, 4A.
New York: Doubleday, 1993–2000.
—— and W. Hallo, “Offerings to the Temple Gates at Ur.” HUCA 38 (1967): 17–58.
11:50:13:06:14 Page 217

Page 218
218 Bibliography
Lévi-Strauss, Claude. Anthropologie Structurale. Paris: Plon, 1958.
—— Totemism. Translated by Rodney Needham. Boston: Beacon, 1963; originally
published as Le Totémisme aujourd’hui. Mythes et religions 42. Paris: Presses
universitaires de France, 1962.
—— Structural Anthropology. Translated by C. Jacobson and B.G. Schoepf. 2 vols.
New York: Basic Books, 1963–76.
—— “The Structural Study of Myth.” In Lévi-Strauss 1963a: 206–31.
—— “Structural Analysis in Linguistics and Anthropology.” In Lévi-Strauss 1963a:
31–54; originally published as “L’analyse structurale en linguistique et en anthro-
pologie.” Word 1 (1945): 1–21.
—— The Savage Mind. The Nature of Human Society Series. Chicago: University of
Chicago Press, 1966.
—— “Four Winnebago Myths: A Structural Sketch.” In Structural Anthropology,
translated by Monique Layton, 198–210. Vol. 2. Chicago: University of Chicago
Press, 1976.
—— The Naked Man. Translated by John and Doreen Weightman. New York: Harper
& Row, 1981; originally published as: L’homme Nu. Paris: Plon, 1971.
—— The Jealous Potter. Translated by Bénédicte Chorier. Chicago: University of
Chicago Press, 1988.
Levine, Baruch. “Ugaritic Descriptive Rituals.” Journal of Cuneiform Studies 17 (1963):
105–111.
Lipiński, Edward. “Burnt offering of Head, Peder, and Kidneys.” In Birkat Shalom:
Studies in the Bible, Ancient Near Eastern Literature, and Postbiblical Judaism
Presented to Shalom M. Paul on the Occasion of His Seventieth Birthday, edited by
Chaim Cohen et al., 59–68. Winona Lake: Eisenbrauns, 2008.
Liss, Hanna. “ ‘Describe the Temple to the House of Israel’: Preliminary Remarks on
the Temple Vision in the Book of Ezekiel and the Question of Fictionality in
Priestly Literatures.” In Utopia and Dystopia in Prophetic Literature, edited by Ehud
Ben Zvi, 122–43. Helsinki: Finnish Exegetical Society; Göttingen: Vandenhoeck &
Ruprecht, 2006.
Loretz, O. “Der hebräische Opferterminus kljl ‘Ganzopfer.’ ” UF 7 (1975): 569.
Lupu, Eran. Greek Sacred Law: A Collection of New Documents (NGSL). Religions in
the Graeco-Roman World 152. Leiden: Brill, 2005.
Machiela, Daniel A. The Dead Sea Genesis Apocryphon: A New Text and Trans-
lation with Introduction and Special Treatment of Columns 13–17. Leiden: Brill,
2009.
Magen, Yitzhak. Mount Gerizim Excavations. Vol 2: A Temple City. JSP 8. Jerusalem:
Israel Antiquities Authority, 2008.
Maier, Gerhard. Das dritte Buch Mose. Wuppertaler Studienbibel. Wuppertal:
Brockhaus, 1994.
Maimonides, Moses. The Guide of the Perplexed. Translated by Shlomo Pines. Chicago:
University of Chicago Press, 2005, 1964.
Malinowski, Bronislaw. Magic, Science and Religion and Other Essays. Boston: Beacon,
1948.
Margueron, Jean Claude. “L’espace sacrificiel dans le Proche-Orient ancient.” In
L’Espace sacrificiel dans les civilisations méditerranéennes de l’antiquité : actes du
11:50:13:06:14 Page 218

Page 219
Bibliography 219
colloque tenu à la Maison de l’Orient, Lyon, 4–7 juin 1988, edited by Roland Étienne
and Marie-Therese Le Dinahet, 235–42. Publications de la Bibliothèque Salomon-
Reinach. Université Lumière–Lyon 2. Paris: Diffusion de Boccard, 1991.
Marx, Alfred. Les offrandes végétales dans l’Ancien Testament: du tribut d’hommage au
repas eschatologique. Leiden and New York: Brill, 1994.
—— “The Theology of Sacrifice according to Leviticus 1–7.” In Rendtorff and Kugler
2003: 103–20.
—— Les systèmes sacrificiels de l’Ancien Testament: formes et fonctions du culte
sacrificiel à Yhwh. VTSup 105. Leiden: Brill, 2005.
Matthews, Peter H. A Short History of Structural Linguistics. Cambridge: Cambridge
Mazar, B. “The Excavations in the Old City of Jerusalem.” Eretz Israel 9 (1969):
161–174. (Hebrew Section)
Mazzotta, Giuseppe. Dante’s Vision and the Circle of Knowledge. Princeton: Princeton
McClymond, Kathryn. Beyond Sacred Violence: A Comparative Study of Sacrifice.
Baltimore: Johns Hopkins University Press, 2008.
Meshel, Naphtali S. Review of Roy Gane, Cult and Character: Purification Offerings,
Day of Atonement, and Theodicy (Winona Lake: Eisenbrauns, 2005). Journal of the
American Oriental Society 127 (2007): 203–5.
—— “Food for Thought: Systems of Categorization in Leviticus 11.” HTR 101
(2008): 203–29.
—— “P1, P2, P3, and H: Purity, Prohibition, and the Puzzling History of Leviticus 11.”
Hebrew Union College Annual 81 (2010): 1–15.
—— “The Form and Function of a Biblical Blood Ritual.” Vetus Testamentum 63
(2013): 276–89.
—— “Towards a ‘Grammar’ of Sacrifice: Hierarchic Patterns in the Israelite Sacrificial
System.” JBL 132 (2013): 543–67.
Michaels, Axel. “Sam. kalpa: The Beginning of a Ritual.” In Words and Deeds, edited by
Jörg Gengnagel, Ute Hüsken, and Srilata Raman, 45–63. Wiesbaden: Harrassowitz,
2005.
—— “Ritual and Meaning.” In Kreintah et al. 2006: 247–61.
—— “ ‘How Do You Do?’: Vorüberlegungen zu einer Grammatik der Rituale.” In Der
Mensch––ein “animal symbolicum?” Sprache––Dialog––Ritual, edited by Heinrich
Schmidinger and Clemens Sedmak, 239–58. Darmstadt: WBG, 2007.
—— “The Grammar of Rituals.” In Michaels and Mishra 2010: 7–28.
—— and Anand Mishra. Grammars and Morphologies of Ritual Practices in Asia.
Ritual Dynamics and the Science of Ritual 1. Wiesbaden: Harrassowitz, 2010.
Milbank, John. “Stories of Sacrifice: From Wellhausen to Girard.” Theory, Culture and
Society 12 (1995): 15–46.
Milgrom, Jacob. “Two Kinds of H . at.t.āt.” VT 26 (1976): 333–7.
—— “The Priestly ‘Picture of Dorian Gray.’ ” RB 83 (1976): 390–9.
—— “The Offering of Incense in Second Temple Times.” In Sefer Ben-Zion Luria:
Meh.qarim ba-Miqra uve-Toledot Yisrael, 330–4. Jerusalem: Kiryat Sefer, 1979.
—— Numbers = [Ba-midbar]: The Traditional Hebrew Text with the New JPS Trans-
lation. Philadelphia: JPS, 1990.
11:50:13:06:14 Page 219

Page 220
220 Bibliography
Milgrom, Jacob. “Two Biblical Hebrew Priestly Terms: S̆eqes. and T.āmē.” Ma‘arav 8
(1992): 107–16.
—— “Does H Advocate the Centralization of Worship?” JSOT 88 (2000): 59–76.
—— Leviticus: A New Translation with Introduction and Commentary. 3 vols. AB 3,
3A, 3B. New York: Doubleday, 1991–2001.
Mishra, Anand. “On the Possibilities of a Paninian Paradigm for a Rule-Based
Description of Rituals.” In Michaels and Mishra 2010: 85–97.
Moyer, James C. “Hittite and Israelite Cultic Practices: A Selected
Comparison.” In Scripture in Context II: More Essays on the Comparative Method,
edited by W.H. Hallo, James C. Moyer, and Leo G. Perdue, 19–38. Winona Lake:
Eisenbrauns, 1983.
Murphy, James G. A Critical and Exegetical Commentary on the Book of Leviticus.
Andover: Draper, 1874.
Naiden, F.S. Smoke Signals for the Gods: Ancient Greek Sacrifice from the Archaic
through Roman Periods. New York: Oxford University Press, 2013.
Narasimhachar, S., ed. Āpastamba Śrauta sūtra with the bhās.ya of Dhūrtaswāmi and
the vr.tti of Rāmāgnicit. Mysore: Oriental Library Publications, 1944–60.
Neusner, Jacob, trans. The Tosefta. 6 volumes. New York: Ktav, 1977–86.
Nihan, Christophe. From Priestly Torah to Pentateuch: A Study in the Composition of
the Book of Leviticus. FAT/II 25. Tübingen: Mohr Siebeck, 2007.
—— “Israel’s Festival Calendars in Lev 23 and Num 28–29.” In Römer 2008: 177–232.
No am, V . ered. Megilat Ta anit: Ha-nusah.im, pishram, toldotehem: Be-tseruf
mahadurah bik.ortit. Jerusalem: Yad Yitsh.ak. Ben-Tsevi, 2003.
Noss, Philip A., and Kenneth J. Thomas. A Handbook on Ezra and Nehemiah. New
York: United Bible Societies, 2005.
Noth, Martin. Leviticus: A Commentary. Translated by J.E. Anderson. Philadelphia:
SCM, 1965.
—— Numbers: A Commentary. Translated by James D. Martin. Philadelphia,
Westminster Press, 1968.
—— The Chronicler’s History. Translated by H.G.M. Williamson. JSOTSup 50.
Sheffield: Sheffield Academic Press, 1987.
Olyan, Saul M. “What Do Shaving Rites Accomplish and What Do They Signal in
Biblical Ritual Contexts?” JBL 117 (1998): 611–22.
—— Rites and Rank: Hierarchy in Biblical Representations of Cult. Princeton:
Princeton University Press, 2000.
Paran, Meir. “Two Types of ‘Laying Hands Upon’ in the Priestly Source” in Beer Sheva
2 (1985): 115–19.
—— Forms of the Priestly Style in the Pentateuch: Patterns, Linguistic Usages, Syntactic
Structures. Jerusalem: Magnes, 1989.
Pardee, Dennis. Les textes rituels. 2 vols. Ras Shamra-Ougarit 12. Paris: Editions
Recherche sur les civilisations, 2000.
—— Ritual and Cult at Ugarit. Atlanta: SBL, 2002.
Parker, Robert. “What are the Sacred Laws?” In The Law and the Courts in Ancient
Greece, edited by E.M. Harris and J. Rubenstein, 57–70. London: Bristol Classical
Press, 2004.
Pascal, Blaise. Pensées. Paris: Le Livre de poche, 1972.
11:50:13:06:14 Page 220

Page 221
Bibliography 221
Patton, Kimberley Christine. Religion of the God: Ritual, Paradox, and Reflexivity.
Oxford and New York: Oxford University Press, 2009.
Payne, Richard K. “The Shingon Ajikan: Diagrammatic Analysis of Ritual Syntax.”
Religion 29 (1999): 29–215.
Paz, Octavio. Claude Levi-Strauss: An Introduction. Translated by J.S. Bernstein and
M. Bernstein. London: J. Cape, 1970.
Penner, Hans H. “Language, Ritual and Meaning.” Numen 32 (1985): 1–16.
Péter-Contesse, René. “RW$ et RP: Note de lexicographie hébraïque.” VT 25 (1985):
486–96.
—— “Quels animaux Israël offrait-il en sacrifice?” In Schenker 1992: 67–77.
Petropoulou, Maria-Zoe. Animal Sacrifice in Ancient Greek Religion, Judaism, and
Christianity, 100 – 200. Oxford: University Press, 2008.
Pillai, K. Raghavan. The Vākyapadı̄ya: Critical Text of Cantos I and II. Studies in the
Vākyapadı̄ya 1. Delhi: Motilal Banarsidass, 1971.
Pinker, Steven. The Language Instinct. New York: W. Morrow and Co., 1994.
Popper, Karl R. Conjectures and Refutations: The Growth of Scientific Knowledge.
London: Routledge and K. Paul, 1969.
Porten, Bezalel and Yardeni, Ada. Textbook of Aramaic Documents from Ancient Egypt.
Vol. 1, Letters. Jerusalem: The Hebrew University, 1986.
Propp, William H. C. Exodus 19–40: A New Translation with Introduction and
Commentary. Anchor Bible 2A. New York: Doubleday, 2006.
Puhvel, Jaan. Hittite Etymological Dictionary. Vols. 1–7. Berlin: Mouton, 1984–2011.
Puri, Baij Nath. India in the Time of Patanjali. Bombay: Bharatiya Vidya, 1957.
Putnam, H. “Comment on Chomsky’s Reply.” In Block 1981: 359–60.
—— “The ‘Innateness Hypothesis’ and Explanatory Models in Linguistics.” In Block
1981: 292–9.
—— “What Is Innate and Why.” In Block 1981: 339–48.
Qimron, Elisha. The Temple Scroll: A Critical Edition with Extensive Reconstructions.
Beer Sheva: Ben Gurion University of the Negev Press, 1996.
—— and John Strugnell. Miqs.at Ma‘aśe ha-Torah. Discoveries in the Judaean Desert
10. Oxford: Clarendon, 1994.
Quack, Johannes. “Bell, Bourdieu, and Wittgenstein on Ritual Sense.” In Sax et al.
2010: 169–88.
Radcliffe-Brown, A.R. Structure and Function in Primitive Society. London: Cohen and
West, 1952.
Rainey, Anson F. “The Order of Sacrifices in Old Testament Ritual Texts.” Biblica 51
(1970): 485–98.
Rashi, [Solomon ben Isaac]. Perushei Rashi al ha-Torah. Edited by Charles B. Chavel.
Jerusalem: Mosad ha-Rav Kuk, 1982.
Regev, Eyal. “The Controversy Regarding the Eating of the Cereal Offering of the
Well-Being Sacrifices and the Thanksgiving Cakes in 4QMMT, the Temple Scroll,
and the Scholion to Megillat Ta anit.” Tarbiz. 65 (1996): 375–88 [in Hebrew].
Rendsburg, Gary. “Late Biblical Hebrew and the Date of ‘P’.” JANES 12 (1980):
65–80.
—— “The Two Screens: On Mary Douglas’s Proposal for a Literary Structure to the
Book of Leviticus.” JSQ 15 (2008): 175–89.
11:50:13:06:14 Page 221

Page 222
222 Bibliography
Rendtorff, Rolf. Studien zur Geschichte des Opfers im alten Israel. WMANT 24.
Neukirchen–Vluyn: Neukirchener, 1967.
—— Leviticus. BKAT III.1–III.4. Neukirchen-Vluyn: Neukirchener, 1985.
—— “Two Kinds of P? Some Reflections on the Occasion of the Publication of Jacob
Milgrom’s Commentary on Leviticus 1–16.” JSOT 60 (1993): 75–81.
—— and Robert A. Kugler, eds. The Book of Leviticus: Composition and Reception.
Supplements to Vetus Testamentum 93. Formation and Interpretation of Old
Testament Literature 3. Leiden and Boston: Brill, 2003.
Renou, Louis. “Les Connexions entre le rituel et la grammaire en sanskrit.” In Staal
1972: 432–69; originally published in JA 233 (1942): 105–65.
Reventlow, Henning Graf. “Rituelle Überlieferungen in Lev 1–4 als kanaanäisches
Erbe.” In Kämmerer 2007: 291–302.
Roberts, Mark D. Ezra, Nehemiah, Esther. The Communicator’s Commentary Series.
Old Testament 11. Dallas, TX: Word Books, 1993.
Rogerson, John W. Old Testament Criticism in the Nineteenth Century: England and
Germany. London: First Fortress Press, 1985.
Römer, Thomas, ed. The Books of Leviticus and Numbers. BETL 215. Leuven: Peeters,
2008.
—— “De la périphérie au centre: Les livres du Lévitique et des Nombres dans le débat
actuel sur le Pentateuque.” In Römer 2008: 3–34.
Rütersworden, Udo. Deuteronomium. BKAT V/3.1. Neukirchen–Vluyn: Neukirchener,
2011.
Ryle, Herbert Edward. The Books of Ezra and Nehemiah. Cambridge: Cambridge
Saraogi, Olga Serbaeva. “When to Kill Means to Liberate: Two Types of Ritual Action-
ism in Vidyāpı̄t.ha Texts.” In Michaels and Mishra 2010: 65–84.
Sarup, Lakshman. The Nighan.t.u and the Nirukta: The Oldest Indian Treatise on
Etymology, Philology, and Semantics. Delhi: Motilal Banarsidas, 1967.
Sax, William, Johannes Quack, and Jan Weinhold. The Problem of Ritual Efficacy.
Oxford Ritual Studies. Oxford: Oxford University Press, 2010.
Scheid, John. An Introduction to Roman Religion. Translated by Janet Lloyd. Blooming-
ton: Indiana University Press, 2003.
Schenker, Adrian, ed. Studien zu Opfer und Kult im Alten Testament. FAT 3. Tübingen:
J.C.B. Mohr, 1992.
Schiffman, Lawrence H. “ ôlâ and h.at.t.ā t in the Temple Scroll.” In Wright et al. 1995:
39–48.
—— “Sacrificial Halakhah in the Fragments of the Aramaic Levi Document from
Qumran, the Cairo Genizah, and Mt. Athos Monastery.” In Reworking the Bible:
Apocryphal and Related Texts at Qumran, edited by E. G. Chazon, D. Dimant, and
R.A. Clements, 177–202. Studies on the Texts of the Desert of Judah 58. Leiden:
Brill, 2005.
Schmid, Konrad. “Has European Scholarship Abandoned the Documentary
Hypothesis? Some Reminders on Its History and Remarks on Its Current Status.”
In Dozeman 2011: 17–30.
Schwartz, Baruch J. “The Bearing of Sin in the Priestly Literature.” In Wright et al.
1995: 3–21.
11:50:13:06:14 Page 222

Page 223
Bibliography 223
—— Torat ha-K.edusha: ‘Iyyunim bah.uk.k.a ha-kohanit she-ba-Torah. Jerusalem:
Magnes, 1999.
—— Leviticus. In The Jewish Study Bible, edited by A. Berlin and M.Z. Brettler,
203–80. New York: Oxford University Press, 2004.
—— “Does Recent Scholarship’s Critique of the Documentary Hypothesis Consti-
tute Grounds for Its Rejection?” In Dozeman et al. 2011: 3–16.
—— David P. Wright, Jeffrey Stackert, and Naphtali S. Meshel. Perspectives on Purity
and Purification in the Bible. LHBOTS 474. New York: T. & T. Clark, 2008.
Schwartz, Joshua. “The Temple Cult without the Sages: Prolegomena on the
Description of the Second Temple Period Cult according to Sources of the Second
Temple Period.” In New Studies on Jerusalem (h.idushim be-h.ek.er Yerushalayim)
edited by Eyal Baruch, Ayelet Levy-Reifer, and Avraham Faust, 14:7*–19*. Ramat-
Gan: Ingeborg Rennert Center for Jerusalem Studies (2008).
Schwemer, Daniel. “Das alttestamentliche Doppelritual ‘lwt wšlmym im Horizont
der hurritischen Opfertermini ambašši und keldi.” In Studies on the Civilization and
Culture of Nuzi and the Hurrians, edited by D.I. Owen and G. Wilhelm, 81–116.
Bethesda. Vol. 7, Edith Pofada Memorial Volume: CDL, 1995.
Scurlock, JoAnn. “Animal Sacrifice in Ancient Mesopotamian Religion.” In A History
of the Animal World in the Ancient Near East, edited by Billie Jean Collins, 389–403.
Handbook of Oriental Studies 64. Leiden: Brill, 2002.
—— “The Techniques of the Sacrifice of Animals in Ancient Israel and Ancient Meso-
potamia: New Insights through Comparison.” AUSS 44 (2006): 13–49, 241–64.
Searle, John R., Ferenc Kiefer, and Manfred Bierwisch, eds. Speech Act Theory and
Pragmatics. Synthese Language Library 10. Dordrecht and Boston: D. Reidel,
1980.
Seebass, Horst. Numeri. BKAT 4.1–4.3. Neukirchen–Vluyn: Neukirchener Verlag,
1993–2007.
Sen, Chitrabhanu. A Dictionary of the Vedic Rituals: Based on the Śrauta and Grhya
Sutras. Delhi: Concept, 1978.
Severi, C. “Language.” In Kreinath et al. 2006: 583–93.
Shachter, Jay F. The Commentary of Abraham Ibn Ezra on the Pentateuch. Vol. 3.
Leviticus Hoboken: Ktav, 1986.
Shulman, David. “Axial Grammar.” In Axial Civilizations and World History, edited by
Johann P. Arnason, S.N. Eisenstadt, and Björn Wittrock, 369–97. Jerusalem Studies
in Religion and Culture 4. Leiden: Brill, 2005.
Sifra (Weiss) Isaac H. Weiss, Sifra de-Ve Rav. New York: Om, 1947.
Śivadatta, M.P, Parab, K.P., and Śāstrı̄ , W.L. (eds.) The Patañjali Charita of Ramab-
hadra Dikhit. Kāvyamālā 51. 2nd edition. Bombay: Nirn.aya Sāgar, 1934.
Singer, Itamar. “The Hittites and the Bible Revisited.” In I Will Speak the Riddles of
Ancient Times: Archaeological and Historical Studies in Honor of Amihai Mazar,
edited by A.M. Maier and Pierre de Miroschedji, 723–56. Winona Lake:
Eisenbrauns, 2006.
Ska, Jean-Louis. Introduction to Reading the Pentateuch. Translated by Sr. Pascale
Dominique. Winona Lake: Eisenbrauns, 2006.
Smith, Brian K. Classifying the Universe: The Ancient Indian Varn.a System and the
Origins of Caste. New York: Oxford University Press, 1994.
11:50:13:06:14 Page 223

Page 224
224 Bibliography
Smith, Frederick M. The Vedic Sacrifice in Transition. Bhandarkar Oriental Series 22.
Poona: Bhandarkar Oriental Research Institute, 1987.
Smith, Jonathan Z. “The Bare Facts of Ritual.” In Imagining Religion: From Babylon to
Jonestown, 53–65. Chicago Studies in the History of Judaism. Chicago: University
of Chicago Press, 1982.
—— “The Domestication of Sacrifice.” In Hamerton-Kelly 1987: 191–205.
—— To Take Place: Toward Theory in Ritual. Chicago Studies in the History of
Judaism. Chicago: University of Chicago Press, 1987.
Smith, W. Robertson. Lectures on the Religion of the Semites. London: Adam and
Charles Black, 2nd edn. 1894.
Snaith, Norman H. Leviticus and Numbers. London: Nelson, 1967.
—— “The Sprinkling of Blood.” The Expository Times 82 (1970): 23–4.
—— “A Note on Ezra viii.35.” JTS  22 (1971): 150–2.
Sommer, Ferdinand, and Hans Ehelolf. Das hethitische Ritual des Pāpanikri von
Komana. Leipzig: J.C. Hinrichs, 1924.
Sperber, Dan. Rethinking Symbolism. Translated by A.L. Morton. Cambridge:
Cambridge University Press, 1975.
—— “Pourquoi les animaux parfaits, les hybrides et les monstres sont-ils bons à
penser symboliquement?” L’Homme 15 (1975): 5–34.
Staal, J.F. A Reader on the Sanskrit Grammarians. Cambridge, MA: MIT Press, 1972.
—— “The Meaninglessness of Ritual.” Numen 26 (1979): 2–22.
—— “Ritual Syntax.” In Sanskrit and Indian Studies: Essays in Honour of Daniel
H. H. Ingalls, edited by M. Nagatomi et al., 119–42. Studies in Classical India 2.
Dordrecht: D. Reidel, 1980.
—— ed. AGNI, The Vedic Ritual of the Fire Altar. 2 vols. Berkeley: Asian Humanities
Press, 1983.
—— Rules without Meaning: Ritual, Mantras, and the Human Sciences. Toronto
Studies in Religion 4. New York: Peter Lang, 1989.
—— “Artificial Languages across Sciences and Civilizations.” JIPh 34 (2006):
87–139.
Stackert, Jeffrey. Rewriting the Torah: Literary Revision in Deuteronomy and the
Holiness Legislation. Tübingen: Mohr Siebeck, 2007.
Staubli, Thomas. Die Bücher Levitikus, Numeri. NSK Altes Testament 3. Stuttgart:
Katholisches Bibelwerk, 1996.
—— “Hühneropfern im Alten Israel: Zum Verständnis von Lev 1,14 im Kontext der
antiken Kulturgeschichte.” In Römer 2008: 355–69.
Stroumsa, Guy G. The End of Sacrifice: Religious Transformations in Late Antiquity.
Translated by Susan Emanuel. Chicago: University of Chicago Press, 2009.
Swiggers, Pierre, and Alfons Wouters, “Content and Context in (Translating) Ancient
Grammar.” In Ancient Grammar: Content and Context, edited by Pierre Swiggers
and Alfons Wouters, 123–61. Leuven: Peeters, 1996.
Tambiah, S.J. “The Magical Power of Words.” Man 32 (1968): 175–208.
—— “A Performative Approach to Ritual.” Proceedings of the British Academy 65
(1979): 113–79.
Thite, G.U. Āpastamba-Śrauta-Sūtra: Sanskrit Text with English Translation and
Notes. 2 Vols. Delhi: New Bharatiya, 2004.
11:50:13:06:14 Page 224

Page 225
Bibliography 225
Thomas, Günter. “Communication.” In Kreinath et al. 2006: 429–70.
Tisserant, E. “Fragments syriaques du Livre des Jubilés.” RB 30 (1921): 55–86,
206–32.
Turner, Victor. The Forest of Symbols: Aspects of Ndembu Ritual. Ithaca: Cornell
—— The Ritual Process: Structure and Anti-Structure. Ithaca: Cornell University Press,
1969.
Urie, D. “Sacrifice among the West Semites.” PEQ 81 (1949): 75–7.
van Bekkum, Wout et al. The Emergence of Semantics in Four Linguistic Traditions.
Studies in the History of the Language Sciences 82. Amsterdam: John Benjamins,
1997.
Vanderkam, James C., ed. and trans. The Book of Jubilees: A Critical Text. Lovanii: E.
Peeters, 1989.
Varro, On the Latin Language. With an English Translation by Roland G. Kent. Vol. 1,
Books V–VII. Loeb Classical Library. London: Heinemann, 1938.
Vom Orde, Klaus. Die Bücher Esra und Nehemia. Wuppertaler Studienbibel.
Wuppertal: R. Brockhaus, 1997.
Vos, Howard F. Ezra, Nehemiah, and Esther. Bible Study Commentaries. Grand
Rapids: Zondervan, 1987.
Waschke, E.-J. “H$oe .” In TDOT, 14.46–49.
Watts, James W. “The Rhetoric of Ritual Instruction in Leviticus 1–7.” In The Book of
Leviticus: Composition and Reception, edited by Rolf Rendtorff and Robert Kugler,
79–100. Leiden: Brill, 2003.
—— Ritual and Rhetoric in Leviticus: From Sacrifice to Scripture. Cambridge:
Cambridge University Press, 2007.
Weinfeld, Moshe. Deuteronomy and the Deuteronomic School. Oxford: Clarendon,
1972.
—— “Social and Cultic Institutions in the Priestly Source against their Ancient Near
Eastern Background.” In Proceedings of the Eighth World Congress of Jewish Studies.
Jerusalem: Magnes, 1983: 95–138.
Wellhausen, Julius. Prolegomena to the History of Ancient Israel. Translated by J.S. Black
and A. Menzies. Gloucester: Peter Smith, 1973.
Wenham, Gordon J. The Book of Leviticus. NICOT 3. Grand Rapids: Eerdmans, 1979
—— Numbers: An Introduction and Commentary. Leicester and Downers Grove, IL:
Inter-Varsity Press, 1981.
Werman. “Appointed Times of Atonement in the Temple Scroll.” Meghillot, 4
(2006): 89–119 [in Hebrew].
Wevers, John William. Notes on the Greek Text of Deuteronomy. SCS 39. Atlanta:
Scholars, 1995.
—— Notes on the Greek Text of Leviticus. SCS 44. Atlanta: Scholars, 1997.
Wilke, Annette. “Basic Categories of a Syntactical Approach to Rituals: Arguments for
a ‘Unitary Ritual View’ and the Paraśurāma-Kalpasū¯tra as ‘Test–Case.’ In Michaels
and Mishra 2010: 215–62.
Williamson, H.G.M. Ezra, Nehemiah. WBC 16. Waco: Word Books, 1985.
Wright, David P. “The Gesture of Hand-Placement in the Hebrew Bible and in Hittite
Literature.” JAOS 106 (1986): 433–46.
11:50:13:06:14 Page 225

Page 226
226 Bibliography
Wright, David P. The Disposal of Impurity: Elimination Rites in the Bible and in Hittite
and Mesopotamian Literature. Atlanta: Scholars Press, 1987.
—— David Noel Freedman, and Avi Hurvitz, eds. Pomegranates and Golden Bells:
Studies in Biblical, Jewish and Near Eastern Ritual, Law and Literature in Honor of
Jacob Milgrom. Winona Lake: Eisenbrauns, 1995.
Yadin, Yigael, ed. The Temple Scroll. 2 vols. Jerusalem: Israel Exploration Society, 1983.
Zenger, Erich, and Christian Frevel. “Die Bücher Levitikus und Numeri als Teile der
Pentateuchkomposition.” In Römer 2008: 35–74.
Zer-Kavod, Mordekhai. Ezra u-Neh.emyah: Meforashim bi-yede Mordekhai Zer-Kavod.
Jerusalem: Mosad ha-Rav K.uk. 1980 [in Hebrew].
Zevit, Ziony. “Converging Lines of Evidence Bearing on the Date of P.” ZAW 94
(1982): 481–511.
—— The Religions of Ancient Israel: A Synthesis of Parallactic Approaches. New York:
Continuum, 2001.
Zimmerli, Walther. Ezekiel: A Commentary on the Book of the Prophet Ezekiel.
Translated by Ronald E. Clements. 2 vols. Philadelphia: Fortress Press, 1979–83.
Zobel, H.-J. “ZE
a.” In TDOT, 10.577–83.
11:50:13:06:14 Page 226

Page 1
A “Grammar” of Σ
This  is intended as a concise presentation of the Israelite sacri-

ficial system (Σ), integrating the findings of the present study with the
results of previous research. It has been abstracted primarily from the
Priestly texts in the Pentateuch (P); post-biblical data are referred to only
where they elucidate the rules abstracted from P. However, it is designed to
account for every licit (that is, grammatical) sacrificial sequence in Σ and
to exclude every illicit (ungrammatical) sequence in it.
In order to invalidate this  it should suffice, in principle, to
offer a single example of a sacrificial sequence in an Israelite text (or in
historical practice in antiquity, such as pictorial or other archaeological
findings) that is here considered ungrammatical. As in natural languages,
grammatical rules should allow for exceptions, but for a grammar to have
a strong explanatory power its rules must be formulated in a manner
that minimizes such exceptions. On the other hand, many sequences
that are considered grammatical here are not (and need not be) attested
in any text. This in itself does not invalidate the , but only
suggests that it is too permissive and that further restrictions might
be included in it. This grammar will be validated if new findings reveal
previously unknown sacrificial sequences that are herein considered
grammatical.
For the sake of concision, the sigla defined in the section “Sigla Used for
Formal Analysis” (see pp. 000–000) are used throughout the .
ZOEMICS
1. Sacrificial animals are categorizable according to three binary-

branching criteria: zoological class, sex, and age (§ 2.2.1).
(a) Zoological class is consistently binary; the total of five species is
arrived at by bifurcation of so-called “flock animals”, as in the
diagram below.
(b) Sex is binary (due to biological restrictions).
(c) Within P, age is binary.
11:50:13:06:14 Page 1
Page 2
2 A “GRAMMAR” OF R
....................................................................................
2. The intersection of these criteria results in twenty blocks, the small

shaded squares in the following diagram:
3. A zoeme is a combination of one or more of these twenty blocks;

hence there are theoretically 220 zoemes.
4. A potential (=grammatical) zoeme is a zoeme that does not violate
any of the following restrictions:
(a) Zoemic combinations within a zoological species may consist of
one, two, or four blocks (but not three).1
(i) A zoeme of two blocks will have contiguous blocks––that is,

blocks that share a border in the zoemic map.2
1
A possible exception is created artificially by the distinction between S웧 on the one
hand (Num 15:4–7) and the three other members of ! (along with all other members of !,
Num 15:11b) on the other hand (see p. 000 and n. 000). See, however, rule (12) in Jugation.
2
E.g., a zoeme consisting of {g웧, G웨} is ungrammatical; a zoeme consisting of {g웧, g웨}
is grammatical (and is named: g); a zoeme consisting of {g웧, G웧} is grammatical (and is
named: 웧). Note that the blocks of f웧 (for the paschal offering) are contiguous on the
three-dimensional diagram in Fig. 4.
11:50:13:06:14 Page 2
Page 3
A “GRAMMAR” OF R 3
....................................................................................
(b) Among quadrupeds, zoemes comprising blocks from more than

one zoological class never consist of only the immature of one
class and only the mature of another.3
(c) Zoemes comprising blocks from more than one zoological class
will never consist of only the male of one class and only the
female of another.4
3
E.g., a zoeme consisting of {s, G} would be ungrammatical.
4
E.g., a zoeme consisting of {웧, 웨} or of {s웧, 웨} would be ungrammatical.
11:50:13:06:14 Page 3
Page 4
....................................................................................
5. An attested zoeme is a grammatical zoeme that is attested in a text;

only twenty to thirty zoemes are attested in P (see Table 1). The
following observations pertain to Σ:
(a) Within no distinction is made between 웨 and 웧 p. 000.
(b) consists of two zoemes only.
(c) There is a tendency towards symmetry in zoemes comprising
blocks from more than one zoological class: the number of
blocks from each class is usually the same.5
(d) No attested zoemes consists exclusively of mature female
specimens.6
JUGATION
1. The following diagram maps the jugational pattern of a single zoeme

(without salt):
2. In the zoemic system, jx is to be read “(the) jugate j on jugational Level

x”:7
(a) j must be one of the materials listed in J1 (in P); x indicates levels
A, B, C, and so on.
3. With the jugate salt, the full jugational map (stemming from a
zoeme) is as follows:
5
E.g., the zoeme consisting of {s웧, g웧} constitutes the zoeme f웧 (for the paschal ritual,
Exod 12:5) (see p. 000). There are, however, exceptions to this rule. E.g., Josephus (Ant. Jud.
3.9.11–12) considers the zoeme {s웧, g웧, 웧} grammatical; see p. 000.
6
The “red heifer” is not a sacrificial offering and therefore not a zoeme strictly speaking
(nn. 000, 000).
7
E.g., oilB is to be read: “the jugate oil on jugational Level B”, or in short “B-level oil.”
11:50:13:06:14 Page 4
Page 5
....................................................................................
4. A zoeme is never found to be directly or indirectly subordinate to

another zoeme
(a) But it may be subordinate to non-animal materials (see § 3.4.3),
in which case one finds:
(b) In these cases the jugates proceed from zoemeB as in rules 1)

and 3).8
5. The jugational map is referred to as a “tree” with “branches” and
“nodes.”
(a) Every word in this map is a “node.”
(b) A chain of two or more nodes (proceeding rightwards on the
map) is a “branch.”
(c) The relation between nodes can be expressed in terms of a family
tree, for example: wineB in 3) is a “son” of zoemeA and an “uncle”
of oilC.
(d) Each node in the tree represents a potential jugate: where a node
appears in the chart, a jugate has the potential to be expressed
praxemically.
(i) In praxemic sequences, every node may be expressed or
unexpressed, that is in a potential jugational pattern it is un-
necessary for every jugate to be expressed praxemically,9 but:
(ii) If a node is unexpressed, then its descendants are excluded
with it from the jugational tree.
8
TS apparently refers to [wineA, zoemeB, wineC] on a single branch (19:11–20:8); and to
[oilA, zoemeB, grainC, oilD] on a single branch (21:12–22:8); see n. 000.
9
In other words, beginning with the initial node and proceeding to the right, one can
stop at any node along the branch; the terminal nodes need not be all on the same jugational
level.
11:50:13:06:14 Page 5
Page 6
....................................................................................
6. Grammatical sequences can be formed by beginning at any node on

this tree and following the branches of the tree (to the right).
(a) The initial node is considered to be placed on Level A. For
example, starting from grain (p. 000), the jugational tree would
be:
7. Jugates of the same material may appear more than once in a single
jugational tree.
(a) E.g., grainA may have a “brother,” another grainA; or a “nephew,”
which would then be grainB.10
(b) A jugate may even reappear on the same branch, that is, as a
direct descendent––a son, a grandson, a great-grandson, and so
on––of another jugate of the same material.
(i) For example in the cases described in rule 4)a) one finds:
grainA zoemeB grainC
(ii) However, with the possible exception of salt, jugates of the
same material are never found to be adjacent on the same
branch, for example,
grainA grainB
8. Since there appears to be no limit on the size of jugational patterns,
there are potentially an infinite number of finite jugational patterns.
Moreover,
(a) The jugational tree is potentially unlimited on the vertical axis.11
(b) If salt can be an immediate subordinate jugate to another salt,
then an infinite chain of subordinate salts is formed (§ 3.9.2).
9. Nothing but salt may be subordinate to salt.
10
See Exod 29:2 and Num 6:15a.
11
This is primarily due to a lack of limitation on A-level cojugates. For example,
Solomon’s hecatombs (1 Kgs 8:63) could be conceived as a large wellbeing+1 offering, com-
prising 22,000+120,000 (not including the wholeburnt offerings), all of which are co-
jugates.
11:50:13:06:14 Page 6
Page 7
....................................................................................
Attested Diachronic Developments in Jugation
10. The following diachronic developments are attested:

(a) Either coordinate or subordinate jugation of grain to zoeme
(grainB and grainA cannot be “uncle and nephew”)
Both coordinate and subordinate jugation of grain to zoeme
(grainB and grainA may be “uncle and nephew”)
(b) Only wholeburnt offerings entail subordinate jugation
Wholeburnt and wellbeing offerings entail subordinate jugation
(c) Only calendric offerings entail subordinate jugation
Calendric and non-calendric offerings entail subordinate
jugation
(d) Wholeburnt and wellbeing offerings entail subordinate juga-
tion
Wholeburnt, wellbeing, and calendric purification offerings
entail subordinate jugation
(e) Wholeburnt and wellbeing offerings entail subordinate juga-
tion
Wholeburnt, wellbeing and some non-calendric purification
offerings entail subordinate jugation
The Interface between Jugation and Zoemics
11. The relation between animals and their jugates is often explicit in
the texts.12
12. According to these texts, jugation is determined by the age, sex, and
species of the animal––regardless of whether it is being offered in
fulfillment of a requirement to offer one zoeme or another.13
12
See Num 15; ADL 9.
For example, even if one offers a ram as an (e.g., as a volitional wellbeing offering) or
13
as an 웧 (e.g., as a volitional wholeburnt offering) and not as an S웧 (e.g., on a calendric

occasion where a ram, S웧, is actually required) one must still offer with it 0.2 ephah of
grain–1 and not 0.1 ephah as for a male lamb (s웧), a female lamb (s웨), or a mature ewe (S웨)
offered as an .
11:50:13:06:14 Page 7
Page 8
....................................................................................
The Interface between Jugation and Praxemics
13. From a praxemic point of view, the link between a zoeme and its
subordinate grain offering is stronger than the link between a zoeme
and its subordinate libation.
(a) Therefore, from a diachronic point of view it is more difficult to
sever (p. 000).
14. Jugates may undergo praxemic transformations depending on their
location in the jugational tree.
(a) E.g., in P grainx is turned into smoke on the altar (p. 000 and
n. 000).
(i) in its entirety if x = B, C, D, and so on.
(ii) partially if x = A (and no zoemeB is present)
(iii) not at all if x = A and a zoemeB is present
(b) According to rabbinic tradition,14 oilx is turned to smoke
(i) in its entirety if x = C, D, and so on (i.e., where oil is sub-
ordinate to a cereal offering that is in turn subordinate to a
zoeme)
(ii) partially, if x = A, or if x = B (subordinate to grainA).
(c) According to rabbinic tradition (n. 000), winex
(i) is poured on the base of the altar if x = B
(ii) is spattered on the upper surface of the altar if x = A.15
15. Jugates may undergo further jugational transformations depending
on their location in the jugational tree (§ 3.8).
(a) There are some instances where if jugate j is found on juga-
tional Level A, it will entail subordinate jugate kB; if jugate j is
located on jugational Level B, subordinate jugate kC will be absent.
(i) For example in P, it appears that a raw freewill grainA
offering entails frankincenseB but a raw freewill grainB does
not entail frankincenseC.
(ii) In rabbinic tradition, any freewill independent grainA
offering entails frankincenseB but grainB does not entail
frankincenseC.
(iii) Similarly, according to late rabbinic tradition, wineA entails
saltB but wineB does not entail saltC.
14
See HMQ 16:14.
15
Some authorities opine that since this cannot be done due to external restrictions,
wineA, too, is poured on the base of the altar (see n. 000).
11:50:13:06:14 Page 8
Page 9
....................................................................................
HIERARCHICS
1. zp+1 = [(xp)a , (xp)b ,. . .,(xp)m , (yp)a , (yp)b ,. . .,(yp)n] is grammatical

Read: A sacrificial complex of type z on hierarchic Level p+1 comprising m
offerings of type x and n offerings of type y, each on hierarchic Level p, is
grammatical (p. 000).
Since the number of lower-level sacrificial types need not be limited to
two, this rule can be formulated more generally as follows:
2. tx+1 = [(j1)xt , (j2)xt , . . . (jn)xt ]
1 2 m
is grammatical
x indicates the hierarchic level; j1, j2, . . ., jn are an n number of jugates; t1,
t2 . . . tm are an m number of sacrificial types
Read: A sacrificial complex (by the name of t) on hierarchic Level x+1 con-
sisting of n jugates offered praxemically as m distinct sacrificial types on
hierarchic Level x is grammatical.
The Interface between Hierarchics and Praxemics
3. Where one or more Level 0 offerings comprise a Level +1 offering,

their praxemics remain unaltered by the fact that they have entered
this hierarchic relation.16
PRAXEMICS
1. A praxeme consists of up to six components: (1) atomact, (2) agent,

(3) object, (4) target, (5) location, and (6) time-frame (§ 5.5).
(a) A praxeme comprises a minimum of two of these components,
atomact and agent.
(b) A praxeme is designated as an atomact with four qualifiers, in
the following form: a(g, o, t, l) where the a = atomact; g = agent;
o = object; t = target; and l = location.17
16
E.g., the blood of the /0purification and /0wholeburnt comprising a reparation+1 offering
in Lev 5:7 is manipulated in the same manner as that of any other /0purification or
/0wholeburnt that are not in such a hierarchic relation.
17
E.g., daub (high priest, blood, four horns of bronze altar, by/on bronze altar) designates a
praxeme in which the High Priest daubs blood on the four horns of the bronze altar while
standing by the altar or on it.
11:50:13:06:14 Page 9
Page 10
....................................................................................
(i) If one of these components is irrelevant, it is not

mentioned.18
2. The inventory of atomacts comprises (§ 5.5.1):
(a) Bringing forth
forth/in
in
(b) Hand leaning
(i) Single hand (WDY VMS)
(ii) Two hands (WYDY YT$ VMS)
(c) Immolation
(i) Slaughter (UX$o)
(ii) Pinching (QLM)
(d) Acquisition, or “taking
taking”
(i) Of liquids:
(1) By collecting (supposedly in a vessel)
(2) By dipping (finger) (XQL, LBU)
(ii) Of solids
(1) Extraction RYSH, OYRH
(2) Separation OYRH
(e) Dismemberment
(i) Tearing (bird, ESd $o)
(ii) Cutting apart (XTd N)
(f) Washing (JXR)
(g) Raising (FYNH)
(h) Placement
(i) Of liquids
(1) Tossing (QRZ)
(2) Pouring (VP$o)
(3) Squeezing/draining (HCM)
(4) Daubing (ITN)
dH)
(5) Flicking (HZ
(ii) Of solids
(1) Sprinkling
(2) Flinging (VYL$oH)
(3) Setting (VRE, O$ o , ITN)
18
E.g., slaughter (anonymous, zoeme, northern flank of altar) designates a praxeme in
which a zoeme is slaughtered on the northern flank of the altar ( not mentioned).
11:50:13:06:14 Page 10
Page 11
....................................................................................
(i) Cooking
(i) Boiling (L$oB)
(ii) Roasting
(j) Ingestion (LKA)
(k) Incineration (FR$ o , RYUQH)
3. There are internal restrictions on atomacts:
(a) Atomacts designated by Roman numerals are mutually exclusive
if subsumed under the same lower-case letter.19
(b) Atomacts designated by Arabic numerals are not mutually
exclusive (even if subsumed under the same Roman numeral).20
4. Each praxeme must be performed by one of three possible agents
(§ 5.5.2).
(a) Offerer
(b) Officiant
(c) Unspecified party (=“Anonymous”)
5. The offerer may be:
(a) A single offerer
(i) A lay Israelite
(ii) A chieftain (AY$o N)
(iii) A priest
(1) Ordinary priest
(2) The High Priest (XY$oMH IHKH)
(b) Joint offerers
(c) The community (HDEH)
6. The officiant may be:
(a) An ordinary priest
(i) Single priest (IHKH)
(ii) Several priests ([IRHA YNB] OYNHKH)
(b) The High Priest (XY$oMH IHKH)
7. Some particular rules pertain to agents:
(a) The offerer is fixed for each zoeme. That is, all praxemes pertaining
to a single zoeme pertain to the same offerer, but not vice versa.
19
E.g., (b)(i) precludes (b)(ii) and vice versa; (c)(i) precludes (c)(ii) and vice versa, for a
single .
20
Example: Zoemic blood may be taken first in a vessel (d)(i)(1) and then on a finger
(d)(i)(2). This may be true for (h)(ii)(1)–(3) as well, but no examples are documented.
11:50:13:06:14 Page 11
Page 12
....................................................................................
(b) The officiant is not fixed for each zoeme. That is, diverse
praxemes pertaining to a single zoeme may be assigned to
different officiants.
(c) A High Priest may always substitute for an ordinary priest, but
not vice versa.
(d) If the offerer is a priest, some praxemic shifts may apply.
(i) grainA is burnt on the altar in its entirety in this case (Lev
6:16).21
(ii) Also, in this case, offerer = officiant is possible.
(e) Rarely, it is required that the offerer and the officiant be the same
person (this may occur when the person in question is the High
Priest).
8. The inventory of   includes:
(a) Zoeme (in toto)
(b) Blood (OD)
(c) Suet (collective) (! only, BLX, OYBLX)
(i) The suet that covers the entrails (BRQH TA HSKMH BLXH)
(ii) All of the suet that is around the entrails (LE R$oA BLXH
BRQH)
(iii) Both kidneys (TYLKH YT$o)
(iv) Suet around kidneys, on the sinews (IHYLE R$oA BLXH
OYLSKH LE R$oA)
(v) Caudate lobe (DBKH TRTY)
(vi) Broad tail (! only, HYLAH)
(d) Flesh and bones
(i) Right hind leg (! only, IYMYH QW$o)
(ii) Breast (! only, HZX)
(iii) Foreleg (! only, EWRZ)
(iv) Innards (! only, RDP)
(1) Entrails (! only, BRQ)
(a) Stomach (! only, HBQ)
(v) Head (any zoeme, $oAR)
(1) Lower jaws (! only, OYYXL)
(vi) Shins (! only, OYERK)
21
According to Lev 4:3–12; 9:11, the flesh of a purification offering is incinerated outside
the camp in its entirety if a High Priest is the offerer. Gane 2005:89 generalizes that a priest
may not partake of the flesh of his own purification or reparation offering.
11:50:13:06:14 Page 12
Page 13
....................................................................................
(vii) Crissum (with adjacent feathers, ! only, HTCNB WTARM)

(e) Hide (! only, RWE)
(f) Any agent may be an object manipulated
(i) but this is rarely attested (p. 000)
9. The inventory of  includes:
(a) The tent of meeting
(i) Inner sanctum ($oDQH, OY$oDQH $oDQ)
(1) The cover of the ark (TRPKH LE)
(2) Area located before the cover of the ark (TRPKH YNPL)
(ii) Outer shrine ($oDQH)
(1) The veil (TKRPH)
(2) Incense altar (TRUQH XBZM)
(a) Horns (TRUQH XBZM TNRQ)
(b) Upper surface (BHZH XBZM LE)
(3) Other areas in the shrine, or the tent in general (LHA
DEWM)
(b) The bronze altar and its vicinity (T$oXNH XBZM)
(i) Horns (XBZMH TNRQ)
(ii) Upper surface (XBZMH LE)
(iii) Walls
(1) Single wall (XBZMH RYQ)
(2) Several walls (XBZMH TWRYQ*)
(iv) Base (DWSY)
(v) By the altar (XBZMH LCA)
(c) Human bodies
(i) (Offerer’s) body in general (e.g., IRHA LE)
(1) Right earlobe (TYNMYH WNZA VWNT)
(2) Right thumb (TYNMYH WDY IHB)
(3) Big toe of right foot (TYNMYH WLGR IHB)
(ii) (Offerer’s) clothing in general (e.g., WYDGB LEW)
(d) The   in one praxeme may become the
 in the next.22
22
Only very few cases of 9.(d) are attested, e.g., Lev 9:20–21, where the breasts are the
 in v. 20, but the   in v. 21. Note that since every  may be
an   according to 8)f), every  may also be a .
11:50:13:06:14 Page 13
Page 14
....................................................................................
(i) But a comparable transformation cannot occur within a

single praxeme.23
10. The inventory of  in P includes:
(a) The sacred precincts
(i) The tent of meeting (DEWM LHA)
(1) The inner sanctum ($oDQH, OY$oDQH $oDQ)
(2) The outer shrine ($oDQH)
(ii) The outer court (DEWM LHA RCX)
(1) The bronze altar and its vicinity
(a) On the altar (XBZMH LE)
(b) By the altar (XBZMH LCA)
(c) The northern flank of the altar (XBZMH VRY LE
HNWPC)
(2) Temporary dumping grounds for ashes (I$oDH OWQM)
(3) (Presumably) every other area in the outer court
(RWHU OWQM within DEWM LHA RCX)
(b) The camp
(i) Within a house (DXA TYBB)
(ii) Unspecified location (=“Anywhere”)
(c) Outside the camp (HNXML JWXM)
(i) Dumping grounds for the ashes (I$DH VP$o)
(ii) Unspecified location (=“Anywhere”)
11. The inventories of  and  may be expanded in
an unpredictable manner according to the real or imaginary archi-
tecture of the sanctuary and altar where the sacrifice is performed.
12. Pairing denotes the creation of a logical link between a member of
one inventory and a member of another inventory, namely, their
coexistence within a single praxeme.24
13. The pairing of members of the five inventories (atomact, agent,
object manipulated, target, location) with one another is highly
restricted.
(a) Since there are five inventories, a (g,o,t,l), ten sets of pairs must
be considered: a+g, a+o, a+t, a+l, g+o, g+t, g+l, o+t, o+l, and t+l
23
Example: a(o=flesh,t=flesh) cannot occur (it is logically impossible).
E.g., in the praxeme pinch (priest,!) a member of the inventory  (priest) is
24
paired with a member of the inventory  (pinching). This pair is attested in P, since a
priest performs the pinching in Lev 1:15; 5:8.
11:50:13:06:14 Page 14
Page 15
....................................................................................
(i) In P, target and location are closely linked and are only
rarely considered separately.25
14. : +: any agent can perform any atomact, with
the following restrictions:
(a) hand-leaning must be performed by the offerer or by the
offerer’s representative.
(i) Atomact leaning two hands can be paired only with agent
high priest in P but in later traditions no such restriction is
found.
(ii) In LXX a restriction according to sacrificial type appears
to be active: For ! wellbeing both hands of the offerer are used,
for other sacrificial types, one (see n. 000).
(iii) Representatives may substitute the offerer if the offerer is
the community.
(b) For atomact bring forth and target priest or sanctuary the agent
must be offerer.
(c) Atomact raise may be paired only with agent officiating priest.26
(d) Atomact ingestion is paired only with agent priest (see
15(a)(iii)(8))
(i) The paschal sacrificial ritual is an exception.27
15. : + : any atomact may be paired
with any object manipulated.28
(a) The following restrictions may apply:
(i) Physical restrictions; for example it may be impossible to
perform atomacts that pertain to solid materials (for
example cutting) if the object is liquid.
(ii) External legal restrictions; for example the law prohibits
one to ingest blood, though this is physically possible.
25
In later traditions, this link is loosened, and the location and target may be some thirty
cubits apart (e.g., b Zeb 64a).
26
In Num 8:8, where Moses officiates, both Moses and Aaron raise the Levites.
27
Most of the apparent restrictions on the pair agent+atomact, e.g., the fact that atomacts
toss, pour, squeeze, daub, and flick are always paired with the agent priest, are not in fact
intrinsic restrictions on the pair agent+atomact. Rather, they arise in a roundabout manner
from the intrinsic restrictions on the pairs agent+target and agent+object manipulated, and
from the fact that these atomacts are attested only in relation to blood (and libations). The
operative restrictions in this case are probably that the target altar and the object blood can
be paired only with the agent priest.
28
E.g., the pairing of the atomact lean hands with object flesh; or of atomact toss
with object flesh is grammatical, though never required in P. Both are found in Tannaitic
traditions (m. Tam. 7:3).
11:50:13:06:14 Page 15
Page 16
....................................................................................
(iii) Internal grammatical restrictions

(1) For example, hand-leaning can be paired only with
certain zoemes
(a) With !
(i) hand-leaning (zoeme, court) is always required,
unless
1. type = reparation (hand-leaning is not
required for ! reparation, according to one
opinion) (see p. 000)
2. agent = woman or non-Israelite (Tannaitic
tradition; m. Men. 5:7; 9:6–7; regarding
public offerings, see n. 000)
(b) But not with !
(c) Nor with non-zoemic materials (in P).
(2) slaughtering can be paired only with certain zoemes:
(a) With !, where slaughter(zoeme) is always required
(i) Except for the 웧 dispatched to the wilder-
ness, which is not slaughtered (and is not itself
a sacrifice, strictly speaking).
(ii) slaughter (anonymous, zoeme, northern flank) is
required in the case of !wholeburnt/reparation/purification
(b) Not with !
(i) Where, instead, pinch (priest, zoeme) is
required
1. For !wholeburnt pinching is thorough (head
detached)
2. For !wholeburnt pinching is partial (head not
detached)
(3) Dismemberment is found only for ! wholeburnt.
(a) Sometimes for !purification (late tradition; see
n. 000)
(b) Its form differs from one zoeme to another:
(i) Detachment of head and tearing for !
(ii) Cutting apart for !
(4) The atomact wash is paired with some objects that are
to come into direct contact with the sancta.
11:50:13:06:14 Page 16
Page 17
....................................................................................
(a) In the case of wash (anonymous, entrails and

shins) and, outside P, in the case of wash (flesh of
!wholeburnt)
(b) In a case where object = agent (reflexive, priest
washes parts of body), before the agent approaches
the altar or the tent of meeting.
(5) The atomact raise can be paired with any object,
including live zoemes, materia sacra, and agents.
(a) After an object is raised, it is either incinerated
on the bronze altar or else it becomes a priestly
perquisite.
(b) After having been raised, an object will not be
incinerated outside the camp.
(6) Atomacts subsumed under placement are found pri-
marily in the context of the placement of materia sacra
on the sancta, and are treated under +
and +.
(7) Cooking is mentioned in sacrificial texts, but it is not
considered an atomact where it is optional and not
required.
(a) It often precedes ingestion where ingestion is itself
permitted but not required, for example in the
context of wellbeing offerings.
(b) There are, however, restrictions on its location.
(c) In one rare case, cook (foreleg) precedes raise (fore-
leg), and appears to be an atomact. It appears that
the offerer is to perform it (Num 6).
(d) The atomact roast is paired only with the flesh of
the paschal zoeme as a whole and is a prerequisite
for its ingestion.
(i) Though it is physically possible to cook or
roast suet, for example before placing it on the
altar, this is not attested in the Israelite system
and is probably ungrammatical.
(8) Ingestion is an atomact strictly speaking only in the
context of the eaten purification offering, and possibly
in the context of the reparation offering (and in
the paschal rite); elsewhere it is permitted but not
required.
11:50:13:06:14 Page 17
Page 18
....................................................................................
(a) There are, however, restrictions on its location

even where it is not an atomact.
16. From a diachronic point of view, there is a tendency to project––
(a) atomacts performed in relation to one category of materia sacra
onto other categories of materia sacra:
(i) toss (priest, blood ) gave rise to toss (priest, flesh) in later
traditions (m. Tam. 7:3)
(ii) wash (entrails and shins of !wholeburnt) may have given rise
to wash (flesh of !wholeburnt) (see n. 000).
(iii) hand-leaning (offerer, zoeme) engendered hand-leaning-
(priest, flesh) and hand-leaning (priest, suet) (m. Tam. 7:3)
(b) atomacts performed in the context of one sacrificial type onto
other sacrificial types:
(i) dismember (anonymous, zoeme) for !wholeburnt gave rise
to dismember (anonymous, zoeme) for !burnt purification (see
n. 000)
17. : +  : +:
(a) Since most praxemes do not contain the component ,
only a few atomacts are paired with a target. These atomacts
are:
(i) bringing forth: the acts designated in P by BYRiQ v iH and AYB
iaH
should probably be viewed as bring (agent, object, target)
rather than present (agent, object, location)
(1) The objects that are paired with this atomact are
(a) Zoeme (in toto), in which case the praxeme is
bring (offerer, zoeme, courtyard), or bring (offerer,
zoeme,t=priest)
(b) Parts of a zoeme (see 20.(a))
(c) Non-zoemic materials
(ii) Atomacts subsumed under placement
(1) With regard to the placement of solid materials, the
set of attested atomacts is relatively limited, but few
grammatical restrictions seem to apply.
(a) Atomact sprinkle is found only with object salt and
with target materia sacra in the praxeme sprinkle-
(priest, salt, materia sacra)
(b) Atomact fling is found in P only with object
crissum and feathers (of !wholeburnt) in the praxeme
11:50:13:06:14 Page 18
Page 19
....................................................................................
fling (priest, crissum and feathers, t=temporary

dumping grounds for ashes)
(c) Atomact set is found to be paired with the target
upper surface of bronze altar in the following
praxemes:
(i) set (priests, wood, upper surface)
(ii) set (priests, flesh, wood), where the flesh is that
of a !wholeburnt
(iii) set (priests, suet, upper surface)
(d) And with various targets in preparatory atomacts
for the atomact raise
(i) set (offerer, suet, t=breast), where the zoeme is
a wellbeing offering
(ii) set (priest, suet and non-zoemic materials, t=
[hands of]offerer) 29
(2) Where the placement is of liquid materials, the object
in P is either oil,30 or wine, or blood.
(a) The application of liquid oil is treated only in
Leviticus 14 (person purified from scale disease)
and in Exodus 29 || Leviticus 8 (consecration ritual).
(b) Elsewhere in P, the oil is treated as part of the
cereal offering.31
(c) P does not indicate the precise manner in which
wine is to be offered.
(i) But later texts distinguish between two atom-
acts and three targets (not all combinations
are documented)
1. flick (priest, upper surface)
2. pour (priest, t=zoeme)
3. pour (priest, upper surface)
4. pour (priest, base)
(d) The libation of water is not mentioned in P
29
E.g., Lev 8:27; Num 6:19.
30
The placement of oil on semolina and similar acts described in Lev 2 as part of the
preparation of the semolina offering probably take place outside the sacred precincts and are
not included here.
31
Later traditions refer to separate atomacts that pertain to voluntary offerings of oilA,
e.g., taking a handful (HCYMQ), followed by toss (priest, oil, upper surface). See HMQ 16:14.
11:50:13:06:14 Page 19
Page 20
....................................................................................
18.  

(a) Due to its centrality in the sacrificial system, a separate section
is dedicated to the application of blood. Five atomacts pertain
to the application of blood:
(i) flick
(ii) daub
(iii) squeeze
(iv) toss
(v) pour
(b) Where more than one of these atomacts occurs within the
offering of a single zoeme, the atomact listed first in rule (a)
above normally precedes the atomact listed later in this list.
(1) But daub may precede flick.
(c) Each atomact may occur more than once even within the
offering of a single zoeme.
(d) Therefore, there are, theoretically, an infinite number of
combinations.
(e) However, it appears that several restrictions apply:
(i) pour and squeeze are followed by no further blood-related
atomacts.
(ii) squeeze is limited to !.
(iii) pour and squeeze are mutually exclusive: within the
manipulation of the blood of a single zoeme, if one occurs,
the other cannot occur.
(iv) toss and squeeze are mutually exclusive.
(f) The adverb BYBS applies (in P) only to tossing (passim) and
daubing (Lev 8:15; 16:18), but this does not seem to be an
inherent grammatical restriction
(i) In fact, in TS one finds pour+BYBS
(1) And other atomacts may be qualified by BYBS to form
new grammatical combinations.
(g) It would appear that in the context of blood application, any of
the five atomacts listed in 18.(a) may combine with any target,
with the following restrictions:
(i) If the target is an altar, then
(1) flick+base is ungrammatical
(2) daub+base is ungrammatical
11:50:13:06:14 Page 20
Page 21
....................................................................................
(3) daub+wall is probably ungrammatical

(4) squeeze+horns is ungrammatical
(5) pour+horns is ungrammatical
(6) squeeze+golden altar is ungrammatical
(7) toss+horns is grammatical (but only due to
the semantic shift of QRZ, and
therefore one would expect to
find this only in late rabbinic
texts)
(ii) If the target is an agent, then
(1) daub+agent is grammatical (see Lev 14:14)
(2) flick+agent is grammatical (see Exod 29:21)
(3) toss+agent is probably grammatical
(4) squeeze+agent is probably ungrammatical
(5) pour+agent is probably grammatical
19. : +
(a) P is relatively silent concerning the precise location of
praxemes.
(i) In praxemes that contain a target, it appears that any spot
from which it is physically possible to reach the target is an
acceptable location.
(1) P does not even seem to be concerned about whether
the priest ascends the altar or not in order to perform
the blood applications.
(2) In rabbinic literature the spatial relation between
 (the spot where the priest stands) and
 becomes more complex, and is often explicit in
the texts.
(ii) In praxemes that do not have targets (the vast majority of
praxemes), some restrictions are found concerning location
(1) Immolation of ! (according to one opinion, !),
as sacrificial types wholeburnt, purification and
reparation must take place on the northern flank of
the bronze altar.32
32
According to Tannaitic tradition, this restriction pertains also to public wellbeing
offerings (m. Zeb. 5:1,5), but there is no indication of this in P.
11:50:13:06:14 Page 21
Page 22
....................................................................................
20. : +

(a) There appears to be a clear gradation among the three zoemic
materials: blood, suet, flesh
(i) object blood can be paired only with the agent officiating
priest(s)
(ii) The object suet can be paired with agents officiating priest
or offerer.
(1) offerer is found in the following praxemes:
(a) remove (offerer, suet)
(i) This is performed by the offerer in the well-
being offering (Lev 3:3; 7:30), and perhaps in
reparation and purification offerings as well
(Lev 4:31, 35).
(b) set (offerer, suet, t=breast)
(c) bring (offerer, suet, t=priest)
(i) The last two praxemes must be performed by
the offerer in the case of wellbeing offerings,
not in purification and reparation offerings.
(2) officiating priest is the agent in the praxeme set/
toss (priest, suet, upper surface).
(3) Both officiating priest and offerer participate in the
praxeme raise (suet); but it is the priest who performs
this atomact.
(iii) The object flesh may be paired with the agents priest or
offerer:
(1) in the sacrificial types purification and reparation,
ingest (priest, flesh)
(2) in the sacrificial type wellbeing, bring (offerer, breast,
t=priest) 33
(b) The object zoeme (in toto) is paired with
(i) the agent offerer in the praxeme bring (offerer, zoeme)
(ii) the agent anonymous in the praxemes slaughter (anonymous,
zoeme) and cut apart (anonymous, zoeme).
33
In the context of !wholeburnt, one finds the praxeme cut apart (anonymous, zoeme), which
implies that contact with the flesh is not limited to the offerer and to the officiating priest.
11:50:13:06:14 Page 22
Page 23
....................................................................................
21. : + is highly restricted

(a) For targets subsumed under sanctuary and bronze altar, the
agent must be officiating priest(s).
(i) If the target is subsumed under inner sanctum, the agent
must be high priest.
(b) In P the component target is only rarely found in other
contexts, for example:
(i) t=offerer (specifically, offerer’s earlobe, thumb, and toe)
in the context of daub (priest, blood, offerer), where g=
officiating priest (Lev 14:10–31).
(ii) t=breast in the context of set(g=offerer,o=suet,t=breast)
(iii) t=priest in the unique context of Lev 8 (|| Exod 29), where
Moses is the agent
(c) If atomact bring is taken into consideration, and the place or
agent to which the object is brought is considered a target, then
one might include:
(i) t=officiating priest (e.g., Lev 2:2)
(ii) t=entrance to sanctuary (e.g., Lev 1:3)
22. : +, like +, is highly restricted,
and for similar reasons.
(a) The locations on altar and in sanctuary, are paired only with the
agent officiating priest.
(i) If the location is inner sanctuary then the officiating priest
must be high priest.
(b) There is no restriction on the category agent in P, if the location
is by altar.34
Standard Sequences
23. It is convenient to speak of several standard praxemic sequences

with relation to zoemic offerings, recurring in P and in later
systems. These are termed sacrificial types.
(i) Within the animal sacrificial system, there are four common
sacrificial types:35
34
Restrictions concerning the encroachment of impure persons on the sancta are
external to the grammar of sacrifice, and are not considered here.
35
In P these are not terms for praxemic sequences, but rather terms for zoemes offered
according to such sequences.
11:50:13:06:14 Page 23
Page 24
....................................................................................
(1) wholeburnt (HLWE),

(2) wellbeing (OYML$o XBZ),
(3) purification (TAUX), and
(4) reparation (O$oA) offerings.
(ii) In a skeletal form, the following praxemic sequences define
these four sacrificial types:
(1) wholeburnt: toss (blood) #incinerate (flesh and suet, altar)
(2) wellbeing: toss (blood) #incinerate (suet, altar) #ingest
(offerer, flesh) 36
(3) reparation: toss (blood) #incinerate (suet, altar) #ingest
(priest, flesh)
(4) type: purification subdivides into two major subtypes
(exceptions exist):
(a) the eaten purification offering
daub (blood) #pour (blood) #incinerate (suet, altar)
#ingest (priest, flesh),
(b) the burnt purification offering
flick (blood) #daub (blood) #pour (blood) #incinerate
(suet, altar) #incinerate (flesh, outside camp) 37
(5) Others types are found in P, but these are usually viewed
as special cases, not as general types, for example the
ordination offering (OYAWLM) and the paschal rite.38
24. There are zoemic restrictions pertaining to standard sacrificial types
(many of these were noted by Maimonides):
(a) !웨wholeburnt is ungrammatical (Lev 1:3, 10)
(b) ! wellbeing is ungrammatical (Kafah. 10, 18)
(c) !ordination is ungrammatical
(i) but presumably birds can be part of an ordination+1
complex
burnt purification
(d) ! is never attested
36
As noted above, ingest (offerer, flesh) is optional, and therefore not a praxeme strictly
speaking; but consumption by the offerer is almost invariably assumed in P in the case of
wellbeing offerings.
37
See p. 107 n. 6. On the fundamental difference between these two acts of incineration,
see n. 000.
38
On the goat dispatched to the wilderness and the “red heifer” see nn. 000 and 000. Still
others are found outside P as distinct types although in P they appear to be subcategories of
type:wellbeing, namely, firstborn (RWKB) and tithe (R$ o EM) animals in rabbinic literature.
11:50:13:06:14 Page 24
Page 25
....................................................................................
25. There are also zoemic restrictions on the category kind (and on the
combination of categories kind and type)
(a) x웨 (kind:public) is ungrammatical (Kafah. 1963:19)39
(b) ! (kind:public) is ungrammatical (Kafah. 1963:10)
(c) !purificaion (kind:public) is ungrammatical (Kafah. 1963:20)
(d) !wholeburnt (kind:public) is ungrammatical (Kafah. 1963:20)
(e) A layperson’s !purification is ungrammatical (Kafah. 1963:20)
26. The interaction between the categories type, genus, class, and kind is
as follows:
(a) Type:wellbeing and type:wholeburnt, and these alone, combine
with both genera to create four grammatical combinations:
wholeburnt,votive; wholeburnt,volitional; wellbeing,votive;
wellbeing,volitional (Kafah. 1963:18)
(i) In all of these cases, in P, the class must be non-calendric
and the kind private.40
(b) Type:wholeburnt and type:purification occur as class:calendric
and as class:non-calendric.
(c) Type:reparation occurs only as class:non-calendric and only as
kind:private (Kafah. 1963:10).
(i) But other combinations may be grammatical.
(d) A layperson’s x웧purification is ungrammatical (Kafah. 1963:19; with
the exception of Num 15:27 in its original context, on which see
§ 2.4.10).41
(e) In P there is but one clear case of a private calendric offering,
the paschal f웧.
(f) In non-P material there is mention of OYMYH XBZ, which,
according to its name, is calendric (OYMY = year), and according
to the contexts in which it appears, is private.
(g) Both genera––votive and volitional––are by definition non-
calendric.
39
See, however, p. 200 n. 4.
40
In rabbinic literature, one finds non-calendric volitional public wholeburnt offerings
(e.g., m. Sheq. 4,4); similarly, one finds type:reparation offered, for all intents and purposes,
as genus:volitional (b Zeb 76b).
41
Cases in which a prominent figure like Noah offers the zoemes mentioned in
Num 29:1–5 (including a male goat for purification) on behalf of himself and his household
(Jub 7:3) can hardly be considered instances of a layperson’s offering.
11:50:13:06:14 Page 25
Page 26
....................................................................................
(h) Type:wholeburnt and type:purification occur as kind:private

and as kind:public.
(i) A layperson’s xpurification is always an eaten purification offering.
(j) A public xpurification may be eaten (Lev 10:17 in P) or burnt (e.g.,
Lev 4:21).42
(k) xwellbeing (kind:public) is grammatical but rarely attested (Kafah.
1963:13)
(i) It is attested in P and in H:
(1) In P as a non-calendric offering (though its time is
decreed, Lev 9:4)
(2) In H as a calendric offering (Lev 23:20)
(a) Some scholars opine that the latter originated as a
private offering.43
(ii) It is also found in TS 20:1–21:10; 21:12–22:13.44
(iii) !wellbeing (kind:public) is unattested
(l) Type:ordination is non-calendric in P
(i) No offering of this type is found in the calendric lists in P
and H.45
(ii) Outside P, there is evidence of recurrent ordination
offerings,46 particularly when the bronze altar is re-
constructed.
42
In later traditions, public calendric purification offerings in general are eaten; see
Josephus, Ant. Jud. 3.249 (regarding the festival of unleavened bread, see Feldman 2000:303);
m. Zeb. 1:3, m. Men. 11:7; Milgrom 1990:325.
43
See Milgrom, Leviticus 2:2005–8.
44
In the context of the Wine and Oil Festivals. In all appearances, these are wellbeing
offerings. See particularly 20:1–8, 14–16; 22:8–13. Qimron reconstructs [OYM] L$ in l. 1. See
also Yadin 1983, 2.81–2 on 19:3–4, but cf. Qimron’s collation.
45
The inauguration ritual complexes (Lev 8–9) are similar to calendric sacrifices
inasmuch as their timing was specified by  in the biblical text. There are uncertain
hints in P that the ordination offering was not perceived to be historically unique but
periodic, occurring perhaps every time a new High Priest was ordained (Exod 29:30;
Lev 7:37).
46
Ezekiel does not mention ordination offerings, but the coordinate purification
and wholeburnt offerings in ch. 43 may be viewed as constituting a historically unique
ordination+1 offering (see 43:26). See TS 15:1–14 and 15:15–17:4, which may even suggest a
calendric ordination offering. According to rabbinic traditions, the ordination0 offerings of
P were historically unique, but various ordination+1 complexes do recur on rare occasions
(b Men 45a; HMQ 2:14).
11:50:13:06:14 Page 26

Meshel N. S. - The Grammar of Sacrifice PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Meshel N. S. - The Grammar of Sacrifice PDF

Uploaded by

Copyright:

Available Formats

Page 1

The “Grammar” of Sacriﬁce

for Maya, Eyal, Ellah, Hadas, and Itai

THE “GRAMMAR” OF SACRIFICE

List of Figures and Tables

Figure 1. Structure of an agnicayana 15

List of Figures and Tables xiii

Table 1. Zoemes attested in P 37

ABBREVIATIONS USED IN BIBLIOGRAPHY

xvi List of Abbreviations

GENERAL SIGLA AND ABBREVIATIONS

List of Abbreviations xvii

The pronunciation symbols follow the simpliﬁed system used by Merriam-

: (small caps) is used in contradistinction from “grammar” (lower-

TECHNICAL SACRIFICIAL TERMS

asham see reparation oﬀering

List of Deﬁnitions xix

SETS USED IN THE GRAMMAR

sacriﬁcial class = {calendric, non-calendric}

SIGLA USED FOR FORMAL ANALYSIS

Sigla for Zoemics (see p. 000 Table 1)

bovine (“cow” in colloquial English, e.g., RWo$o)

List of Deﬁnitions xxi

Sigla for Jugation

jx In words: a sacriﬁcial material by the name of j on Level x

xxii List of Deﬁnitions

Sigla for Hierarchics

hx In words: a sacriﬁcial material by the name of h on hierarchic Level x

Sigla for Praxemics

a(g,o,t,l) a praxeme consisting of atomact a, agent g, object o, target t and

Masked in the form of a many-headed snake, according to an Indian

2 The “Grammar” of Sacriﬁce

granted that rituals have such “grammars.” Thus, structure of sacriﬁcial

4 The “Grammar” of Sacriﬁce

allegory, explains, “Attainment of faultless speech is the attainment of

1.2 MAIMONIDES’ GENERALIZATIONS

One example may demonstrate the applicability of Patañjali’s insights to a

6 The “Grammar” of Sacriﬁce

1.3 MODERN ANTHROPOLOGY

8 The “Grammar” of Sacriﬁce

1.3.1 Hubert and Mauss

In its modern anthropological and comparative context, the association of

10 The “Grammar” of Sacriﬁce

discovering certain universals of sacriﬁce. If a particular structure is shared by

1.3.2 Lévi-Strauss, Structuralism, and Neo-Structuralism

The application of linguistic concepts to the study of non-linguistic cultural

12 The “Grammar” of Sacriﬁce

analogy to language, here in a form of graphic representation, is retained:

1.3.3 Frits Staal and the Shift to Generativism

14 The “Grammar” of Sacriﬁce

concentrated on the relationship between this structure and syntax in natural

In an article published in 1979, “The Meaninglessness of Ritual,” Staal

Fig. 1. Structure of an agnicayana.

16 The “Grammar” of Sacriﬁce

able relevance of his empirical data, and a number of inaccuracies.67

1.3.4 Recent Scholarship and the Heidelberg Ritualdynamik School

Staal’s provocative work is not intended as a grammar of any ritual system;

18 The “Grammar” of Sacriﬁce

1.4 METHODOLOGICAL CONSIDERATIONS

1.4.1 The Siglum “P”

Throughout our discussion of biblical passages, I will use a number of sigla

20 The “Grammar” of Sacriﬁce

22 The “Grammar” of Sacriﬁce