Professional Documents
Culture Documents
(Studies in The History of The Language Sciences 23) Fred W. Householder - The Syntax of Apollonius Dyscolus-John Benjamins Publishing Company (1981)
(Studies in The History of The Language Sciences 23) Fred W. Householder - The Syntax of Apollonius Dyscolus-John Benjamins Publishing Company (1981)
(Studies in The History of The Language Sciences 23) Fred W. Householder - The Syntax of Apollonius Dyscolus-John Benjamins Publishing Company (1981)
General Editor
E. F. KONRAD KOERNER
(University of Ottawa)
Volume 23
Fred W. Householder
by
FRED W. HOUSEHOLDER
Indiana University
1981
The preparation of this volume was made possible in part by a
grant of the Translations Program of the National Endowment
for the Humanities, an independent federal agency.
I must first thank Indiana University for the sabbatical leave on which I began
this task, and the National Endowment for the Humanities for their generous
financial support. T h e n Mr. O. Stephen Pillinger, besides p r e p a r i n g the index,
has helped in many ways, and Professor W. Keith Percival of the University of
Kansas has supplied various kinds of assistance. Professor Gregory Nagy of
H a r v a r d University provided me with help to speed my work in Widener Library.
Professor Luigi Romeo of the University of Colorado, besides giving valued
encouragement, lent me his copy of Egger's interesting book. My son-in-law, Dr.
Dennis Prindle of Ohio Wesleyan University borrowed a copy of Buttmann's
G e r m a n version for me. T h e Indiana University Linguistics D e p a r t m e n t se
cretaries, particularly Ms. Nancy Quinn, have been helpful in many ways, from
dealing with the manuscript and various typed versions to p r e p a r i n g the applica
tion to N E H . And my wife, Clarice Householder, has been a constant support
throughout the project, and has given particular help with the index. T h e
camera-ready copy has been ably p r e p a r e d by the typesetters and paste-up people
at the Indiana University Printing Plant.
1. Introduction
This translation is made, in the main, from Uhlig's text, with about seven or
eight minor changes, mostly (a) where I have preferred a reading which Uhlig
kept in his apparatus, (b) where I have simply omitted a bracketed passage, (c)
where I have given the text of a quotation in a form closer to the accepted text of
the poet being quoted (provided, of course, Apollonius' point is not affected by the
change). T h e goal of this, as of any translation, is to say in contemporary English
exactly what the author said in Greek, but, in this case, without allowing him the
knowledge of any linguistic theory m o r e recent than his own. This means, neces
sarily, that some technical terms may have a slightly different set of associations for
us than they would have had for him, but I have tried, in my comments, to call the
attention of the r e a d e r to this fact. Words (generally only a few) placed in single
brackets are to be read directly as part of the English text, but have no localized
equivalents in the Greek. Within double brackets are longer comments, including
comparisons of various m o r e recent linguistic theories with those of Apollonius,
and occasional references to the scholarly literature. Surprisingly, perhaps, the
literature dealing directly with the Syntax is very modest in extent, most of it prior
to Uhlig's edition. T h e only complete translation into a m o d e r n language is
Buttmann's (1877) into German, which helped me a bit in spots, despite Uhlig's
unrestrained contempt for it. Uhlig mentions three Latin versions, only one of
which, that of Portus, was ever published. This I was able to secure only after my
translation was virtually finished. Uhlig's text is accompanied by an Argumentum or
summary which often contains long stretches of verbatim Latin translation, and
many passages were translated much earlier by Priscian, whose version is likewise
published by Uhlig at the foot of the page. Most of these aids were of use all
t h r o u g h the task; a scattering of articles a n d m o n o g r a p h s were occasionally
helpful, particularly Egger (1854), who gives French versions of several long
passages a n d a few shorter bits; C a m e r e r (1965), for Apollonius' treatment of
mood, tense a n d aspect; Egli (1970), for Stoic terminology; Donnet (1967), for
various points, but particularly the technical terminology. For this last, however,
the most helpful aids were the two Indices Verborum (really also vocabularies with
Latin equivalents) in Uhlig's edition: one, the briefer one by Uhlig which comes
first (in vol. II, technically), to Uhlig's commentary on the Syntax; the other, at the
end (in vol. III), R. Schneider's lexicon to the whole Apollonian corpus. T i m e a n d
again these saved me from foolish errors, mostly caused by the dangerous as
sumption that Greek words in Dyscolus have meanings closely related to those of
the same words in Plato or Aristotle. Schneider's brief sketch ("Tractatus d e
Apollonii consuetudine", vol. I I I , p p . 141-161) of the chief grammatical idiosyn
crasies of Apollonius, which shows many features c o m m o n to Greek of the
imperial era a n d foreshadowing M o d e r n Greek, was also often helpful.
T h e degree to which Dyscolus was a Whorfian prisoner of his technical
vocabulary is striking. Suppose we consider some of the simplest a n d most basic of
2 THE SYNTAX OF APOLLONIUS DYSCOLUS
1
I wish here to thank the audiences before whom I have delivered earlier versions of this paper, at
the University of Colorado and at Indiana University, as well as the audience at the University of
Hawaii, and at the APA-MGSA meetings in Atlanta, 1977.
APOLLONIUS DYSCOLUS AND THE ORIGINS OF DEEP STRUCTURE 5
are also interesting: most often (20 times or so) either "man" or "grammarian";
less often (10 or m o r e times) "philosopher", "slave", "boy", "X's son", "X's friend";
least often "teacher", "rhetor", "poet" and "gymnasiarch". T h e . situations con
j u r e d u p by the sentences are often instructive; consider the following sample.
(1.19) Aristarchus is reading. (1.23) Studying is fine. (1.23) I enjoy writing. (1.31)
Who's moving? Who's walking? Who's talking? (1.46) This is an A, this is a B. (1.46)
This is a man, this is a horse. (1.46) T h e teacher calls this an A, this a B. (1.46) T h e
boy erased the A. (1.47) H e erased the point of the A. (1.64) T r y someone else, not
me. (1.93) We finished r e a d i n g Book I of Alcaeus. (1.101) My father is a philoso
pher. (1.106) T h e o n wants to be called a g r a m m a r i a n . (1.111) Let the boy who was
disorderly be beaten. (2.6) H e beat me personally. (2.16) egoge is an Attic form.
(2.29) T r y p h o n is teaching T h e o n . (2.32) Whoever is present, let him read. (2.61)
Let's study today, you in school and I at home. (2.70) I'm beating you. (2.70) Give it
to me. (2.77) T o d a y I watched you. (2.78) You must beat him while I watch.,(2.86)
H e blamed you, yourself. (2.106) I'm teaching my son. (2.141) I flayed you. (2.149)
T h e y beat each other. (3.8) Who beat you? (3.26) Write and get u p ! (3.29) I got
tired of writing. (3.40) I am writing a n d he is writing, too. (3.52) We children work
hard. (3.57) If you read, pay attention. (3.57) If you talk, t u r n a r o u n d . (3.75) Since
studying is deficient, let's study. (3.81) T r y p h o n should reward you. (3.92) If you
wish, I will not read. (3.101) Let the door be shut. (3.126) Give me p a p e r to write
on. (3.152) Alcaeus' poems are in Aeolic, Alcman's in Doric. (3.157) I am flayed by
T r y p h o n . (3.180) So-and-so isn't listening to what is being said. (4.22) Why are you
h u r t i n g T h e o n ? (4.23) Why did T r y p h o n make a mistake? (4.23) In what respect is
the sentence incomplete? (4.26) Because I read more than you, I'm smarter than
you. (4.52) It's possible to study.
All these (and many more) seem to clearly evoke a classroom in which a
teacher is trying to keep o r d e r and teach a g r o u p of boys to read poetry (especially
lyric) a n d to write, sometimes asking the boys grammatical questions about the text
being studied. Whether or not Apollonius was poor, he was certainly a school
teacher. Not all sentences, of course, are from the classroom, e.g. (1.110) "Ptolemy
the ex-gymnasiarch was honored." or (1.124) "How big is the crowd in the
market-place?" or (1.137) "A horse kicked the man." or (2.105) "You, who are my
servant, are running." (2.138) "I d u g a field of my own." or (3.99) "May my son
t u r n out to have won!" A few are typical Stoic examples, as at 3.89 "Because I'm
walking, I'm moving." But even some of these may be linked u p to a school
situation in one way or another. Apollonius may have been "Difficult" in still
another way: p e r h a p s he was a h a r d teacher.
And g r a m m a r , as Dyscolus taught it, originated in the school situation. What
was taught? How was it taught? What was its purpose? It is sometimes said that
Greek boys were taught to translate old writers into contemporary language, but
this seems, at best, a partial view, and, in general, quite implausible. But the boys
were certainly taught these old writers. In the fifth century they were apparently
all poets (to j u d g e by allusions especially in Aristophanes a n d Plato), and the boys
memorized a lot of this verse, and were taught to explain what h a r d passages
meant, in various ways. Paraphrase was surely one way. Another was the classifi
cation by figures of speech and tropes.
APOLLONIUS DYSCOLUS AND THE ORIGINS OF DEEP STRUCTURE 7
2
Robins' version (p. 31 of A Short History of Linguistics, Indiana University Press, 1967) of the
definition goes: "Grammar is the practical knowledge of the general usages of poets and prose writers.
It has six parts: (1) accurate reading aloud with due regard to the prosodies; (2) explanation of the
literary expressions in the works; (3) the provision of notes on phraseology and subject matter; (4) the
discovery of etymologies; (5) the working out of analogical regularities; (6) the appreciation of literary
compositions, which is the noblest part of grammar."
8 THE SYNTAX OF APOLLONIUS DYSCOLUS
tion) "For a man, indeed, to become good truly is hard." But other discussions of
such figures, like Aristotle's treatment of poetic words, clearly suggest that the
poet is deliberately altering the natural or p r o p e r structure. Something of this
ambiguity can also be found in Apollonius. These four relations or types of
linguistic change are clearly relatable to the general classification of changes m a d e
by Aristotle in a n u m b e r of places, e.g. in Categories 15a 13 or On Coming-to-be and
Passing-away 314b27: movement (= transposition), growth or increase (= inser
tion), diminution or decrease (= deletion) a n d alteration or change p r o p e r (=
substitution). These are in addition to the two basic operations of coming-into-
existence (i.e. insertion at a different level) a n d death or destruction (i.e. deletion
at a different level). 3
T h e three chief theoretical notions of ancient linguistics, besides the notion of
transformation a n d its subclasses which we have just discussed, are (1) the idea of
sentence-types or grammatical illocutionary forces, (2) the notion of two or m o r e
analogous layers (like Martinet's double articulation or Hockett's dual patterning)
of structure (Plato in the Cratylus passage—432A-C—cited above has three levels:
letter in word, word in sentence, sentence in discourse; Apollonius Dyscolus at the
outset of Book I of the Syntax inserts the syllable between letter a n d word), a n d (3)
the theory of what we mistranslate as "parts of speech", which should be "parts of
the sentence". V a r r o a n d Sextus Empiricus have a great deal of talk about a
quarrel between "anomalists" and "analogists", and the definition of Dionysius
mentions both "usage" (the war cry of the anomalists) a n d "analogy". Apollonius
Dyscolus has no comment on this debate at all; he never uses the word anomaly
(anomalia) at all, and his few uses of analogos, analogia, analogein are all quite
non-polemical, except perhaps the sentence in 1.60 (p. 36b) arguing that one must
know both by intuitive grammatical skill ("usage", h e r e tribe) a n d by explicit
knowledge of tradition (paradosis, i.e. the usage of great writers) a n d the rules
implicit in the practice of these writers. Clearly neither anomaly nor analogy alone
3
Ancient names for all these things may interest some readers. I will give both nouns (like
"deletion") and verbs (like "delete" or "disappear"). Delete/deletion: aphairein/aphairesis in the Cratylus
and in the Poetics and in etymological contexts generally, where it is sometimes restricted to initial
position; huphaireinlhuphairesis in Apollonius Dyscolus, especially for non-initial letters; apokopteinl
apokope for phonological deletion in final position; phthineinlphthisis "diminution" or phtheireinlpthora
"destruction" in Aristotle elsewhere, demereldemptio—of letters—in Varro, de Ling. Lat. 5.6, elleipeinl
elleipsis generally for deletion of words in the grammarians, including Apollonius Dyscolus. Add/
addition = insert/insertion = exceed/excess etc.: prostithenailprosthesis (addition) in the Cratylus,
epekteineinlepektasis (prolongation) in the Poetics, emballeinlemboli (insertion) in the Protagoras, adderel
additio in Varro, pleonazeinlpleonasmos (redundance, mainly of words) in Apollonius Dyscolus and other
grammarians, parelkeinlparolke (prolongation, redundance of words), auxaneinlauxesis in Aristotle, for
whom both insertion and deletion would be changes kata poson "in quantity". Substitute/substitution,
alter/alteration, change, etc.: exallassein/exallage in the Poetics and sometimes in later grammarians,
including Apollonius Dyscolus once or twice, commutarel commutatio in Varro, alloiounlalloiosis in
Aristotle (more often in a general sense of "change" of any kind); for Aristotle this is also a change kata
ousian "in substance" or kata poion "in quality"; enallassein/enallage and hupallasseinlhupallage in gram
marians, including Apollonius Dyscolus, and sullambanein/sullepsis, specialized for the substitution of
one for many, an "amalgamation" transformation, also in Apollonius Dyscolus and other grammar
ians. Transpose/transposition or move/movement: metatithenailmetathesis in the Cratylus and generally
for movement of letters,\kata topon "in place" for Aristotle, huperbibazeinlhuperbibasmos or huperbaton for
movement of words in the Protagoras and later in all grammarians including Apollonius Dyscolus,
anastrepheinlanastrophe also for movement of words or for movement of the accent (to the left) in
Apollonius Dyscolus and other grammarians. Generic words for all types of change or transformation
include alloiounlalloiosis in Aristole and metalambanein/metalepsis in Apollonius Dyscolus, who may also
sometimes use the terms specifically for a substitution transformation, commutolcommutatio in Varro.
APOLLONIUS DYSCOLUS AND THE ORIGINS OF DEEP STRUCTURE 9
4
Gregory Enos, in a recent unpublished paper, argues that we may derive from this both
Apollonius Dyscolus' notion of the origins of language and his implicit phrase-structure rules. This is,
to some extent, true, but it is hard to believe that Dyscolus considered this matter explicitly at any time.
10 THE SYNTAX OF APOLLONIUS DYSCOLUS
(III) only after all case-inflected words (I-II) and before prepositions, adverbs and
conjunctions (IV).
As for the last topic, the illocutionary force-sentence type theme, this is, again,
implicit in Apollonius Dyscolus' treatment of the moods in Book III of the Syntax
(88-147, p p . 244-276b), and in the paraphrases he offers in several passages (e.g.
1.51, 3.21, 3.61) for commands, wish-prayers and assertions. But it is sometimes
hard for us to avoid confusing his inflectional category (mood) with illocutionary
force, although as a rule (as argued by E. Adelaide Hahn in TAPA) he restricts
psuchiche diathesis ("mental disposition") to illocutionary force, as in 3.190 (p.
301b), whereas enklisis ("inclination, mood") refers to the actual verbal inflection,
as in 3.125 (p. 265b) (but note 3.95 and 125). T h u s there is an enklisis called
subjunctive, but it has no psuchiche diathesis corresponding, and is invariably
derived by transformation from an underlying indicative, the transformation
being triggered by a variety of conjunctions (mainly ean " i f and hina "in o r d e r
that"). This doctrine is e x p o u n d e d in 3.132-134 (pp. 269-271), and owes its
beginnings in part (but only in part) to the fact that a phonological change which
had eliminated the o-o a n d e"-ei contrasts by Apollonius Dyscolus' time also
eliminates the phonological difference between present indicative a n d subjunc
tive. This marks clearly the beginnings of the M o d e r n Greek pattern in which the
particles an, na, as and tha when combined with verb forms having primary
endings give these forms the n a m e of subjunctive. Apollonius Dyscolus even
comments (in 3.134) that when an (not the same as M.G. an) is combined with an
aorist indicative, that indicative—which has secondary endings—retains the name
of indicative.
However, it is for his use of the transformational devices (especially ellipsis,
hyperbaton and enallage) that Dyscolus deserves o u r greatest attention. It is from
him, directly or by way of Priscian, that these procedures reached the g r a m m a r
ians of the Renaissance a n d of m o d e r n times.
In these later grammarians it is often possible to find a clear distinction
between two uses of the rules: one for explaining poetical expressions, the other
for ordinary language. T h e first regularly starts with a grammatical, normal-
sounding structure and converts it into a poetical, somehow abnormal one; the
second starts with an underlying form which may or may not be grammatical as it
stands a n d converts it into a normal, everyday structure. T h e distinction is not
quite the same as that between a transformation of the Zellig Harris type and one
of Aspects form, because both input a n d o u t p u t of Harris transformations are
grammatical and may be normal (not necessarily figurative or poetical), while only
the output of Aspects transformations is grammatical. But in Dyscolus some trans
formations closely resemble the Harris type, as for instance the passive transfor
mation (Syntax 2.141, p p . l74-5b, 3.157, p p . 282-3b, both passages translated by
Priscian in 17,135, p . 176 and 18.137, p . 271), which clearly derives a normal,
grammatical passive sentence from an equally normal grammatical active. But
Apollonius' language is not always clear and unambiguous. H e uses a considerable
variety of expressions to relate pairs of sentences (or, sometimes, pairs of phrases),
some of which clearly are merely m e a n t to indicate synonymy or paraphrase-
relation rather than derivation, and in some cases the same verb may be used for
both right-pointing and left-pointing arrows. T h e only cases where we can be
positive that we know the correct direction is where one m e m b e r of the pair is
APOLLONIUS DYSCOLUS AND THE ORIGINS OF DEEP STRUCTURE 11
that two of these rules are commonly violated in ordinary speech as well as in
poetry is taken care of by the important observation (with reference to the rule
about enclitic pronouns) of 2.77, 132b: "These rules are p u t forward so that we
may recognize cases of hyperbaton even when they occur in ordinary everyday
speech, not just as something deviant in poetry."
Another such is enkeitai or enesti "is present in", i.e. "is implicit in", for instance
in 2.51, 118b "the singular n u m b e r a n d the nominative case (of the subject) are
implicitly present in grapho ("I write"), even t h o u g h we do not claim that grapho is
elliptical (for ego grapho)", 2.142, 175b (a similar case); 1.51, 31b (where the verb
"(I) declare" is said to be implicit in any indicative verb form); 1.41, 25b (where a
feminine n o u n "female" is taken to be implicit in phrases where a feminine article
occurs with an indeterminate noun); 1.51, 117b ("I affirm" implicit in positive
indicative sentences). If these are to be interpreted generatively, they would be
cases of sullepsis ("amalgamation"), where several underlying words or ideas are
reduced to one word. Apollonius Dyscolus uses the word sullepsis in 3.38, 216b,
where "we" is said to receive sullepsis of second and third person, so that " T r y p h o n
spoke to me a n d you and Dionysius" becomes, by sullepsis " T r y p h o n spoke to us".
T h e same proposal occurs in 3.46, 220b, where we also have forms of the perfect
sunesteke "is constituted, consists (of)" as well as the n o u n sullepsis. In fact, all of
Dyscolus' uses of sullepsis are in the analysis of "we" and of "you (pl.)", which is said
to be by sullepsis for "you and you" or "you a n d he". O t h e r examples that might be
so classified by other grammarians are explained by Apollonius Dyscolus with
different words, as illustrated above in the examples with enesti and enkeitai.
Similar examples are explained in 2.50, 117b with the expression paruphistatai en
"is implicit in"; the n u m b e r "one" is implicit in the singular name Ajax, "son (of)"
in the patronymic Kronides, and so on. But it is the deloumena "meanings" of "one"
and "son" etc. that are said to be implicit, so it is dubious if Dyscolus is really
proposing a Generative Semantics type d e e p structure. T h e same verb (paruphis
tatai) is used at 1.17, 12b; 2.56, 120b; 3.159, 283b; and 3.176, 293 to allege the
implicit presence of a subject NP ("nominative") for every finite verb, as in the
grapho example with enesti cited above, an example which is mentioned again (with
a form oiparuphistamai) in 2.51, 118b and 3.96, 249b. T h e verb also occurs in 4.25,
318b, in 4.68, 338b (which has "day" implicit in the words for "today", "tomorrow"
and "yesterday").
One other word occurs in similar situations, analuetai (eis)/analusis "is analyz-
able into", "analysis". This is contrasted with paragetai (apo) "is derived from" in
passages like 2.117, 158b "possessives of n o u n s and p r o n o u n s are derivedfrom the
same case they are analyzable into, the genitive." (So also in 1.103, 62b; 2.104, 150;
4 . 1 7 4 , 292b.) C o m p a r e also 3 . 6 1 , 2 3 0 b : " F r o m t h e genitive Hektoros is
constituted/derived (apoteleitai) the word Hektorides ("son of Hector"), which in
cludes (prosginetai) an a d d e d huios ("son"), and so also it is analyzable into Hektoros
huios ("son of Hector")." This passage goes on to analyze a comparative gorgoteros
"fiercer" into mallon gorgos "more fierce" and hippon "horse-barn" into hippous
sunechon "containing horses", and then declaratives into "I declare + infinitive",
optatives into "I wish/pray + infinitive", and imperatives into "I o r d e r + infini
tive". In many ways this passage parallels the one cited above from 2.50, 117 where
the expression is paruphistatai en. A n o t h e r similar passage with ten analusin echei
"has the analysis" occurs in 2.170, 192b where, again, Kronides is analyzed into ho
APOLLONIUS DYSCOLUS AND THE ORIGINS OF DEEP STRUCTURE 13
tou Kronou huios "the son of Kronos" and hemedapos "compatriot" into hemdn polites
"our fellow-citizen" (an analysis later rejected by Dyscolus).
But the most intriguing term is metalambanetai (es X) (ek Y) "is transformed
(into X) (from Y)", since it looks to b e s o clearly generative. But in 4.44, 327b it is
said that kathizd ("sit down"), katheudo ("lie down") and enepo ("tell") are trans
formed into the simple verbs hizo ("sit"), heudd ("lie"), epd ("tell"). All Apollonius
Dyscolus seems to mean is that the prefixes are r e d u n d a n t ; if we are to transform
by deleting them, we have an interpretive rule rather than a generative one. And
the passage quoted above with analuetai and analusin echei (3.61, 230b) goes on
thus: "Every mood may be transformed into n o t h i n g else but an infinitive gov
e r n e d by a word conveying the m e a n i n g of the mood . . . peripatd "I walk" (is
transformed) into hdrisamen peripatein "I declare(d) walking" [i.e. "I declare that I
am walking"]." Again this seems interpretive. However, when he discusses a
similar case in 3.86, 243b, he seems to be proposing an e m b e d d i n g transforma
tion: "If from ploutei Truphon ( T r y p h o n is walking) a n d hugiainei Truphon
(Tryphon is healthy) there arises a transformation (metalepsis) to phdsi Truphona
ploutein (They say T r y p h o n is wealthy) a n d legousi Truphona hugiainein (They say
T r y p h o n is healthy) . . . " (He goes on h e r e to discuss the ambiguity when a
transitive sentence is e m b e d d e d with subject accusative, and to claim that—
underlyingly—the first accusative is always the subject.) O t h e r passages with
metalambanetai that may be interpretive include 2.114, 106-7b (possessive adjective
into genitive), 2.165, 190b (Kronides into ho tou Kronou huios "the son of Kronos"),
2.96, 144b (hoi into heautdi "for h i m s e l f ) , 2.98, 145b (similar rule). T h e usage in
2.107, 152b is puzzling: "When he (Menander) could have said pros tais emais
thurais hesteka "I'm standing by my door" and transformed it into pros tais thurais
mou hesteka (same sense), he actually said pros tais emautou thurais hestek' ego ("I'm
standing by my own door"). T h e reason it is puzzling is that n o n e of the three
forms is Dyscolus' regular underlying form, which would be *pros tais emou thurais,
a pattern he demonstrated just two lines earlier. 5 This pattern appears again in
2.109, 153b: "A simple transformation (metalepsis) (sc. of Iliad 16.753 hee te min
olesen alke) will yield *he autou alke auton apolesen ("his courage destroyed him")".
This is clearly an underlying form for the poetic expression, which rearranges the
word o r d e r putting the subject NP together in first position, replacing poetic
p r o n o u n s by normal ones a n d restoring the prose prefix to the verb. T h e reason I
have asterisked it here (and in the example above) is that normal personal pro
nouns in the genitive cannot stand between article and n o u n in ordinary Greek, a
fact which Dyscolus states as a rule in 2.122, 161b: "We d o not say Hon emou agron
(of my fields) or Hoi emou agrdi (for my field) or tes emou oikias (of my house), but
rather we say either ton emon agron or ton agron mou . . .'" Nevertheless Dyscolus
gives underlying forms with this ungrammatical structure nine times in this
passage (2.106-115, 151-76), not only autou "his", but also autes "her" and emou
"my". And there are other examples of interpretive metalepsis. T h e r e are, how
ever, some passages where metalambanetai seems to have a clear generative sense,
e.g. 3.5, 196b: "*eme hubrisa ("I h u r t me") is transformed into emauton hubrisa ("I
h u r t myself)"—the reflexivization transformation; 3.26, 208b on the derivation of
5
This is the regular position for any ordinary NP which is a possessive genitive.
14 THE SYNTAX OF APOLLONIUS DYSCOLUS
does not, on the basis of its literal or primary sense, imply a process ("is r e d u n d a n t "
is the basic sense) whereas the former seems to ("draws out", "drags in extra").
Once or twice pleonazei suggests a process, as in 2.39, 1 1 l b : "Again the Dorians did
not redundantly insert t- in tautai and toutoi ("these" nom. pl. fem, and m a s c , for
which other dialects have hautai and houtoi)... This is s u p p o r t e d by the fact that all
the oblique cases and the nominative neuter forms begin with t." But this is not
always true: 2.55, 120b; 2.79, 134b; 3.129, 267b. I n the case of parelkeilparolke the
situation is not much different; process seems implied in 1.82, 51b; 4.44, 327; 1.4,
5b; 1.20, 15b; but not in 3.94, 248b; 2.92, 142b. T h e verb is often also used in a
non-technical sense, about the a r g u m e n t , m e a n i n g "It would/wouldn't be
irrelevant/a waste of time to say. . . "
Perhaps one should also mention paratithemi/parathesis, which is sometimes
used for the addition of a word, as in 1.128, 78b, where it is matched with leipsis,
omission (or deletion) of the article; 1.73, 46b; 1.78, 49b; 182, 51b; 2.56, 121b;
2.91, 140b; 3.122, 269b; etc., in nearly all of which the point is the combination of a
word with another to form a phrase or construction of some kind, as opposed to a
c o m p o u n d . It is especially c o m m o n in discussing prepositional phrases a n d articu
lar constructions (i.e. NP's beginning with a definite article), but occasionally other
phrase types are mentioned, as in 4.2.304b where eu noō ("I think well") as a
two-word phrase is contrasted with eunoo as a single verb ("I am favorable").
Similarly epentithemi/epenthesis is occasionally used for the insertion of a vowel, as in
2.118, 159b (insertion of E in Doric forms of the p r o n o u n "you") and 2.119, 160b
(insertion of/ in an Ionic form of "me"), or of a word, as in 1.130, 78b (of the
definite article), 1.147-88b (of the conjunction hai before a relative p r o n o u n ) ,
3.77, 240b (of dei or chre "one must" with the infinitive chairein "greet" in the
letter-headings — clearly an interpretive use, not generative), 4.34, 322b (inser
tion of article after a preposition as a test), etc. Evidently neither expression
functions like the others we have examined.
Finally we come to substitution, for which we find both metatithetailmetathesis
(which also may refer to reordering), and enallassetailenallage. First metathesis /
metatithetai, most often used in phonology: 1.92, 55b (deaspiration); 1.103, 62b
(declension); 2.118, 159b (change of S to T); 2.119, 160b (Y to I); 2.161, 187b
(change from common to p r o p e r noun); 2.33, 109b (change of part-of-speech
classification, e.g. n o u n to adverb); 2.55, 120b (change of accent); 2.99, 146b
(change of number); 3.76, 239b (inflection for person). Next enallassetailenallage:
1.57, 35b (nominative instead of genitive); 1.73, 46b (change of ending); 2.131,
168b (change of person); 3.9, 198b (change of gender); 3.50, 223b (change of
gender suffix); 2.128, 165b and 4.9, 308b (change of accent); 4.62, 335b (change
of preposition/deaspiration); 2.115, 157b (change of case, dative for genitive);
1.116, 7 l b (substitution of one word for another); 2.73, 129b (change of order, i.e.
hyperbaton). T h e r e are a few examples of hupallassolhupallage: 3.27, 209b (change
of gender); 3.30, 211b (change of voice, active for passive); and in 3.63, 232b we
have anthupallassō (infinitive for imperative in H o m e r ) .
Now let us see what needs to be established, and what the evidence is. Given
that all the types of change discussed above (except perhaps epenthesis and
parathesis) are used by Apollonius in grammatical explanation, we need to know (1)
are these changes generative or interpretive? and is there a difference between
syntax and phonology? (2) if generative, who makes them? Is it only poets, or is it
16 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
ordinary speakers as well? (and we exclude cases where the change categorizes
errors) (3) From what base are the changes made? Is there m o r e than one such
base? (4) How much of this theory is d u e to Apollonius, a n d how much is
traditional?
(1) We have seen n u m e r o u s examples which might go either way, but all the
phonology examples a n d many of the syntax ones are clearly generative. This
shows u p best in cases where the assumed base form is surface-ungrammatical, so
cannot be a helpful paraphrase. For syntax we saw such a case above which dealt
with possessive constructions involving genitives of personal p r o n o u n s and pos
sessive p r o n o u n s (Syntax 2.106-115, 152-7b). Another instance is the postulation
of infinitives with the definite article to as subjects of such impersonal verbs as melei
("is a care to"): sentences like *melei to philosophein Platōni ("Philosophy is a care to
Plato") are grammatical only if the to is deleted ("Plato cares for philosophy").
Another is the account of what is now called Equi-NP Deletion in 3.162, 285-6b
where sentences like *boulomai emauton peripatein ("I want myself to walk") are
proposed as underlying boulomai peripatein ("I want to walk"). As in English, it is
conceivable that such sentences may be grammatical when " m y s e l f is contrastive,
but not otherwise. In morphology we may note (3.146, 276b) cases like *edon, a
first-person singular (second aorist) to match the plurals edomen, edote etc. rather
than the edoha ("I gave") which is the normal form (regarded as a "first aorist").
(Since all other persons are considered to be derived from the first singular
(active), as all cases from the nominative singular (masculine), this is also a kind of
abstract base.) Similarly, in 2.19, 102b, *gunaix instead of gune ("woman"), *hudas
for hudor ("water"), nominative singulars. But the most instructive is the account of
the definite article in 1.80-1,49-50b, which shows why the underlying forms of the
nominative singular a n d plural, masculine a n d femine, must be Hos, He, Hoi, Hai,
from which the normal Attic ho, he, hoi, hai are derived by rules which we might call
t-deletion, h-insertion a n d s-deletion. These might all be classed as morphological
r a t h e r than phonological abstract forms. All the really phonological cases are
etymologies which might be a r g u e d to be historical r a t h e r than descriptive (e.g.
hudor from *huor by d-insertion —pleonasmos — in 1.4,4b), although some poetical
cases are perhaps meant to be attributed to the poet himself, e.g. lelachosi for lachosi
(third plural aorist subjunctive oflanchano' in Homer) in the same passage. Many
such examples are found in the first eleven chapters of the Syntax.
T h e second kind of evidence is the statement of firm, exceptionless rules
which are not, in fact, true on the surface in Greek. We noted two such above, one
from 3.49, 223b, which says that the relative p r o n o u n must be immediately
followed either by its verb or a participle; a n o t h e r from 2.77, 131b, which says that
any enclitic, non-contrastive personal p r o n o u n must follow its verb a n d any
contrastive one precede. T h e example quoted from 2.128, 166b above ("who sent
us") exemplifies both of these rules. Apollonius Dyscolus' discussion in 2.77,
131 -2b, cited above, shows that he is quite aware of the abstractness of the rules: "It
is stupid to suppose that these arguments can be overturned by citing counter
examples with enclitic p r o n o u n s preceding verbs and contrastive ones following,
because this rule was developed in o r d e r that otherwise overlooked transpositions,
overlooked because they occur in ordinary speech, might have their recognition,
and so that people might stop imagining that such transpositions are something
exceptional which occurs in poetry only."
APOLLONIUS DYSCOLUS AND THE ORIGINS OF DEEP STRUCTURE 17
This quotation clearly answers question (2) above ("who makes the transfor
mations?"), but (3) ("from what base?") deserves a few m o r e lines. Both speech
and poetry start from the same, perfectly regular, partially abstract base; this is
clear from many of the passages already cited. But it is also clear that poets exercise
a second r o u n d of transformations, particularly substitutions (of poetic words like
algea for kaka "evils" or Achaiois for Hellesin "Greeks" in Iliad 1.2, as explained in
3.49, 223b) a n d phonological-morphological changes of all types (most of the
examples discussed in 1.3-11, 4-8b illustrate this), but also including poetical
huperhata (re-orderings) which go beyond what occurs in ordinary speech (as in the
explanation oilliad\A\ which appears in 1.108, 66b; h e r e anastrophe has moved
the article which belongs with "priest" back to the n a m e "Chryses").
T h e fourth question, alas, which a m o u n t s to the assignment of credit (or
blame) for the development of this theory, cannot be answered safely. We have
enough hints about the use of these devices by other grammarians, beginning with
the passage quoted earlier from Plato's Protagoras, to be sure that most of this
theory, as applied to the interpretation of H o m e r a n d other poets, is earlier than
Apollonius. But there seems to be no surviving evidence for the earlier introduc
tion of an abstract base from which ordinary speech is derived, a n d the passage
just quoted from Syntax 2.77, 131-2b suggests that this idea was not yet wholly
familiar to or accepted by Apollonius' contemporaries. So either he invented it, or
took it over from another grammarian of relatively recent date (or, conceivably,
from some Stoic philosopher). But n o n e of his quotations from T r y p h o n , the only
logical candidate, give any clear indication of this theory, (and all o u r knowledge
of Tryphon's syntactical ideas comes from A.D.). Probably, then, Apollonius
Dyscolus was the inventor of the abstract base.
REFERENCES
Apollonius Dyscolus, De Constructions Libri Quattuor (Peri Suntaxeos) ed. Gustav Uhlig, in Grammatici
Graeci, part II, vol. II. Leipzig: B. G. Teubner, 1910. Reprinted at Hildesheim by Georg Olm
Verlagsbuchhandlung, 1965. (Citations are by book, paragraph, and Bekker's pages which are
given in Uhlig's margins.)
______, Scripta Minora (De Pronomine or Peri Antonumias, De Adverbiis or Peri Epirrhematon, and De
Coniunctionibus or Peri Sundesmon), ed. Richard Schneider, in Grammatici Gram, part II, Vol. I,
Leipzig: B. G. Teubner, 1878.
Enos, Gregory (1978, unpublished) Ontogeny Recapitulates Phytogeny: Apollonius Dyscolus' Conception of
Sentence Generation.
Hahn, E. A., "Apollonius Dyscolus on Mood," in Transactions of the American Philological Association, 82,
pp. 29-48. 1951.
Harris, Z. S. (1957) "Co-occurrence and transformation in linguistic structure",Language 33.283-340.
Hoerschelmann, W. (1880) "Kritische Bemerkungen zu A.D. de Pronomine", Rheinisches Museum
35.373-89.
Priscian, Institutiones grammaticae, ed. Martin Hertz. (Vol. 3 of Grammatici Latini, ed. H. Keil). Leipzig:
Teubner. 1859.
Robins, R. H. A Short History of Linguistics, Bloomington: Indiana University Press. 1967.
Citations from classical authors (e.g. Homer, Plato, Aristotle, Sextus Empiricus, etc.) are in standard
form, usable with any edition.
THE SYNTAX
OR
PERI SUNTAXEOS
(DE CONSTRUCTIONE)
OF
APOLLONIUS DYSCOLUS
translated, and with commentary
by
FRED W. HOUSEHOLDER
APOLLONIUS OF ALEXANDRIA
On Syntax, Book I
1. In our previous publications we discussed the theory of words (phone), as the 3
nature of the subject required; the present work will cover the topic of the
combination (suntaxis) of these words, according to appropriate collocation rules,
into independent sentences, a topic which we have chosen as deserving the
greatest precision because of its essential importance for the interpretation of
poetry.
2. We have seen how the unanalyzable substance unit, the p h o n e m e , long ago
was restricted, in that it would not u n d e r g o combination at r a n d o m , but only
according to p r o p e r rules of structural o r d e r i n g (suntaxis), from which it in part
got its name [[the Greek word stoicheion "letter, p h o n e m e " is related to steicho
"march in order"]]. T h e next higher unit, the syllable, adopted the same restric
tion since structures (suntaxeis) of syllables, when filled out according to rule,
constitute words (lexis). A n d it clearly follows that words, too, being the primes of a
regularly constructed complete sentence (logos), must also accept restriction by the 4
structural rules of syntax. For the meaning (noeton) which subsists in each word is,
in a sense, the minimal unit (stoicheion) of the sentence, and just as the minimal
units of sound compose syllables when properly linked, so, in turn, the structural
combining (suntaxis) of meanings (noeta) will produce sentences (sullabai - as it were
"semantic syllables") by combining words. Just as the word is made of syllables, so
the complete sentence is made by the grammatical collocation of meanings (noeta).
[[Why A.D. uses this Stoic t e r m noeton here (and nowhere else) is unclear. In Stoic
theory it is more or less the generic or abstract which is "meanable" or "thinkable"
as opposed to the particular or concrete which is perceptible (aistheton). O n e really
expects the term lekton, "sayable", which he never uses in the Syntax.]]
3. T h e same parallelism can be shown in the properties [of the three units]. E.g.
(a) reduplication or repetition (1) of letters, as in ellaben (for elaben), and ennepe (for
enepe), (2) of syllables, as in Lelex, pampan, and (3) of words:
1.1 Mōs' age Mosa ligeia (Alcman fr. 1)
("Come Muse, shrill Muse")
1.2 barus, barus sunoikos (Sophocles, fr. 686)
"Heavy, heavy mate"
But even (4) sentences are affected by this, when a whole utterance is repeated,
sometimes significantly, sometimes redundantly.
4. Similarly (b) there is meaningless insertion ("pleonasm") (1) of a p h o n e m e
(other than repetition), as the d in hudōr("water"), which comes from huein ("to
rain"), or the a malaletos ("shout"; as if from laleō "speak"), which may also be taken
as pleonasm of a syllable, since in a broad sense even single vowels [with no
20 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
preceding consonant] are called syllables. But h e r e are some clear examples of
insertion (2) of a syllable: theressi ("beasts" dat. pl.; for thersi), kunessi ("dogs" dat.
pl.; for kusi) lelachosi ("get" 3 pl. ao. subj.;Iliad 7.80 for lachōsi), and many more. (3)
Words, too may be inserted [with no change of meaning], Kathezomailhezomai
("sit"), en-epo/epo("speak"), antiosi en-antios ("opposite"). [[All these are what we 5
would call prepositional prefixes, b u t they counted as words—kata, en—for the
ancient grammarians.]] So also ge in Homer Iliad 5,812 ou su g'epeita/Tudeos ekgonos
essi ("You are not then the offspring of Tydeus!") [and everywhere else], which we
call an "expletive conjunction". [[We would say it has meaning, and is an "empha
tic particle".]] (4) We say also that sentences are sometimes introduced where they
contribute nothing to the meaning, as Aristarchus shows in many cases where he
marks a line [of Homer] as spurious for such reasons.
5. (c) Conversely, there are deletions, (1) of a letter, gaia/aia ("earth") skeptouchos
("sceptre-holding") from skeptron, and similarly phaidimos alongside phaidros
("bright"). Such pairs are almost unlimited, there are so many of them. (2)
Likewise of a syllable, when we get lō for thelō ("wish"), aipolos for aigiopolos
("goatherd"). (3) And also of a word, as in Iliad 9.649 erchesthe ("you are going")
where apo ("back") is required, and 11.271 odinds echousai ("having pains") where
we require the preposition para to complete the sense (i.e. parechousai, "causing").
A n d in Odyssey 9.550:
1.3 arneion d'emoi oioi
("The r a m to me alone. . .")
T h e article is missing. For the sentence requires referential identity {anaphora) for
the ram, and hence shows the deletion of the article.
6. Most of the rest of this book will contribute to the clarification of such matters.
Why did Zenodotus conjecture hdlloi [for hoi alloi "the others"] (in Iliad 2.1, 10.1),
with crasis (sunaloiphe) of the article, unless he was influenced by the needs of the 6
syntax? A n d why did Aristarchus never say that the article was redundant (pleonas
tic, inserted) but describe the poet as habitually deleting articles? So whenever there
was a possibility of using the article, when the sense of the context required one,
Aristarchus, taking into consideration the habitual usage of the poet, rejected it.
For instance in Iliad 10.408, he preferred the conjunction dai to the article hai in
the phrase:
1.4 pos dai ton allōn Troon
("How [the guards] of the other Trojans. . . ?"]
This will be fully explained later, in the appropriate place [ch. 127].
7. It is clear that examples like the following are not cases of apocope, as some
people say:
1.5 all'ana, ei memonas ge (Iliad 9.247)
("But u p , if you are mindful. . .")
1.6 para d'aner, hos katathesei (Odyssey 16.45)
("There's a m a n here who will p u t it down. . .")
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 21
[[where ana " u p " means anastethi "get u p " , a n d para "present" means paresti "is
present".]] but are defective sentences with the verb deleted. How can there be
apocope of a whole word? T h e very n a m e of the change (pathos) proves this, since
any apocope ("cutting o f f ) implies some part of the whole thing being left behind.
8. (d) It seems plausible to propose the following comparison. It sometimes
h a p p e n s to words that they are spelled wrong, but may still either be clearly
understood by hearing or, if that is uncertain, may be corrected by the logical
analysis which we call the theory of spelling. I propose that a similar property
applies also in the case of sentences. W h e n words are defectively linked together,
we call the p h e n o m e n o n solecism, the elements (stoicheion) of the sentence being 7
ungrammatically combined. Now if it is possible to determine the correct spelling
(as suggested above), so also it is possible to determine the correct structure of the
sentence. [[I.e., probably, just as there are categories u n d e r which one can classify
poetical distortions of spelling so as to recover the underlying, "true" form, so also
there are rules of distortion of syntax which may be used to recover the underlying
structure.]]
9. (e) (1) A m o n g both vowels and consonants there are, as we know, some which
regularly precede [[e.g. a, o, e in diphthongs,s initial before mute plus liquid]] [and
others which follow] [[e.g. i, u in diphthongs,s final in words]]; (2) the same thing is
true of syllables, e.g. the syllable eu occurs only at the beginning of words [since it
arises only in a u g m e n t e d verb forms or as the result of crasis], a n d similarly all
syllables beginning with h in ordinary Greek. A n d postpositive syllables include,
for instance, all those beginning with the clusters -gm, -km, -chm-[which cannot
begin words, unlike dm-, tm-, gn-, kn-, chn-, which can], or those e n d i n g in word-
final clusters such as -ls, -rs, -ns, etc. (3) T h e analogy holds also with words; h e r e we
n a m e a set of words "prepositions" ("standing before"), and speak of "prepositive
articles" [[i.e. the definite article, which regularly precedes at least some word of its
phrase, usually the head noun]] and "post-positive articles" [[i.e. relative pro
nouns, which tend to follow their antecedents]] a n d likewise adverbs (epi- + remata,
"upon + verbs"] which are so-called m o r e because of their syntax than their
meaning. (4) A n d a similar thing can be found also for sentences [i.e., here,
clauses], when a hypothetical syllogism is true, with antecedent before consequent,
just as it is contained in the first premise of the a r g u m e n t , as, for instance, if
someone should say:
1.7 ei peripatei Dionusios, kineitai
("If Dionysius is walking, he is moving. [He is walking,
he is moving]")
But it is false [if one says]:
Dionusios kineitai ("Dionysius is moving"),
peripatei (".'.he is walking")
W h e n the two statements are reversed, the whole is no longer true. [[cf. 3.89 and
the Stoic account in Sextus Adu. Math. 8.244.]]
10. (f) (1) A single vowel is sometimes divided into two syllabically, e.g. hedeiheade
("it was pleasing"), a n d also two may be contracted (sunaleiphetai) into one, beleal
bele. (2) A n d a syllable may be divided into two (taking "syllable" in its p r o p e r
22 THE SYNTAX OF APOLLONIUS DYSCOLUS
sense) koilon/koilon, and, conversely, two may be contracted (suneisi) into one, as
when we change three-syllabled gerdi to two-syllabled gerdi, or Aides into Hdides. 8
[[From a m o d e r n point of view, all these examples are cases of the same p h e n o m
enon, contraction, but some of the contractions became established as normal
variants at an earlier date than others.]] (3) Similarly with words, when com
pounds are resolved as polls akre for akropolis, or chorōi kale ("fair in the dance", Iliad
16.180) for kallichores, a n d furthermore sun agrion ("wild boar", Iliad 9.539) for
suagron. And also separate words may be joined, when we read as single words
pasimelousa ("all famous" Odyssey 12.70) and keressiphoretous ("fate-driven" Iliad
8.527). [[Both of these are c o m p o u n d s whose first m e m b e r is a dative plural form
which could also be taken as an i n d e p e n d e n t word.]] (4) In sentences, also, two or
more may be united by accompanying conjunctions, as for instance, conditional
(sunemmenos) or causal (parasunemmenos) or c o m p o u n d (sumpeplegmenos) sentences;
or else, conversely, when conjunctions are omitted, a sentence may be broken into
two, as, for instance:
[[This implies that each of the conjoined verbs is a sentence, in some way, though
in the first case both are participles.]]
12. (III) (h) F u r t h e r m o r e , just as (1) some p h o n e m e s are vowels, which are 9
complete sounds even in isolation, and others consonants, whose pronunciation is
not possible without vowels, so also (2) one can consider words of two kinds. Some
words are, like vowels, independently speakable, e.g. verbs, nouns, p r o n o u n s ,
adverbs, when they can be applied to actions in the situational context, as when we
shout kallista ("very well"—i.e. "bravo") at performers who are doing something
just right, or hugios ("soundly") or kalōs ("well"). O t h e r words resemble consonants,
and just as they require vowels, so these require the presence of some of the
aforesaid parts of speech; this is the case with prepositions, articles a n d conjunc
tions. These words always co-signify; for instance, we say di'Apolldniou ("by Apol-
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 23
lonios") with the genitive, implying that Apollonios was aware, but di'Apollonion
("because of Apollonios") with the accusative, implying that he was to blame. [[I.e.
the preposition dia by itself doesn't mean anything, but dia + gen. means one thing,
and dia + acc. means another.]] Conjunctions, too, may vary in force according to
their position in the sentence or the context, since etoi is understood conjunctively,
{sumplektikos) [[i.e. "on the one hand", not "either", which it could also mean]] in
Iliad 1.68:
1.11 etoi ho g'hos eipon kat'ar'hezeto
("He said this and sat down.")
Here etoi is taken as equivalent to men, as is shown by the immediately following
context, where de appears:
toisi d'aneste
("Then Kalchas stood up.")
Elsewhere etoi is disjunctive (diazeuktikos):
1.12 etoi neos estin ee palaios
("He is either young or old.")
T h e case is similar for articles; when conjoined with nouns they have their normal 10
force, but when they have no noun with them, thev become a sort of p r o n o u n , as
will be explained in detail later [ch. 28], where we will also show the reason, since
other parts of speech, too, may behave the same—e.g. very often nouns may be
heard used adverbially.
13. (i) Perhaps also the traditional order of the letters of the alphabet, with A first,
followed by B (etc.), will suggest the order of the parts of speech, in which noun
appears first, followed by verb and the rest of the parts of speech and likewise the
order of the cases, with nominative, genitive, and the rest, or of the tenses of the
verb, with present (enestōs) first, then imperfect (paratatikos) and the rest; or the
order of genders—masculine, then feminine, and third the negation of these, the
neuter, and similarly in many other cases which we will reason about individuallv.
But, for the above-mentioned reasons, it's important to compare and justify the
ordering of the parts of speech. Perhaps someone, flaunting his own ignorance of
such matters, may foolishly urge that there's no need to bother with such investi
gations, suggesting that these matters have been a r r a n g e d arbitrarily and by
chance. But such people can also propose that, in general, nothing is o r d e r e d and 11
there are no errors of ordering; but this would be quite idiotic. And if you concede
order in some cases, you must grant it in all. [[Note that the o r d e r in question here
is merely the order in which parts of speech etc. should be discussed in a grammat
ical treatise. Over the centuries a more or less traditional a r r a n g e m e n t had
become standard.]]
[[This is now the end of the section discussing the parallels between different
levels, which began in Chapter 2 and was carried through 9 successive sections,
which I have lettered a-i.]]
14. T h e ordering is a reflection of the complete sentence, quite properly placing
the n o u n first, and after it the verb, since any sentence which lacks (either of) these
is not complete. [[It is curious that this argument, which works pretty well for
24 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
English, is quite unconvincing in Greek unless you follow the Stoics in requiring
specific definite subjects for completeness.]] It's easy to test this with a sentence-
structure containing all the parts of speech; if either n o u n or verb is deleted, then
the sense of the sentence is indeterminate, but if any of the others is removed,
there's no defect in the sentence at all.
1.13 ho autos anthropos olisthesas semeron katepesen
("The same man slipping today fell down.")
[i.e. "slipped and fell". T h e parts in o r d e r in Greek are: article,
p r o n o u n , n o u n , participle, adverb, preposition, verb.]
This contains everything except a conjunction, which would entail the addition of
another sentence. Now suppose you delete (elleipein) the n o u n or the verb. T h e n
the sentence is defective, d e m a n d i n g in one case a verb, in the other a noun:
(2) ho autos olisthesas semeron katepesen
("The same (one) slipping today fell down.")
[[But in Greek this is a perfectly O.K. sentence, since ho autos "the
same", means, in such cases, "the same o n e " or "the same man."]]
(3) ho autos anthropos olisthesas semeron kata
("The same man slipping today down. . . ")
But if you delete the adverb, the sentence is not made deficient at all:
(4) ho autos anthropos olisthesas katepesen
("The same man, slipping, fell down.")
And if you take away the participle, too:
(5) ho autos anthropos katepesen
("The same man fell down.")
there is no defect. T h e same holds for the preposition:
(6) ho autos anthropos epesen
("The same man fell.")
and for the p r o n o u n ; for then what remains is:
15. I am not here claiming that you can't have a complete sentence with p r o n o m
inal subjects, such as:
1.14 ego peripato ("I'm walking.")
su peripateis ("You're walking.")
[[Here, too, the ego "I" and su "you" are normally present only if emphatic; peripato
is the ordinary way of saying "I'm walking".]]
For then, too, completeness is achieved, when a p r o n o u n is used in place of a
noun, which gives an essentially identical structure. T h e question when a p r o n o u n
may be substituted will be taken u p in the appropriate sections [2.43], as well as the
question which verbs occur only with nominative [nouns], and which ones require
oblique cases [as well] [3.148,158].
16. T h e n o u n necessarily precedes the verb, since influencing and being influ
enced are properties of physical things, and things are what nouns apply to, and to
things belong the special features of verbs, namely doing and experiencing. [[I.e.
physical objects exist before any events or experiences which they cause or under-
go.]]
17. T h e nominative [subject] is implicitly present in [finite] verbs, and it is
definite (i.e. has definite reference) in the first and second persons, but indefinite
in the third because of the unlimited n u m b e r of possible referents, except in the
case of such unique actions as astraptei ("it is lightening") and brontdi ("it is t h u n d e r
ing')• [[Or perhaps we should say "he is thundering", i.e. Zeus, as explained in
2.17.]]
18. So the term onoma ("noun" or "word"), because of its primacy, acquired
general applicability, so that all parts of speech may be called onomata. For if
anyone should want to claim the opposite, that onoma became the word for nouns
as a result of having been the general name for all words, this argument, too would
give it first prize, assigning it first place a m o n g the parts of speech, just as we show 13
how, because of the invention of the alphabet, all the letters might have been called
by the name "alpha" from alphein ("to be profitable"; some grammarians say it
means "to invent"), but A took the name of all the letters to itself because of its
initial position, and the coincidence of the initial sound of the word. T h e initial
sound of alphein coincided with the pronunciation of the letter, [a], and all the
other letters acquired particular extensions [to make their names: e.g. b- gets the
extension -eta, to give the name beta.]
19. Someone could, I think, quite reasonably object "Why on earth shouldn't the
next place after the n o u n be filled by the one word-type which may replace it
syntactically, i.e. the p r o n o u n , since it, too, by substitution for a noun, may
combine with a verb to make the sentence complete?" O n this matter the clearest
evidence is this, that p r o n o u n s were invented for the purpose of accompanying
verbs. [Therefore verbs must preexist.] Since nouns indicate third persons, apply
ing them in a sentence can neither serve for the person speaking, i.e. the first
person —for children do not assign names to themselves — nor do we assign
names in addressing them — and the person addressed is second person. But
verbs are concerned with three persons, and therefore p r o n o u n s were invented to
refer to definite persons [i.e. persons specified by the speech act] thus taking the 14
26 THE SYNTAX OF APOLLONIUS DYSCOLUS
place of nouns and completing the syntax of the verb, e.g. ego grapho, su grapheis ("I
write, vou write"). And so there is nothing to block saying, in the third person,
Aristarchos anaginoskei ("Aristarchos is reading"), which is a collocation of two
third-person forms. And I'm not trying to suggest that third-person p r o n o u n s are
unnecessary because nouns are always available to accompany third-person verbs.
T h e reason why there are third person p r o n o u n s will be revealed in the a p p r o
priate sections [2.45]. And if this is correct, it is obvious that the placement of verbs
must come ahead of pronouns, which were invented for them.
20. T h e strongest proof is that [finite] verbs are inherently marked for the
persons understood in the nominative, when thev are used absolutely. For if
anyone savs peripato ("I'm walking") or grapho ("Fm writing"), he is not indicating a
contrast with some other person [i.e., he's not saving "I' m walking" etc.]. Pro
nouns, which were devised for contrast of persons, occur in the oblique cases also
as enclitics, which represent persons absolutely [i.e. are not contrastive, not accen
ted], but when they are introduced in the nominative case, they are always
emphatic and contrastive. So, for instance, emoi ("me") in:
1.15 paida d'emoi lusaite (Iliad, 1.20)
("But for me, free my daughter")
is in contrast with humin ("you") in:
1.16 humin men theoi doien (Iliad, 1.18)
("May the gods grant you. . . ")
and erne ("me") in:
1.17 he d'eme cheiros helousa {Odyssey 12.33) 15
("And she took me by the h a n d . . . ")
with hetairon ("companions"):
1.18 phildn apaneuthen hetairon {Odyssey 12.33)
("away from my companions").
Now how does it h a p p e n that nominative p r o n o u n s cannot occur in enclitic form?
Obviously it is because [finite] verbs preexisted and contained in themselves the
absolute [i.e. unemphatic] person, so that enclitic nominative p r o n o u n s could
contribute nothing useful to the sense, and would be more genuinely r e d u n d a n t
than the so-called "expletive conjunctions" [emphatic particles, mainly, like ge or
de], which we will show [3.127] are often required [for the sense]. (When the Doric
nominative second-person pronoun tu became enclitic, it necessarily destroyed its
own nominative character, changing to the minimally different accusative.) And it
is evident that personal pronouns of the nominative case became established in use
for no other reason than that verbs, not being able to represent contrast [of
subject], adopted the p r o n o u n for the sake of its characteristic feature, namely
contrast.
21. And the participle, in turn, took its necessary place immediately after the
verb, showing by its position its direct transformation from the verb (as we have
explained at greater length in our treatise "On the Participle", it was necessary to
be able to transform verbs into forms which could be inflected for case and could
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 27
agree in gender, since [finite] verbs could not enter into such a grammatical
construction).
22. It is also clear that the etymological application of the n a m e "participle"
["sharing" — i.e. sharing features of both n o u n a n d verb] could not otherwise be 16
kept in view unless it were introduced after both n o u n and verb h a d been, since
this part of speech depends on its acceptance of some of their features, just as the
neuter ["neither", i.e. neither masculine nor feminine] must come after masculine
and feminine, as a negation of both of them. For if we did not accept the prior
o r d e r i n g of these categories (noun a n d verb in one case, masculine a n d feminine
in the other), we would be left without a basis for calling one "participle" a n d the
other "neuter"; of what two preestablished entities could the neuter be a denial [as
"neither of the two"]? A n d what ones would the ("sharer") participle share?
Neither would it be possible to insert any other part of speech — p r o n o u n , adverb,
conjunction, anything else — before the participle. For it (the participle) would
not have any share in their unique features.
23. T h e article didn't just get tossed in next by chance, either, but reflects the fact
that it is regularly conjoined with the preceding nominals [noun a n d participle],
but avoids the [personal] p r o n o u n , which cannot take the article; indeed the
article may even be combined with verbs [infinitives], as in:
1.19 to philologein kalon estin
("The study of literature is good")
1.20 tōi graphein hedomai
("I enjoy writing.")
24. T h e placing of the p r o n o u n next won't require any consideration, since it
barely missed being placed right after the n o u n . It's obvious that what can be
substituted for something naturally accepts a later place [than what combines with
that something]. A n d since the article is used with a n o u n a n d the p r o n o u n
substitutes for a n o u n , it must be granted that the article, which co-occurs with 17
nouns, outranks the p r o n o u n .
2 5 . Perhaps also we might consider thp fact that referential p r o n o u n s are equiva
lent to n o u n s combined with articles [i.e. to definite NP's, not just N's]. For n o u n s
by themselves are not referential, b u t only when they are accompanied by the
article, whose special function is reference (anaphora). So functionally these
p r o n o u n s are equivalent [in part] to articles. Incidentally, the article itself, when it
lacks its normal association with a n o u n , changes into the next lower r a n k i n g part
of speech, the p r o n o u n , as in the following examples:
1.21 ho gar elthe thods epi neds Achaion(Iliad 1.12)
("For he came to the ships of the Achaeans.")
1.22 hos gar ra malistalhendane kerukon (Odyssey 17.172-3)
("For he was the best-liked of the heralds").
1.23 ton d'apameibomenos (Iliad, 1.84, etc)
("And answering him, he. . . ).
28 THE SYNTAX OF APOLLONIUS DYSCOLUS
[[Note that the term "article" includes both what we call "definite article" and
"relative p r o n o u n " , and in H o m e r both forms are also used as anaphoric third-
person pronouns.]] We will explain the details of this construction fully in the
appropriate place [2.28].
26. It is clear that the preposition also does not rank first or ahead of any of the
parts of speech we've already discussed, and hence is placed here. T h e n a m e
"preposition" ("placed before") was not given to it because of its own inherent
meaning, but because it is preposed [in speech] to the previous word-types (morion)
[noun, verb, participle, article, p r o n o u n ] , so that if they were not preestablished, it
could n o t be combined with them as we have explained above [ch. 22]. From this,
then, it gets its position, being preposed sometimes in composition [e.g. before
verbs], sometimes in construction [before the other parts] with the aforesaid parts
of speech, so that is is in rank later by nature, but initial in a structure (as can be said 18
also about the so-called preposed article [i.e. the definite article]; for it is not the
case that it ranks higher than nouns because it is preposed to them syntactically.
Note that when the article is combined with a n o u n , it functions to recall some
prior knowledge of that n o u n [i.e. is anaphoric].)
27. A n d since the adverb (ad-verb) has a structural position a d d e d to [i.e., after]
the verb, just as the placement inherent in its n a m e testifies, and since the verb is
second to the n o u n , so also the adverb is second to the preposition, which in
composition and construction is placed before [e.g.] nouns.
28. After all the parts that have been listed we take the conjunction, which
conjoins them, and cannot convey any m e a n i n g by itself without the substance of
these words, just as physical bonds are no use if there are no physical objects [to
connect].
29. T h e r e are still m o r e a r g u m e n t s which could be used to demonstrate the
o r d e r [of the parts of speech], but since o u r proposed topic is not that, let's
consider this discussion sufficient.
30. There's one m o r e point to be established before we take u p the details of the
constructions of each of the parts of speech: why it is that interrogative words are
found in only two parts of speech, nominal and adverbial, and why there is not just
one nominal interrogative and one adverbial, but several, e.g. tis ("who"), ti
("what"), poios ("which"), posos ("how much"), postos ("the what'th"),pelikos ("how
big"),podapos("what nationality") on the one h a n d and pos ("how"), pote or penika
("when"), pou ("where"),pei ("in what way"),pothen ("whence"), on the other [[all 19
these, a n d others as well, are single words in Greek and Latin, so Priscian can
duplicate the list. This is related to Bloomfield's problem of "class cleavage"
(1933), 258-270.]] Is this not good proof that the most vital parts of the sentence
are the first two, n o u n and verb, which naturally, when they are in doubt,
frequently have questions asked about them? [[Since answers to the adverbial
interrogatives are adverbial expressions, not verbs, the a r g u m e n t seems a little
compressed.]] This is the reason for the many nominal and adverbial interroga
tives.
motion, walking or speech, but the identity of the actor (energon) is unclear. T o this
we find nominal answers, common n o u n s or p r o p e r nouns — and p r o p e r nouns
also convey the generic essence. T h e answer may be "A man is walking" or "A
horse" or " T r y p h o n " ; and if you say " T r y p h o n " , you also signify "A man". Or we
may have a word which is used instead of a noun, specifically instead of a p r o p e r
noun, namely [a personal p r o n o u n ] , e.g. "I [am walking]."
32. And since all the attributes or qualities are not m a d e clear by the suggested
nominal answers (for by itself the word "who" inquires only about the essence
(ousia), to which quality and quantity apply), the additional device was invented of
interrogation about these matters, so that when we ask about quality we say poios
("what sort?"), a n d about quantity posos ("how much?") [or posoi "how many?"], a n d
about size pelikos ("how big?"), and to elicit a word derived with the nationality
suffix [e.g. "Athenian"] we ask podapos ("where from?"). With the same suffix is
formed hemedapos " O u r fellow-townsman", with personal p r o n o u n base, and also
humedapos "your compatriot", and similarly from alios "other", we get allodapos
"foreigner," in which alios with the nationality suffix signifies the negation of [both
first and second] personal p r o n o u n s [i.e. a foreigner is neither my fellow-citizen
nor yours]. A n d so the response to a question poios ("what kind?") which has been
preceded by a "who?" [question and its answer], is adjectival information, e.g. "the
grammarian", "the musician", "the r u n n e r " [all these words are adjectival in form
in Greek], and the dialogue might go like this:
35. Interrogative adverbs are used in cases where features of the event are not
known, either the quality of the act, as when we say "How did he read?" to which
we may get an adjectival (i.e. manner) adverb in response, e.g. "Beautifully",
"Rhetorically", "Philosophically", or when we're not asking about that, but about
the time when the event took place, e.g.pote? ("When?") penika? ("At what time?"),
to which we may answer "Yesterday", "Day before yesterday", "Long ago", or
about the place in which the event occurred, e.g. (Iliad 10.406):
1.26 pou nun deuro kion lipes Hektora
("Where did you leave Hector when you came here?")
or about the difference of route in getting from place to place, as in (Iliad 6.377):
1.27 pei ebe Andromache
("Which way (i.e. in what direction) did A n d r o m a c h e go?")
[or the place from which]:
1.28 pothen paregenou
("Where did you come here from?").
But [in Odyssey 11.57] Odysseus knows where Elpenor came from, and so doesn't
ask him that, nor does he ask him about his destination [which he also knows,] but:
1.29 pos elthes 22
"How did you come?"
Why these interrogative words (morion) are all either adverbial or nominal has now
been explained. Later on [Ch. 119] we will discuss their constructions.
36. Since the rest of the parts of speech are used in construction with verbs and
nouns, from which fact they have received their names, we must now discuss each
of those which are combined [with these - e.g. article, preposition, adverb, con
junction] and those which are both substituted and combined, such as p r o n o u n s ,
which are either used instead of nouns or with nouns, and likewise participles,
used instead of verbs, and with verbs, and after these, the remaining parts of
speech.
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 31
37. (IV) Articles are used in n o u n constructions and also in certain verb con
structions, since it is possible to combine them with infinitives:
1.30 to philosophein ophelimon ("The philosophizing is useful".)
1.31 toi peripatein hedomai ("I enjoy (the) walking".)
and furthermore the article may be used with any part of speech when it signifies
nothing more than the name of the word [i.e. what is called "mention" or "hypos
tasis"], where the article really applies to the n o u n understood from the context:
1.32 ho "men" protaktikos esti tou "de" ("men is placed before "de")
H e r e the word sundesmos ("conjunction') is understood [i.e. ho men sundesmos, "the
conjunction men", tou de sundesmou," of the conjunction de."]
1.33 to "lege" prostaktikon esti
{"Lege" is imperative)
H e r e we understand rhema "verb" [to "lege" rhema, "the verb 'lege'"]. A n d similarly
with nominals (ptotikon) [nouns, participles and adjectives, mainly]; when we say:
1.34 ho Aristarchos ("Aristarchus")
we are using the (masculine) article with reference to the gender of the referent,
but when we say
1.35 to "Aristarchoi" proparoxunetai
("Aristarchoi" is accented on the antepenult)
1.36 to "Aristarchoi" eis oi legei 23
("Aristarchoi" ends in -oi)
we are referring to the name of the word, u n d e r s t a n d i n g onoma("noun"); and
similarly one may say:
1.37 he semeron, he chthes
("today", "yesterday")
[meaning he semeron hemera, he chthes hemera, with hemera "day" understood. But
these examples are really use, not mention.] So such articles are always singular,
for every word is just one, even though it may represent a plurality. So the word
anthropoi ("men") is one word, considered just as a word, but represents a plural
sense and referent. And when people say that p r o n o u n s do not take articles, they
are not referring to their use as names for themselves, but to their p r o p e r use to
convey their inherent meaning, which is representation of persons with deixis [i.e.
they point to the speaker, the hearer, etc.]. But when they are used hypostatically,
they take the article [feminine, to agree with antonumia, " p r o n o u n " ] :
1.38 he ego monon orthotoneitai
("Ego takes only the acute on the ultima.")
1.39 he soi enklinetai
("soi is enclitic")
32 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
38. (V) Some grammarians have fallen into the serious error of supposing that
the function of articles is to distinguish the gender of the nouns they are used with.
I have given a detailed reply to them in my treatise On Genders. But here, too, since
the discussion of syntax requires it, we will set out a few arguments to show how
utterly false their opinion is.
39. In the first place, no part of speech is designed to resolve the ambiguity of
another part of speech, but each is used for its own p r o p e r meaning, as we will
show in what follows. In fact, ambiguity of gender is resolved by the accompanying
parts of the sentence [which may be articles, adjectives, participles, p r o n o u n s ] , just 24
as any word with multiple meanings in the same form has its ambiguity resolved by
the prior context. T h e word sophron, without any article, has its ambiguity resolved
by the context in:
1.40 sophron ousa Helene herpage hupo Alexandrou.
("Helen was virtuous when she was abducted by Alexander.")
[[Here both the participle 'ousa and the n a m e 'Helene' give the gender.]]
However, in the sentence:
1.41 ek ton theon epeklosthe Odussei to me thanein kata thalassan.
("Odysseus was destined by the divinities not to die at sea.")
Even though the article {ton) is present, nevertheless it does not suffice to eliminate
the ambiguity of gender, whether the gods are meant or the Fates. [[The genitive
plurals ton and theon can be either masculine or feminine.]] Obviously, then, it's
ridiculous to say that the article is used to distinguish gender, when (a) gender is
ambiguous, even though the article is present, and (b) the ambiguity is resolved,
even though the article is not used.
40. In the second place, it hardly seems right for the word which was introduced
for the distinction of gender to fall itself into the same difficulties for which it was
invented, namely confusion of gender, since ton (gen. pl. article) and toin (gen. or
dat. dual) and some others like it are not limited to one gender. [[In fact all dative
or genitive neuter forms are the same as masculine forms; the only other form
besides ton and toin which is the same for all three genders is to, the nom.-acc.
dual.]] And if we should grant that ō [the particle used with vocatives] is an article,
since it is the same for all genders and n u m b e r s , still other "articles" will have to be
invented to distinguish these forms. Indeed, since often an ambiguity of gender
[in the article] is resolved by the accompanying nouns, e.g. ton Mouson ("of the 25
Muses"), there's no reason why we shouldn't claim that nouns were invented to
discriminate the gender of articles, which would be stupid. It is a p u r e coincidence
that articles sometimes resolve ambiguities of gender.
4 1 . Thirdly, we can add this, that [if the article's function is to discriminate
gender], articles should only be added in phrases where there is real ambiguity of
gender, not in those which are able to show their gender unassisted, e.g. [it should
always occur with] theos ("divinity"), since we say ho theos ("the god") and he theos
("the goddess"); similarly ho hippos ("the stallion"), he hippos ("the mare"), but [it
should] not be added to gune ("woman"), which semantically contains what might
be regarded as the essence of the feminine gender. But, in fact, it will occur with
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 33
gune in some sentences where the syntax requires it, and in other contexts, where
theos or hippos or similar words occur, the article cannot possibly be a d d e d . Let's
consider examples, first of the first case, e.g.:
1.42 pos he gune se hubrise?
("How did your wife [or "the woman"] h a r m you?")
H e r e if the article is removed the sentence becomes incomprehensible [or else
poetical, a possibility which he should not ignore]. T h e n an example of the second
case:
1.43 theos tis se eleesen
("Some god took pity on you.")
H e r e it is impossible for the article to be added (*ho theos tis = *the some god), as we
observed in On Genders with regard to Homer Iliad 8.7:
1.44 me te tis oun theleia theos to ge me te tis arsen
("Let no female devinity [try to disobey] nor any male, either.")
Where H o m e r explicitly added theleia ("female") [and arsen "male"]. T h e prohibi
tion was directly mainly against Hera and Athena, but the me te tis arsen was a d d e d
so that he would not seem to be opposing these goddesses specifically. And it's
clear that where the article, if present, would have distinguished gender inciden
tally, h e r e H o m e r used the words "male" and "female", because the construction,
being an indefinite NP [with tis "any", an indefinite] would not allow the definite 26
article.
42. Perhaps someone may say: "Well, isn't it the case anyway that this sentence
has u n d e r g o n e deletion (ellipsis) of the articles?" T o him we will reply that deleted
items must be present potentially [i.e. recoverable], clearly conveyed by the sen
tence which calls for them. But in the passage in question the genders are confused
even though the grammatical structure does not call for [articles], and this is
precisely why (I think) H o m e r completed the sentence with explicit mention of
genders, namely "female" and "male". [[Though a little unclear, this passage
seems to imply a theory of recoverability of deletion.]]
4 3 . (VI) T h e special function of the article, as we have explained elsewhere, is
anaphora (reference), that is, the presentation of a person [i.e. the referent of an
NP] previously identified [i.e. known to speaker and hearer]. A n a p h o r a of nouns
may be (1) par excellence, as when we say:
1.45 houtos estin ho grammatikos
("This is the grammarian.")
indicating by this " T h e one who stands above all", " T h e superlative g r a m m a r i a n "
so to speak. In this way the poet [= H o m e r ] got the article as a permanently
prefixed syllable, thus winning recognition of rank above all other [poets], and
recognition of prior knowledge on the part of all [i.e. everyone knows who H o m e r
is]. Or else (2) it marks unique possession. For if one says:
1.46 doulos sou tauta epoiese
("A slave of yours did this."),
34 THE SYNTAX OF APOLLONIUS DYSCOLUS
one indicates that you have several slaves, but if it is said with the article:
1.47 ho doulos sou tauta epoiese
('Your slave did this.")
unique possession is implied. [[Of course Apollonius should have noticed that this
applies to other modifiers besides possessive genitives—e.g. adjectives—and even,
in a sense, to unmodified nouns; nor did he note that the same distinction applies
in the plural—"slaves of yours" vs. "your slaves."]] O r else (3) it may be mere
anaphora, as when we say:
1.48 ho anthropos ēlthe se zeton
("The man came looking for you.")
or
1.49 ho grammatikos se ezetei 27
("The grammarian was looking for you.")
understanding "the grammarian" now not in the par excellence sense, but just as
already mentioned.
44. (4) Sometimes also the anaphora is anticipatory, when the phrase is used
indefinitely (i.e. generically), as when we say:
1.50 ho turannoktonesas timastho
("Let the tyrannicide be honored.")
H e r e the person referred to is potential, just as in Odyssey 6.158-9:
1.51 keinos gar peri keri makartatos exochos allon
hos ke se hednoisin brisas oikon d'agagetai
("For that man will be far happiest of all men, beyond fate, who
loads you with bride gifts and brings you home.")
H e r e of course, it is keinos "that man" which is anaphoric. This is not similar to
Odyssey 4.145:
1.52 keinos aner, hot'emeio kunopidos
("that man, who for my [sake] shameless that I am")
since here keinos refers back to a previously mentioned (4.143) person, whereas in
6.158 it refers to one who does not yet exist.) I am quite aware that it is also possible
to refer to an already identified person in:
1.53 ho turannoktonesas timastho
("Let the one who kills/killed the tyrant be honored.")
Suppose, for instance, that someone has killed a tryant but has not yet been
honored, then, referring to such a person, someone could say: "Let the tyran
nicide be honored." This use will become most obvious with a verb in the indicative
mood, particularly with a past tense verb:
1.54 ho turannoktonesas tetimetai
("The tyrannicide has been honored.")
(5) Finally we will show [in Ch. 111] that the article sometimes suggests a plurality.
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 35
45 (VII) In the light of the above ideas about the definite article, we must
consider (a) whether all nominals (case-inflected words) may take it, (b) if not,
which of them do not allow construction with the article, (c) and which ones not 28
only accept the article but keep it permanently, like an inseparable prefix syllable.
A n d first let us begin with the construction of names of letters of the alphabet with
the article.
46. Names of letters are used in nominative and accusative both with and
without the article. (a) Without the article when we say:
1.55 touto alpha esti, touto beta esti
("This is (an) A, that is (a) B.")
where they are understood as nominative, in analogy with sentences like:
J
T.62 ton charaktera apeleipse [ho pais]
("The boy erased the letter.")
47. In the genitive and dative, however, it's impossible to omit the article in
construction with a letter, as in:
1.63 toi alpha parepetai to ekteinesthai kai sustellesthai
("It is a property of (the) A to be lengthened and shortened.")
[of (the) A = toi A]
36 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
52. Besides, the infinitives that he [sc. T r y p h o n ] says are used without articles,
can also occur with articles, e.g.:
1.85 to philologein boulomai eper to ploutein
("I'd rather be a scholar than be rich." or
"I prefer [the] being a scholar to [the] being rich.")
Evidently, then, the rules for using the article with infinitives are the same as those
with letters of the alphabet. So we must conclude that there are two types of
articular infinitive, (1) one when the word is being mentioned, as we have shown is
the case with every part of speech, and (2) the other when it is being used to refer to
an event. Examples of the first type are:
1.86 to graphein protes esti suzugias
("Graphein belongs to the first conjugation.")
1.87 to graphein dia tes ei diphthongon graphetai
("Graphein is spelled with the d i p h t h o n g EL")
[[Remember, in Apollonius' day EI was p r o n o u n c e d identically the same a s / . ] ]
and other similar cases which refer to the morphology or phonology. Examples of
the second type are:
1.88 to graphein epiponon esti
("Writing is a nuisance.")
1.89 to graphein philoponou andros estin
("Writing is a j o b for an industrious man.")
53. (IX) It isn't a waste of time, either, to show that in such constructions the 33
article is not an adverb when it is used before [infinitive] verb forms. Since the
a r g u m e n t that has been offered to show that such articles are not to be interpreted
as articles, but as adverbs, is not implausible, we'd better discuss the point a little.
54. [Someone might argue as follows:] "When a neuter adjective form is placed
before a verb, it is more often taken adverbially than nominally (=adjectivally),
e.g. tachu paregenou ("You arrived quick(ly)."):
1.90 hos t'euru rheei Pulion dia gaies (Homer, Iliad 5.545)
("Which flows wide(ly) t h r o u g h the land of the Pylians")
1.91 ou men kalon atembein oude dikaion (Od. 20.294)
("It is not nice or right to mistreat [Telemachus' guests].")
[[This doesn't strike us as an adverbial use of kalon, but Apollonius so treats it in
other passages as well. How he interpreted the line is not obvious; if it were atembete
it could m e a n , p e r h a p s , "you are unpleasantly a n d unjustly m i s t r e a t i n g
[Telemachus' guests]", but that is h a r d to do with an infinitive, unless perhaps he
took it to d e p e n d on keklute mou in 292—"Let me tell you that you are etc."]]
1.92 tachion dialegou
("Talk faster.")
[[In the comparative, as in Latin, the neuter singular form is the regular adverbial
form.]] T h e same analysis applies to articles, which have three genders, just like
adjectives. So if:
40 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
1.117 holloi men rha theoi te etc.- [holloi by crasis for hoi alloi]
("The other gods etc.")
But it's clear that where Zenodotos couldn't emend, we can't either, e.g. Iliad
1.295:
1.118 alloisin de taut' epitelleo
("Give these orders to the others.")
[where the article would have to be in the form toisin, which won't fit into the line].
But let's d r o p this matter, since it is more time-wasting than instructive about
linguistic tradition. [[We might rather say that in Homeric Greek the article was
not required in combination with alloi, in spite of the rule in classical Attic.
T h e r e f o r e its absence in H o m e r would not be ellipsis for us. But for A.D. Homer's
underlying structure was almost identical with Attic surface structure.]]
63. T h e article will always be used with alloi [in classical, not Homeric Greek]
when it is inclusive of the totality of what is referred to and makes a division [of that
totality]. For when this is not the case, then the article will not be used at all. Let's
take as examples sentences about dialect pronunciation:
1.119 hoi men alloi Hellenes dasunousi ta en tei lexei phoneenta, Aioleis de
monon psilousi 39
("The other Greeks [all] use h's on [some initial] vowels, but
Aeolians p r o n o u n c e [the same words] only without h.")
For the Aeolians are a subdivision of the Greeks. O r someone might assert:
1.120 hoi men alloi Dorieis terousi to theta, Lakones de kai eis sigma metahallousi
("The other Dorians [all] preserve [general Greek] T H , but the
Spartans even change it to S.")
Again the group is inclusive, and the Dorians are subdivided [into Spartans and
the rest]. T h e same line of reasoning as used here will apply in all similar cases.
But on the other h a n d [without the article], an example is
1.121 allois anthropois homilei, me tois peri Truphona
("Speak to other men, but not to [associates of] Tryphon.")
For the "others" here do not include all mankind.
1.122 alloi men gar kai charizontai hemin, su de kai prosaphairei
("For others even give presents to us, but you take away what we
have.")
1.123 allous hubrize, me hemas
("Mistreat others, not us.")
There's no need here to add an article, where the expression is normal and
acceptable. But (e.g.) the first example becomes unintelligible [or rather, it
changes its meaning] if we say it without the article:
"Other Greeks use h's, Aeolians never."
64. By this test we can distinguish those cases in the poets [quoted in 62]. Where
the word "others" is inclusive of all gods and men, and the subdivision excludes
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 45
Zeus, the article will be required [[in other words, we explain that there has been
ellipsis of the article.]] But in the case of Iliad 23.459 [see 62, 1.116] ("Different
horses seem to be in the lead"), this is not so. Likewise in Iliad 1.295 [see 62, 1.111]
("Give these orders to others") there is no reason at all to supply the article. For it's
as much as to say eniois epitasse ("order some people"). As in everyday use we say: 40
1.124 allois proserchou, me emoi
("Apply to others, not me.")
or
1.125 allous hubrize, me erne
("Mistreat others, not me.")
and do not thereby give orders to mistreat everyone in existence, but only some
appropriate people. In the example from Iliad 15.87 [see 62, 1.115] ("And she
ignored the others, but took a cup from fair-cheeked Themis") there is ellipsis of
the article, since "the others" includes all the gods who were feasting together. But
there is none inlliad 7.473 [see 62, 1.113] ("Some with bronze, some with gleaming
iron"), since here there is no all-inclusive set being divided, nor does that "others
with gleaming iron" subdivide the set; the construction in fact, implies a sort of
rough equality in n u m b e r . T h e article is required in 77. 5.877 [see 62, 1.114] ("all
the other gods who are on Olympus"), like this:
1.126 hoi men gar alloi theoipantes hosoi eisin en Olympoi, pantes soi epipeithon-
tai
("All the other gods who are on Olympus, all of them obey
you . . ."
[This is a prose "underlying form" for 62.114]
because of the construction with hosoi, inasmuch as he [Zeus] had established his
s u p r e m e power.
(XXI) 65. Next, every dative of an adjective modifying the name of the addressee
on the superscription of a letter requires the article, e.g.:
1.127 Dionusios Truphoni tōi agathotatoi chairein, tbi timiotatōi 41
("Dionysius, to the most excellent, most honorable T r y p h o n ,
greetings:")
Without the article the structure is meaningless [["asuneton"; literally this means
"incomprehensible," but A.D. seems to use it merely to mean "unacceptable" or
"ungrammatical", much as contemporary linguists use the asterisk—and no more
accurately or consistently.]] This regular adjunction of the article has the follow
ing explanation. Personal names may be h o m o n y m o u s [different people may have
the same name], and the ambiguity cannot be resolved without the addition of
some qualifier, e.g. we say
" T r y p h o n the grammarian,"
"Dion the philosopher,"
or else use a nationality word—
"Apollodorus the Athenian."
"Apollodorus the Cyrenean,".
46 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
In this way p r o p e r names which have reference to several persons may have their
ambiguity resolved. T h e same thing is involved in the construction we are discus
sing. (Furthermore, for this reason, when Menelaus is s u m m o n i n g Ajax, he uses
two adjectives to distinguish him from his namesake, in Iliad 12.349:
1.128 alia per oios itō Telamonios alkimos Aias
("But let brave Telamonian Ajax come alone.")
and, in another place (Il. 16.358):
1.129 Aias d' ho megas aien eph'Hektori
('Ajax the great was always [aiming spears] at Hector.")
Obviously this construction [at the opening of a letter] requires more lauda
tory adjectives because of the affectionate nature of the accompanying verb
(chairein"rejoice"), whose construction [infinitive] we will discuss later, as well as
the whole context.
F u r t h e r m o r e , it is a p p a r e n t that adjectives may refer to many different
individuals. Therefore, in o r d e r to remove such [uncertainty] from the construc
tion, the article has to be added, indicating (as regularly) unique reference, just as 42
we have shown earlier.
66. O n e can also argue as follows: articles are sometimes added to indicate the
unique specification [of reference] of a n o u n , e.g.
1.130 ek Delou to ploion eis Athenas paragenomenon katepleusen
("The boat arrived at Athens from Delos and put into the harbor.")
H e r e the use of the article in to ploion katepleuse ("the boat arrived") [indicates
known identity], just as the expression ho poietes ("the poet") specifies H o m e r .
Necessarily, then, in such cases [i.e. prescripts of letters,] the addition of the article
joins the adjective to particularize it along with the p r o p e r name.
67. (XIII) T h e following construction illustrates the same point:
1.131 Basileus Makedonon Philippos Athenaibn tei boulei kai toi demoi chai-
rein.
("Philip, King of the Macedonians, to the senate and the people of
the Athenians, greetings")
T o paraphrase the reference, this is equivalent to "Greetings to the Athenians."
And since the same attributes didn't apply to all equally (for some of the Athenians
were "citizens" and the rest were "senators"), and this epistolary construction
requires a distinguishing adjective [[for A.D. "citizen" and "senator" count as
adjectives]] Philip divided [the Athenians] into the senators, for whom he used the
word boule "senate," a n d the citizens, for whom he used the word demos "people,"
so that the whole expression ("the senate a n d people of the Athenians") is equiva
lent to "the senators and the citizens". And it is clear that collective nouns have this
collective function—nouns, that is, which are singular in form but plural in
meaning. And this makes it easy to explain Iliad 20.166
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 47
77. But if anyone doesn't want to grant this, then, as there are many words which
are irregular, but nevertheless don't deviate in sense from their form-class mean- 48
50 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
ing, and.so are counted as p r o p e r members of the form-class (e.g. ego " I " differs
entirely from noi ("we two") as well as from hemeis ("we"), yet, since their sense [as
personal pronouns of the 1st person] remains, they keep their identity as members
of the same form-class), this much, at least, is clearly true, that forms which deviate
from the sense distinctive of their form-class, even if their regularity of form is
perfect to the point of identity [so that they are, in fact, h o m o p h o n e s ] , will not be
counted as the same part of speech {merismos) [and hence, not the same "word" or
lexeme]. T h u s even more emphatically will 5 be excluded from reckoning as a
form of the article, lacking, as it does, any similarity of stem-form or inflection,
and lacking, in addition, the meaning inherent in the article. We can summarize
the argument, then, as follows: parts of speech will neither be established because
of regularity of form, nor because of irregularity but, as has been explained, on
the basis of constant unique distinctive properties [of use or sense].
80. (XX) It will be worth our while to consider also the inflection of the article,
the other forms, that is, in contrast to ō. T h e nominative [sg. masc] besides
d r o p p i n g its t, undergoes another change, the omission of -s [[i.e. the ideal
underlying form is *tos, the surface form is (h)o—A.D. here disregards the h-,
which was no longer pronounced and was not written with a letter]]. T h e reason
for d r o p p i n g -s is to prevent it from being identical with the relative pronoun
("postpositive article"), whose occurrence restrictions are not the same as those of
the article ("pre-positive article"), as we will show in the section on relatives, and
therefore would make the sentence seem ungrammatical.
Now there are four clear arguments to show that -s has been d r o p p e d (i.e. is
underlyingly present): (1) in all other respects the relative p r o n o u n has exactly the 50
same endings as the article, genitive relative hou—article tou [dative] hōi-tōi, [ac
cusative] hon-ton. (2) Also consider the plural [nominative]: all other plurals in-oi
have singulars in -os, even in contract adjectives, like chrusous [i.e. nom. sg. masc.
chrusous, nom. pl. masc. chrusoi, have underlying forms chruseos and chruseoi]; (3)
also clear is the evidence of the oblique cases, since all other words with those
endings [in genitive, dative, accusative] have nominatives in-os (and the similarity
extends to the so-called Thessalian distracted genitive, [article] toio,likt [the adjec
tive] kaloio ("fair, noble" gen. sg. masc. or neuter); (4) formally from the feminine;
for all words with (nom. sg.) feminine in -e, have [nom. sg. masc] in -os.
8 1 . (XXI) [The underlying presence of initial t- is shown by several things.] (1)
T h e feminine article wouldn't end in -e if it hadn't had the t-, just as noete (n.s. fern,
"mental") comes from noetos. [[This seems peculiar, since the relative pronoun,
which he's just contrasted, has he (n.s. fern.) from hos. Schneider thinks he's
revising the rule "-a after vowels and -r-, otherwise -6?" to read "-e after consonants,
otherwise -a" but A.D. surely knows better than that.]] (2) And note that the Doric
dialect replaces the missing t- in the nominative plural, m. and f. toi anthropoi ("the
men"), tax gunaikes ("the women"), quite naturally. For this dialect is full of t's
[where Attic, Ionic and Aeolic have s's], e.g.proti (pros, "to"), legonti (legousi," they
say"), phanti (phasi, "they say"), peruti (perusi,"last year"), tu (su, "thou") etc., etc.
And if, in these cases, the usage is d u e to change [of s into t], then it's even more
likely that they would restore a deleted [t]. (3) T h e initial h- [of ho, he, hoi, hai] also
proves the point; for the article wouldn't have h- otherwise, if it hadn't lost t-.
[[This, also, is a puzzler. T h e ancients know of pairs alternating between 5- and h-,
like suslhus "pig", but the only other word resembling the article on this point is
houtos "this," neuter touto. And surely there are many h's for which no underlying t
is assumed.]] (4) And, finally, the neuter offers convincing evidence, since it
[nom-acc. sg. neuter] begins with t- and ends with -o, unlike the pattern of words
whose masculine terminates in -os. [[The point of this is not very clear, either.
Non-pronominal neuters end in -on.]]
82. Evidently it was for the clarification of ambiguity of gender in the accusative
[singular] that this change [omission of final -n] occurred. [[In the accusative
singular, the masculine is ton, and the analogically expected neuter would be ton;
the form to, then, eliminates a possible ambiguity.]] And the p r o n o u n s which 51
cannot be preceded by the article [ which might otherwise eliminate the ambigu
ity], adopted the same neuter ending as the article: ekeino ("that") touto ("this,
that"), and the p r o n o u n autos, [which can be preceded by the article—though only
52 THE SYNTAX OF APOLLONIUS DYSCOLUS
in the meaning "same"] followed their analogy, auto ("it"). Attic follows a m o r e
rational pattern, then, when it uses this p r o n o u n with the article [to auto "the same
thing"], by restoring the final -n whose omission would now be r e d u n d a n t : tauton
("the same thing").
83. Now we can say that every a r g u m e n t which tests the regular [inflectional]
pattern of the article is also an a r g u m e n t against the form ō [being an article]. (1) If
the inflection of the article obeys the same general laws as the inflection of nouns,
and the vocative is never metrically longer than the nominative, how does it
h a p p e n that ō is longer than ho? (2) [If this is the vocative of the article], why
doesn't it end in -e, as is the case with all other words with genitive in -ou and
accusative in -on? (3) In the feminine and the neuter why isn't it identical with the
nominative? [As is always the case with nouns and adjectives; i.e., why aren't the
forms hē and to instead of unchanging ō?] (4) In the plural, why aren't the vocatives
identical with nominatives, as is true of every ordinary case-word when used as a
vocative? (5) When all forms of the article that don't begin with t- otherwise have
h-, why doesn't 5 also? (6) Why doesn't it have the same accent as the nominative?
There's no reason to suppose that 5 belongs to a special class of syllables that can't
take the acute [i.e. must be circumflexed] [[in standard Greek]]. (7) If it is cir-
cumflexed, why isn't its accent changed to grave in the Aeolic readings of Aristar-
chus, like all the other circumflexed forms of the article (e.g. ton, tots, toi, etc.]? O r
[isn't it clear that] he didn't think of this word [5] [as an article]? (8) Why doesn't it 52
start with t- like all the other [nonnominative] forms of the prepositive article?
There's nothing to cause the d r o p p i n g of t-. (9) Why doesn't the postpositive
article (i.e. the relative p r o n o u n ) have a vocative? Maybe because the prepositive
article doesn't have one either.
like meizōn "greater", voc. meizon (all three genders), and a few like charieis "pleas
ing", whose vocative charien may be masculine or neuter, but not feminine.]]
85. T h e words which are combined [unchanged] with [nouns of] different
number and gender are never articles, but indeclinable words (morion)—namely
conjunctions, adverbs and prepositions. And obviously, since ō is not a conjunction
or a preposition, it must be an adverb, an uninflected vocative adverb. T h e n why
does it go with nouns? Because it's a particle co-occuring with vocatives, just as the
oath particles ne and ma ["by" a god, ne in the affirmative, ma negative] go with the
accusative case, and vet are not at all articles. And their construction resembles the
"syncretism" of ō, since thev, too, occur with all genders and all numbers. Later we 53
will show that even where conjunctions are restricted from occurring with differ
ent cases, like heneken (heneka, heineka, "because o f ) , as in heneken Apolloniou
("because of Apollonius"), heneken Dionusiou ("because of Dionysius"), heinek' ernes
eridos kai Alexandrou {Iliad 3.100) ("because of the quarrel between me and Paris"),
that even that fact of case restriction does not prove that heneka is anything but a
conjunction. Parts of speech do not alter their sense because of restrictions on
their construction [and so the restrictions that say ma and ne must have an
accusative, and 5 a vocative, do not affect the decision.] [[Some linguists disagree
with A.D. and would say that heneka is certainly a preposition, ma and ne almost
prepositions, and even ō might be so regarded. But in Greek, many prepositions
are found with two or all three oblique cases.]]
86. (XXII) Similarly interrogatives do not admit the adjunction of the article,
since the article signifies a previously identified referent, but interrogatives are
used in ignorance of identitv, as we have shown above.
87. (XXIII) No one should imagine that in the word hopoios the initial ho is the
article, because of its form, and the fact that it has a rough breathing (h-) before a.p
[[alludes to a spelling rule that initial o before p should be smooth unless the
meaning is anaphoric]], and because it has an antecedent entity (prosopon) which it
refers to in sentences like:
1.151 hopoios esti Theon, toioutos esti kai Truphon
("Whatever T h e o n is like, T r y p h o n is just the same.")
In response to such an opinion we can say (1) that hopoios is not poios ("what sort?")
plus the article ho, but is a single word-form hopoios, which does not convey the
kind of reference {anaphora) indicated by the article as in:
1.152 ho anthropos paregeneto ("the man arrived")
1.153 ho hippos edramen ("the horse ran")
which is a reference to something known as an individual. T h e kind of reference
shown by hopoios requires a correlation of similarity, just as does hoios [[the plain 54
qualitative relative, like Latin qualis as a relative "(of the sort) of which . . . "]]. T h e
same applies to hosos and helikos [[plain relatives of quantity and size "the n u m b e r
which" and "the size which"]]. For there is no article meaning about t h e m either.
T h e short correlatives [i.e. hoios, hosos, helikos] behave exactly like the ones begin
ning with hop- [hopoios, hoposos, hopelikos], so we can say:
54 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
to the indefinites (like poios, tis), which are identical to the interrogatives except for
accent, starting from the neuter form ho ti ("which something", i.e. "whichever",
"whatever").]]
92. (XXV) For this reason [i.e. that an aspirated vowel initial requires a correla
tive with unaspirated consonant], the Ionic reversal of aspiration does not apply to
these words, (the lonians in general change unaspirated consonants to aspirated
and aspirated to unaspirated, as in: taphos ("astonishment") tethepotes ("as
tonished"; -thep- is the root coresponding to taph-), enthauta ("here" for Attic and
Koine entautha) and also in the r u n n i n g together (sunaloiphe) of prefixes;
1.155 eskatorais polin ("You look down on the city")
[when Attic would have eiskathordis, from eis- "on", kata- "down", and horais "you
see"]. This, then, doesn't h a p p e n to poios and the rest, since the lonians say koios
and kosos. For they couldn't change p- to ph- in these words, because of the
aspirated initial of the correlatives (hoios and hosos). A n d they couldn't accept any
consonant except k-here, because, as we have proved, the correlative [to a rough
vowel] must be a smooth [ = unaspirated] consonant, and the only other smooth
consonant is t-, which can't be used because toios and tosos already exist. [So, if they 56
had to change p-, it had to be into k-.]
93. (XXVI) Similarly, the constructions of book titles [[this must be the sense,
whatever the correct text is]] do not accept definite articles, because they are first
separated from the general category of poems or prose works [by these titles]:
1.156 proton Alkaiou ("Alcaeus' first [book]")
1.157 Phoinissai Euripidou ("Euripides' Phoenissae")
but once their identities are known, they can later be referred to with the article:
1.158 hai Phoinissai Euripidou periechousi ton Thebaikon polemon
("The Phoenissae of Euripides covers the T h e b a n war.")
1.159 to proton AIkaiou anegnomen
("We read the first book of Alcaeus.")
This concludes the discussion of nouns which cannot be used without the
article and of those which cannot take the article.
94. (XXVII) Next we must discuss the construction of articles with p r o n o u n s ,
which some people classify into articular and non-articular [pronouns], as if the
nouns we have just been discussing should not be called non-articular. A n d yet
they perhaps deserve to receive the name even more, since these nouns which
cannot take the article are absolutely resistant to the adjunction of articles. For
adjectives [[such as "non-articular"]] c o m p o u n d e d with alpha privative ("non-",
"un-" "-less") are understood as marking the absence of some expected or a p p r o
priate quality. E.g., when we say:
1.160 atechnos auletes "an unskilled flutist"
we indicate the lack of a skill which is appropriate. Similarly when we say:
56 THE SYNTAX OF APOLLONIUS DYSCOLUS
answer to this one may point out that a similar p h e n o m e n o n occurs with other
word-types, e.g. in the patronymic Priamides("son of Priam") there is implicit the
word huios ("son"), and yet that word may be redundantly added, as inlliadl 11.490 59
1.163 Priamiden nothon huion
("a bastard son of Priam")
and in comparatives there is implicit the word mallon ("more"), and yet often mallon
occurs redundantly with a comparative, as in Iliad 24.243:
1.164 rheiteroi gar mallon Achaioisi de esesthe
("for you will be more easier for the Greeks [to kill, now that
Hector is dead]")
In the same way the anaphoric [article] is implicit in autos;so whenever another
[explicit one] is added, the one implicit in autos is now duplicated by the explicit
article [ho], expressing the same anaphora, which is thus said twice, exactly as the
comparative notion is expressed twice in sentence 1.164. It is clear that this
possibility applies [only] to the ordinary [third-person] p r o n o u n [autos,] for the set
hoi (him, dat.) he, (him, acc.), min (him, acc), and the related forms (i.e., most likely,
some of the ones beginning with sph-, sphoe, sphoin, spheis, sphon, spheōn, sphi, sphe,
etc.] are not even used except by poets, and they belong to the same class as the
pronouns of the first and second persons, with which the article is completely
unacceptable. [[I.e., since ho autos has double a n a p h o r a so that it means "the same
(one)", one might think it possible to say Hoi hoi to mean "to the same (one)", but it
isn't.]]
99. Still, we mustn't forget to mention that Attic usage in exceptional cases
redundantly adds the article to the accusative [only], e.g. ton erne ("(the) me"),tonse
("(the) thee"), for instance, in Callimachus (Aetia I, fr. 28 Trypanis):
1.165 ton se Krotopiaden
("You, the grandson of Crotopos (i.e. Linus)")
or (fr. 114,5):
1.166 nai ma ton auton erne
("Yes, by (the) myself!" spoken by a statue of Apollo).
Unless maybe the article indicates outstanding excellence in example 1.165, since
such a structure is especially liable to the adjunction of the article. We've already 60
remarked [Ch. 43] on the use of the article in the sense par excellence.
100. (XXIX) T h e so-called articular [i.e. possessive] p r o n o u n s , used deictically
in the first and second persons, may be thought by some to take articles, e.g. ho emos
("mine"nom. sg. mase), ho sos ("thine" n.s.m.) or hohemeteros ("ours" n.s.m.). But in
this case the article does not belong to the pronominal person, the one that
indicates the possessor, but rather with the one understood as possessed—e.g.
doulos ("slave") or oikos ("building") or any other such. [[He's chosen singular
masculine nouns as examples of what may be understood with ho by ellipsis. And,
just as in English, there is always 'ellipsis in these constructions. But, unlike
English, the forms may also be used with nouns, where we would have to say "my"
instead of "mine"-h.o emos doulos "my slave" as opposed to "a slave of mine".]]
58 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
101. Habron, however, in his "On the P r o n o u n " argues that the articles do not
belong to the possessed nouns, as follows: "If the articles applied to the thing
possessed, they would have the same degree of grammaticality if they were shifted
to a place immediatelv before the head nouns; but ho emos pater ("my father") does
not mean the same thing as emos ho pater ("the father is mine"). And if in those
instances the article applied to the n o u n , what about cases where emos [as well as
pater] takes an article?
1.167 ho pater ho emos philosophei
("My father is a philosopher.")
T h e r e the second article must belong to the pronoun." And for the sake of that
construction he assigned a special article to the p r o n o u n in that pattern and to the
possessed object as well. [[The "possessed object" in this chapter is regularly called
to hupakouomenon, which in other cases might mean "the understood item", i.e. "the
deleted item." But here the noun is always overt, so this term must mean "the head
noun", or something similar.]]
102. But it's easy to dismiss this specious a r g u m e n t . First, if both words have the
article independently, why can't you begin with ho emos and then add ho pater with
article? [I.e. why isn't *ho emos ho pater grammatical?] It's evident that the two
arrangements are not the same: in the first one, ho pater ho emos, you can have two 61
articles, but this is impossible in the second [only emos ho pater, not *ho emos ho pater].
In the second place it doesn't follow from the facts about transposition of articles
and the resulting differences of meaning that the articles don't belong to the
nouns, since there is a similar difference between:
1.168 hoi nun anthrbpoi agathoi eisi
("Modern men are good.")
and:
1.169 nun hoi anthrbpoi agathoi eisi
("Nowadays, men are good.")
and nobody will have the audacity to claim that hoi does not belong to anthrōpoi.
Therefore you can't argue either from the change of sense caused by transposing
that the article doesn't belong to the thing possessed in both ho emos pater and emos
ho pater. We will clarify the difference between these phrases at the p r o p e r time
[Ch. 132-135].
(XXX) And the use of two articles with a single [head] case-form is not
abnormal. E.g. it's possible to say ho pater ho ekeinou ("His father", i.e. "the father
the of him"), where obviously both articles belong to pater, for how could a
non-articular p r o n o u n [see above, Chapters 95-97] in the genitive case take an
article in the nominative? T h e same thing occurs with p r o p e r names:
1.170 ho doulos ho tou Aristarchou pros erne elthen
("Aristarchus' slave came to me.")
Notice here that the n o u n in the genitive [Aristarchou] has its own article in the
genitive [tou] and the other two articles [ho. . . ho. . . ] apply to the same nominative
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 59
[doulos "slave"]. So there's no reason to claim that in ho pater ho emos the second
article must apply to the p r o n o u n .
103. T h e form emou (gen. sg. of emos, "my") is double genitive, one part being 62
genitive of the possessor, the personal p r o n o u n ("me")—whence the name by
which it is called ("possessive pronoun"), which indicates the deixis of a specified
person, and may be varied to second and third (i.e. "your", "his"), and the other
the genitive of the possessed head, which may be changed by inflection t h r o u g h all
the cases, genders and numbers in concordance with nouns [i.e. emoi d. sg. masc/
neut, erne n. sg. fern., emous acc. pl. m a s c , etc.]. T h e pronominal genitive [implicit
in the form] is always understood only as genitive, and cannot shift into other
cases. For the genitive case determines the possessive meaning. T h e r e f o r e every
possessive p r o n o u n is analyzable as a genitive plus a possessed head. Clearly, then,
the [underlying] personal p r o n o u n s [implicit in the possessives] are not inflected
in other cases or anything else, but the possessive p r o n o u n s are always inflected in
cases, genders and numbers. [[This last sentence r e n d e r s Hilgard's reading.]]
104. For this reason there are two special word-types which can co-occur with the
[possessive pronouns] relating to the two referents involved, the article to go with
the head nominal [in whatever case is required], and, with the inherent pronomi
nal person, forms of the appositive (intensive) p r o n o u n [autos"seif', in genitive
"own"], always in the genitive; as in Odyssey 2.45:
1.171 all' emon autou chreios, ho moi kakon empesen oikoi
("but my own need, the evil which has befallen my house.")
or Odyssey 1.7:
1.172 autōn gar sphetereisin atasthalieisin
("For by their own follies. . . ")
or Odyssey 1.409:
1.173 e heon autou chreios
("Or his own need. . . ")
[[In these examples the possessives emon, sphetereisin and heon agree in case, gender
and n u m b e r with their respective head nouns chreios, atasthalieisin, a n d chreios, but
the intensives autou, auton, and autou are genitive, agreeing with the implicit 63
personal pronouns.]] A n d the article is inflected in all cases, according to the case
of the thing possessed—ho emos ("my, mine" n o m . sg. m.),tou emou (gen. sg. m.),toi
emōi (dat. sg. m.), and it likewise agrees in gender—he erne (nom, sg. f.), to emon
(nom./acc. sg. n.). Clearly the possessing person is always one and the same, but the
things possessed may differ in gender. T h e same applies to n u m b e r : hoi emoi(nom.
pl. m.),hai emai (nom. pl. f.), ta ema (nom./acc. pl. n.). A n d if the [possessed head] is
changed in n u m b e r , the forms of the article change along with it. Suppose several
people own one field, they would say ho hemeteros agros skaptetai ("Our field is being
d u g up") with the article (ho) in the singular to agree with the head n o u n (agros). So
it is clear that in such constructions any accompanying articles have nothing to do
with the deixis of the p r o n o u n s . (Incidentally, this suggests another nice a r g u m e n t
that 5 is not an article, since it can be placed in front of the deictic p r o n o u n [o houtos,
"hey you!", whereas *ho houtos, *hoi houtoi etc. are ungrammatical]).
60 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
105. So much for the [case-words] which occur exclusively with articles and those
which never take them.
(XXXI) T h e remaining structure is nominal forms which take the article or
not according to the r e q u i r e m e n t s of the utterance-situation: hippos trechei
("There's a horse running") [when the horse is new to the discussion], or, where
the horse is one already known [to speaker a n d hearer], ho hippos trechei ("the horse
is running"), ploion katepleusen ("a ship set sail"), to ploion katepleusen ("the ship set
sail). O n this basis we can say that there is ellipsis of the article [i.e. deletion of an 64
underlying article] in Odyssey 9.550:
1.174 arneion d'emoi oiōi euknemides hetairoi
("but my well-greaved comrades [gave] the ram to me alone")
because [the ram] is here mentioned as one which had been described earlier, in
lines 444 and 433:
1.175 hustatos arneios melon esteiche thuraze
("Last of all the sheep, the r a m went to the door.")
1.176 tou kata nōta labōn
("holding on to him a r o u n d his back").
[[Line 432 actually introduces the ram, arneios gar een, melon och' aristos hapantbn
"there was a ram, far the finest of all the sheep."]] For if that were not the case,
there would be no ellipsis of the article [since it was never there to delete]. For it is
explicitly added, e.g. in Odyssey 11.35:
1.177 ta de mēla lahbn apedeirotomesa
("And taking the sheep, I cut their throats")
H e r e the sheep were first mentioned at the m o m e n t when they were p u t on board
by Circe (Od. 10.572).
106. This rule [about the article] is valid for every construction with verbs,
except for verbs which refer to future existence {esomene huparxis) [i.e. verbs which
take a predicate nominative, verbs of becoming etc.]. Consider sentences like:
1.178 Dionusios ho gnorimos mou thelei philologein
("My acquaintance Dionysius wishes to study literature.")
1.179 houtos ho anthrbpos thelei anaginbskein
("This fellow wants to read.")
T h e rule applies to these [i.e. they are already identified, hence take the definite
article]. But suppose we insert verbs like genesthai ("become"), kaleisthai ("be
called"), or others of the same sub-class, then the article will not occur:
1.180 Dionusios thelei moi gnorimos genesthai
("Dionysius wishes to become acquainted with me.")
1.181 Thebn grammatikos thelei kaleisthai
("Theon wants to be called a grammarian.")'
And this is quite reasonable, since verbs of this class signal a relatively future
quality, whereas the article means something relatively past. But if you go ahead
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 61
and add the article [before the adjectives gnorimos and grammatikos in the last two
examples], you will also have to add another adjective, to function as a predicate
adjective complementing the verb, since the addition of the article took the first
adjective away [from the verb phrase, and attached it to the subject n o u n phrase]
as belonging to preexisting knowledge:
1.182 Dionusios ho gnorimos mou thelei genesthai philosophos
("Dionysius my acquaintance wants to become a philosopher.")
Obviously, if one now removes the article [ho from this last sentence], the con- 65
struction becomes ungrammatical, because the verb (genesthai) cannot take two
[predicate] adjectives [unjoined by "and"]. When, in other constructions, two
nouns refer to the same person, there's a difference from the verbs described
here. T r y p h o n began to explain the construction of these with articles, but didn't
offer even a hint about the patterns just mentioned h e r e .
107. (XXXII) W h e n adjectives are construed with p r o p e r names, they always
take articles unless they are predicates with verbs of being or becoming (existence:
huparktikos):
1.183 ho grammatikos Truphon anaginoskei
("The grammarian T r y p h o n is reading.")
1.184 ho philosophos Dion peripatei
("The philosopher Dion is walking.")
Otherwise the article cannot be omitted unless a participle of being or becoming is
inserted:
1.185 grammatikos on Truphon anaginoskei
("Tryphon, being a grammarian, is reading." or
"Tryphon, who is a grammarian, is reading.")
1.186 philosophos on Dion dialegetai
("Dion, being a philosopher, is conversing.")
A n d this is quite reasonable because the quality is now being asserted with the
participle on. T h e same thing could not be accomplished with a finite verb because
of the additional verb, since two verbs cannot form a single construction unless
they are conjoined. This is obvious from the examples, first:
1.187 ho philosophos Dion dialegetai
("Dion the philosopher is conversing.")
second:
1.188 philosophos on Dion dialegetai
("Dion, being a philosopher, is conversing.")
and third:
1.189 philosophos esti Dion kai dialegetai 66
("Dion is a philosopher and is conversing")
62 THE SYNTAX OF APOLLONIUS DYSCOLUS
And obviously it is only logical that verbs which link adjectival predicates should
reject the construction [of the adjective] with the article. And hence it is obvious
that in Iliad 1.11 we have a case of anastrophe [i.e. scrambled word o r d e r — h e r e he
means that ton "the" belongs in front of aretera, and has been figuratively displaced
by H o m e r ] :
1.197 houneka ton Chrusen etimasen aretera
("because he had dishonored the priest Chryses").
for if one were to replace etimasen ("dishonored") by onomase ("called"), it would
alter the construction:
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 63
For if you want to make these [last two] unique, you will a d d an article—he
therapaina sou ("your maid-servant"), ho gnorimos [ho] Aristarchou ("Aristarchus'
friend"). And if you have a lot of ships, someone could tell you naus sou katepleusen,
naus sou hesteken en tōi limeni ("A ship of yours sailed into port", "A ship of yours is at
anchor in the harbor"). But once again, if anyone has in mind unique possession,
then he must add the article, and hence we must suppose ellipsis [of the article] in
Odyssey 1.185:
1.231 neus de moi hēd' hesteken
("My ship is anchored here.")
as we learn [[we must e m e n d here to edidachthemen or the like]] [frorn another]
passage (Od. 1.182):
1.232 nun d' hōde xun nei kateluthon
("Now, as you see, I have just arrived with my ship")
which clearly implies "with one ship". It is clear also from the word hode, signifying
"thus, as you see", not "as befits a king". Obviously also the article is deleted by
ellipsis in Iliad 1.1:
1.233 mēnin aeide thed
("Sing, goddess, the wrath. . . ")
i.e. ten Achillebs oulomenen, "the deadly wrath of Achilles". For this is a uniquely
possessed passion, just like moira ("life-destiny") and psuche ("spirit") in the first
examples above.
119. (XXXVII) Next we must go on to interrogative constructions. W h e n tis
("who?") is used with a verb such as estin ("is") or onomazetai ("is called"), or some
synonym, and a p r o p e r name, it expects as a response a p r o n o u n , e.g.:
1.234 tis Truphōn onomazetai
("Who is n a m e d Tryphon?")
1.235 tis Truphbn legetai
("Who is called Tryphon?")
to which we answer ego ("I") or houtos ("he") or some other word in that category 73
[i.e. personal or demonstrative p r o n o u n s ] . And, conversely, when the question
includes a p r o n o u n [as predicate nominative], the answer will be a name. In
Iliad 10.82, Nestor, being ignorant [of someone's identity] because of the situation
[namely, darkness], says:
1.236 tis d'houtos kata neas ana straton ercheai oios?
("Who are you coming along through the camp by the ships?")
to which he is answered (/Il. 10.88):
1.237 gnbseai A treiden Agamemnona
("You may recognize A g a m e m n o n son of Atreus").
But Priam, when he sees A g a m e m n o n in Iliad 3.166-7 asks for his n a m e , to which
he is answered by Helen, in 178:
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 69
[[By A.D.'s criteria, " p r o n o u n " applies only to the inherently definite personal
p r o n o u n s and demonstratives, though elsewhere he includes possessives (see Ch.
131), which aren't always definite but do imply definite possessors.]]
126. (XXXVIII) If, in sentences of the above types which cannot take articles, we
replace the interrogative nominal [i.e. tis "who", posos "how much", pelikos "how
big"] by an adverb [i.e. pos "how",pote "when", pou "where", etc.] then the NP's will
take articles [[structurally, of course, it is these articles which are replacing the tis,
poios, or whatever within the NP]], for example:
1.258 (a) poios anthropos enikesen
("What man won?")
(b) pōs ho anthropos enikesen 77
("How did the man win?")
1.259 (a) tis dramon enikesen
("Who won the race?")
(b) pos ho dramon enikesen
("How did the man who ran win?")
1.260 (a) posos ochlos paraginetai
("How big a crowd is coming?")
(b) pote ho ochlos paraginetai
("When does the crowd come?")
So for instance, take a sentence of this type:
1.261 poiai ton Troon phulakai phulassontai
("Which guards of the Trojans are on duty?")
and substitute an adverb in this way, and it will absolutely require an article:
1.262 pōs hai ton Troon phulakai
("How do the guards of the Trojans [guard]?")
127. So why wasn't one inserted by Aristarchus in Iliad 10.408 [cf. Ch 6 above]:
1.263 pos toi ton Troon phulakai
("How . . . the [guards] of the other Trojans . . .")
when (a) the sentence-context demands the article, and (b) the text contains the
right letters [AI, in this case] for the article? A n d besides that, not only is the article
required because of the preceding interrogative adverb, but (c) when genitives
occur with the article in possessive constructions, the possessed nouns also take the
article, as in:
1.264 hai ton Hellenon poleis epanestesan tois barbarois
("The cities of the Greeks revolted against the Persians")
This becomes ungrammatical without the article. Both nouns must have the article
because, if the article is omitted with poleis ("cities"), the context requires tines
("some"), to indicate a more limited revolt:
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 73
135. We must not overlook the fact that any nominal [adjective, p r o n o u n or
participle—this term ptotikon represents a kind of super-form class for A.D., like
76 THE SYNTAX OF APOLLONIUS DYSCOLUS
136. (XLI) T r y p h o n says that the nominative article ho ("the") can be set before a
n o u n in any oblique case [with its own article], provided a participle follows and 82
then the p r o n o u n houtos ("this man"), with estin ("is") as main verb, e.g. ho ton
anthropon hubrisas houtos estin ("this is the one who injured the man"). [[This is one
of the Greek patterns most like what are called "cleft" and "pseudo-cleft" sen
tences in English, incidentally.]] But in fact he could have m a d e his statement
more general: [instead of houtos and estin, say] any nominative NP and any verb of
existence, i.e. of n a m i n g or being:
1.282 ho ton anthropon hubrisas Theon onomazetai
("The one who injured the m a n is called Theon.")
ho ton anthropon laktisas hippos estin
("What kicked the m a n is a horse.")
ho ton anthropon hubrisas ego eimi
("The one who injured the m a n is me.")
( = ' T m the one who injured the man.")
[[In fact, of course, T . and A.D. could be even m o r e general; since the pattern ho +
ton + N/acc + Partic/nom. is a complete subject NP, any appropriate predicate at
all can follow—"eats cheese", "gets u p at dawn", "joined the Senate", "spoke in the
Assembly", etc., etc.]] T h e reason for this can be worked out as follows.
137. T h e oblique [forms] are connected with the nominative ones by means of a
verb whose action passes over from the nominative [NP] to the oblique [NP]
1.283 ton anthropon hubrisen Theon
("Theon injured the man.")
ton anthropon elaktisen hippos
("A horse kicked the man").
Now if we adjoin an article before the nominative n o u n in-this construction, the
structure remains grammatical:
BOOK I: THE DEFINITE ARTICLE; THE RELATIVE PRONOUN 77
If the genitive is not used in this way, but is governed by a verb, then the article is
not necessary—anthropou akouo ("I hear a man"), agrou despozo ("I possess a field").
140. (XLII) T h e possessive construction we have been explaining takes two
84
articles with c o m m o n nouns:
ho Apollōniou philos
("Apollonius' friend").
A n d the word basileus ("king") in this respect behaves like a n a m e ; we say:
1.292 ho basileos oiketes paregeneto
("The king's servant arrived.")
although our c o m m o n n o u n s cannot be so used. And p e r h a p s it's natural and
right for common nouns to take two articles, since it's hardly likely that someone's
possession could be known if the possessor wasn't also known even earlier. [[The
a r g u m e n t fails for indefinite possessors—"someone's slave", i.e. "the slave of
somone"—but A.D. ignores this case here.]] Either both n o u n s occur without an
article:
1.293 leontos skumnion edramen
("A lion's cub ran. . . ")
or both alike are anaphoric:
1.294 to tou leontos skumnion edrame
("The lion's cub ran. . . ")
A n d p r o p e r names, because of their u n i q u e application, are not thus b o u n d to the
article. So it seems quite acceptable to say ho basileos oiketes (the King's servant"),
since the word basileus is, in effect, a name. For if one says "the c u r r e n t ruler" one
understands "Ptolemy", a n d indeed this is m o r e specific t h a n saying "Ptolemy",
since there are other Ptolemies, but only one to whom the kingship belongs.
141. T h e r e is still another participial construction, with two articles in a row, like
the above constructions (Ch. 136-7) and with a following infinitive verb:
1.295 ho ton anthropon thelon hubrisai houtos estin
("The one who wants to injure the m a n is this o n e " )
A n d in general the structure with a verb of wishing transformed into a participle
completes the sentence in this m a n n e r . But the reason why verbs of wishing 85
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 79
143. Prepositive articles precede their nouns and accompany them [to form an
NP which is] in some relation to a single verb or participle:
1.298 anthropos peripatei
("A man is walking.")
1.299 anthropou aidontos ekousa
("I heard a man singing")
and similarly, with article:
[ho anthropos peripatei
("The m a n is walking")
tou anthropou didontos ekousa
("I heard the man singing")
tou anthropou tou didontos ekousa
("I heard the singing man.")]
This is impossible in the case of the postpositive article, at least in a simple sentence
construction. For it always requires the presence of a second verb [of which the
relative may be the subject]:
1.300 elthen grammatikos hos dielexatolhos anegno
("A grammarian came who had spoken/who had read.")
and in case the verb involves transitivity, i.e. an additional NP, [then the relative
may be in an oblique case]:
80 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
145. There's still another construction equivalent to that with the 'postpositive 87
article', namely that using anaphoric p r o n o u n s :
1.308 anthropoi homilesa kai autoi pareschon xenian
("I spoke to a man and offered him hospitality.")
grammatikos paregeneto kai houtos dielexato
("A grammarian arrived and he conversed.")
For the anaphoric force {anaphora) which is present in the "article" is shared with
such p r o n o u n s , and so the sentence-types constructed with them are closely
related.
146. Well then, should we propose to unite these two u n d e r one part of speech
[as later g r a m m a r i a n s did]? By no m e a n s . T h e m e r e fact that [pronouns]
claim some of the properties of 'articles' doesn't make them 'articles', since the two
word-types differ in other respects. So [finite] verbs, too, have in themselves the
force of nominative personal p r o n o u n s [i.e. homilo by itself means "I speak"], but
are not therefore the same part of speech. Once again, they differ from each other
in many characteristic features. So we have explained that coordinating conjunc
tions may be substituted for 'postpositive articles', but that doesn't mean that
conjunctions and 'articles' are a single part of speech. [[Many later grammarians,
however, do regard relative p r o n o u n s as implicitly containing conjunctions, or as
cross-classified with conjunctions. See Bloomfield's Language (1933) p p . 262-3.]]
147. But [demonstrative and anaphoric] p r o n o u n s don't even agree [with rela
tives] in this very construction. For they d e m a n d the addition of a conjunction:
1.309 grammatikos paregeneto kai houtos dielexato
("A grammarian came and he spoke.")
It is obvious that if one takes away the conjunction, the sentence becomes non
sense. And besides, [with demonstratives] the person is intended more demon
stratively:
1.310 grammatikos paregeneto kai houtos dielexato
("A g r a m m a r i a n came, and that one spoke.")
with deictic emphasis. And in cases where autos is used [in the nominative], the 88
person is emphasized as the one par excellence:
1.311 grammatikos paregeneto kai autos dielexato
("A g r a m m a r i a n came and he himself spoke.")
as if you were to say "the lord", "the master", etc. Now if you insert a conjunction in
the construction before the relative p r o n o u n , it is clear that the article will be
disassociated and no longer refer back to the original antecedent noun, but to
some other understood indefinitely:
1.312 anthropoi homilesa kai hoi xenian pareschon
("I spoke to a man and to [someone] to whom I gave entertain
ment.")
[[This again depends on the fact that n o u n and relative are in the same case. If
they're not, the resulting sentence will usually be uninterpretable and u n g r a m -
82 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
matical.]] So, when the difference is so great, how can anyone suppose that
'articles' and p r o n o u n s should be reduced to a single part of speech?
148. (XLIV) "Postpositive articles' (relative pronouns) are grammatically unaf
fected in case [[he here ignores the patterns called "attraction" and "inverse
attraction"]] by the case of the antecedent nouns to which they make reference
(anaphora):
1.313 anthropos elthen hon exenisa
("A man [nom.] came whom [acc.] I entertained")
philou ekousa hoi kai prōēn sunebalon
("I heard a friend [gen.] whom [dat.] I had also met the day before
yesterday")
A [relative p r o n o u n ] in an oblique case may take as antecedent a n o u n in the
nominative when the event of the [second] verb has a different entity as subject;
and a [postpositive article] in the nominative may refer back to a n o u n in an
oblique case when the reference {anaphora) of the article assigns the [second] verb
to the [action of the] antecedent in the oblique case. A n d both may be in the
nominative—both the n o u n and the 'postpositive article'—when the same entity
("person") is the subject of both verbs.
149. An example of the first case [relative p r o n o u n obl., antecedent nom.] is:
1.314 grammatikos elthen hoi ho Truphon homilesen
("A grammarian came to whom T r y p h o n spoke.")
Here the action of the verb [homilesen, "spoke"] derives from the nominative 89
Truphon [not from grammatikos; i.e. T r y p h o n is subject of the relative clause].
Examples of the second case [antecedent obi., relative nom.] are in H o m e r , Odyssey
1,1-2:
1.315 andra moi ennepe, Mousa, polutropon, hos mala polla/planchthe
("Tell me, Muse, the m a n of many wiles, who wandered far an
wide.")
and in Iliad, 1,1-2:
1.316 [menin aeide, thea. . . ] oulomenen, he muri Achaiois alge etheken
("[Sing, goddess] the deadly [wrath], which [caused] many [troub
les] for the Greeks.")
T h e antecedents here are oblique, since the subject of the verb in both cases
{ennepe "tell", aeide "sing") is a Muse, but the [relatives] are nominative, since the
actions [of planchthe "wandered" and etheken "caused"] are d e p e n d e n t on "the
man" and "the wrath" respectively [i.e. their referents are the actors]. For it is he
[Odysseus] who wandered, and it is that [wrath] which caused many troubles for
the Greeks.
And, finally an example of the third case [both antecedent and relative in
nominative] is Iliad 7.125, 127:
BOOK I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 83
positive articles", while the position [after the antecedent NP] also shows this, as in
the case of H o m e r , Iliad 21.193-4:
1.322 [Dii Kroniōni machesthai]
tōi oude kreion Acheloios isopharizei
(". . . to fight with Zeus, son of Cronos, for whom even the mighty
Achelous is no match").
H e r e [we take tōi as a relative] (1) because of the position —since it is placed after its
n o u n (Dii)—and (2) because of the construction, since it is governed by isopharizei.
W h e n such forms [apparently definite article forms] are to be taken neither as
postpositive nor as prepositive articles, but as deictic (demonstrative) p r o n o u n s [so
many editors take it in the H o m e r passage just quoted] or as indefinites, we will
explain when we come to the constructions of p r o n o u n s [Bk. 2, Chs. 28-32].
152. (XLVI) In the light of that, let's consider Iliad 18.460:
1.323 kai thorech'; ho gar en hoi, apolese pistos hetairos
(". . . and breastplate; for the one he had had, his trusty comrade
had lost.")
and investigate whether the ho here is (1) a prepositive article going with pistos
hetairos ("trusty comrade"), or (2) is being used in place of hos as a postpositive
91
(relative), referring back to thoreka ("breastplate"), or (3) is used in the postpositive '
construction with reference forward to hetairos. (1) It's impossible for it to be taken
as a definite article; a finite verb can never be construed with a definite article, as en
is here. So it must necessarily be a postpositive article, but cannot very plausibly
refer back to the breastplate; since not only was that lost by Patroclus, but all
together (458-460):
1.324 aspida kai truphaleian
kai kalas knemidas, episphuriois araruias
kai thoreka
("shield and helmet, and fine greaves fitted with ankle-guards,
and breastplate").
A n d if you can overlook this evidence, there's still the ungrammaticality of the
structure of the sentence. For it will be granted that the existence of the breastplate
is presupposed by the construction with the article, as we have just explained. And
if this is so, how can the pistos hetairos ("trusty comrade") be construed as subject of
a verb which belongs to a different subject, which is in t u r n construed with a
different verb, en? [[This is by no means clear, but he seems to be rejecting an
interpretation like this "the breastplate, which he had which his trusty c o m r a d e
lost" or perhaps "which—he had had one—his trusty c o m r a d e had lost".]] W h e n
can two nominatives [with a verb like apolese] make a grammatical sentence?
Rather we should say that an oblique and a nominative make a grammatical
sentence [with such verbs].
153. T h e nominative [NP] marked by the ho must therefore be taken as part of
the same nominative [NP] as hetairos:
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 85
After Aithiopas comes the 'postpositive article' hoi, which H o m e r gives in his 93
customary form with initial t:
1.331 toi dichtha dedaiatai, eschatoi andrōn
("who are divided in two, the farthest of men")
and it is clear that the verb dedaiatai is construed with the plural subject (toi "who"),
which is [appositionally] subdivided in the next line in correspondence with the
separation of the Ethiopians, [into two NP's] with postpositive articles (i.e. hoi men,
hoi de) which have the same verb in common [dedaiatai "are divided"). [[This
construction of partitive apposition with X men. . .X de. . . usually has a definite
(prepositive) article in classical Attic, and many editors read this line of H o m e r the
same way, since hoi, the article, differs From hoi, the "relative", only in the presence
or absence of an accent. T h e masculine singular relative form hos is distinct (as in
A.D.'s first example, about the eagles) and so are the neuter forms ho and ha as well
as all the oblique case forms (t- for article, h- for relative), and in Hellenistic and
later Greek the relative forms become common. T h e grammarians class all of
these uses (ho or hos, ha or ta) as "postpositive articles" because (a) a n o u n may
precede, and (b) a verb may follow. Most editors disallow the hos, ha, etc. relative
forms in Homer.]]
156. When a [plural] verb is present in this figure, then the appositional subdivi
sion after it will be acceptable, [ as in modified form of Odyssey 12.73 (cf. 12.100)]:
1.332 hoi de duo skopeloi hikanousin, hos men achri tou ouranou, hos de achri
tou Olumpou.
("The two rocks reach u p , one to heaven, the other to Olympus.")
But if no such prior verb is present, then the verb must necessarily go with the
postpositive article, and, as we showed above (Ch. 57), the plural nominative at the
beginning must be changed to [a partitive] genitive. For only this case [genitive]
may be subdivided, without the presence of a verb, into the component members
of the class [indicated by the genitive NP], but not so any of the other cases, unless a
[plural] verb is used, as in Iliad 16.317:
1.333 Nestoridai d' ho men outas' Atumnion
("The sons of Nestor, one [of them] w o u n d e d Atymnius. . . ")
H e r e obviously ho [definite article] is substituted for hos ["postpositive"] [[i.e.
H o m e r is deliberately using a schema, ho for hos]], unless perhaps you can say that
he has omitted a noun or p r o n o u n by ellipsis, which was the head n o u n for the
article ho,so that the underlying sentences go like this:
(a) ton Nestoridon ho men heteros outas' Atumnion 94
(("Of the sons of Nestor, the one w o u n d e d Atymnius. . . ")
(b) ton duo skopelon ho men heteros ouranon hikei
("Of the two rocks, the one reaches heaven,. . . ")
157. In the same m a n n e r we can analyze Aratus, Phaenomena 24-5:
B O O K I: T H E D E F I N I T E A R T I C L E ; T H E R E L A T I V E P R O N O U N 87
1. (I) I think I must now follow u p the treatment of the syntax of articles in Book 95
I with a discussion of the syntax of p r o n o u n s . Articles are used with nouns in
sentences, and p r o n o u n s are used instead of nouns, and do not admit of the
structure with prepositive articles for the reasons which we have e x p o u n d e d above
(Bk. 1, Ch. 95). [[Although, when demonstratives are used with nouns, the nouns
must be preceded by articles—houtos anthropos means "this one is a man", houtos ho
anthropos, "this man", etc.]]
2. And, obviously, the reason why p r o n o u n s came to be inflected for case is so
that they might be able to substitute for any n o u n ; and they distinguish all persons
to make u p for a deficiency of nouns, namely that they cannot mark person, when
they are put into nominal constructions. T h u s they share case with nouns, and
person with verbs.
3. (II) T h e r e f o r e they are totally inflected, using the two most general inflec
tions, and doing so without confusion, by an appropriate division [of functions]:
they put the case inflection on the end, and the personal characterization at the 96
beginning. Obviously, if both inflections were attached at the same point, the
change from one case to another would wipe out the distinction of person, and
conversely, the passage through the persons would destroy the characteristic of
the case. Therefore, they make the distinction of case as in emou, emoi, erne ("me" in
genitive, dative, accusative) and of person in emou, sou ("me", "thee", in genitive)
and of both person and case as in emou, soi ("me" genitive, "thee" dative). And in all
instances they are specified as to the two above-mentioned categories by char
acteristic features in each part of the word. Whereas both inflections, when
applied to their special word-types, n o u n and verb, use the word-end for inflec
tion: kalos, kalou ("fine" in nom. sg. masc. and gen. sg. m a s c ) , grapho, grapheis,
graphei ("I write, you write, he writes"), not unreasonably using deletion of the s as
distinctive of the third person, in p r o n o u n as well as verb (soi "to thee", hoi "to him"
[h's don't count as letters] like legeis "thou writest", legei "he writes"). [[This is not
quite like an analysis of s as a second-person m o r p h e m e , zero as third person,
because the Greek g r a m m a r i a n ahvays thought of inflections as changing from one
form to another, not detaching one element and adding another. But it's close.
Priscian, of course, can't take this over into Latin, where final -s is second person, -t
third, but initial t- is second person, s- third.]]
4. I think it's foolish to belabor the question why the end [of the p r o n o u n ] wasn't
used for person-marking, as in verbs, instead of the beginning. T h e end had to be
reserved once and for all for a single part of speech, and, if we may say so, logically
for the nominal, since the noun is prior to the verb, as we have shown (Bk. 1, Ch.
16-18), and obviously its characteristic feature will dominate at the end. Besides, a
p r o n o u n is, by definition, something used in [the syntactic] place of a noun and in 97
imitation of a noun, and clearly the person present in it is not part of its essence.
B O O K II: P R O N O U N S 89
And if endings are the most characteristic bits of parts of speech, and the end of a
p r o n o u n marks the case, it will be called ' p r o n o u n ' predominantly because of this
ending, even if it also puts on a characteristic feature of the verb, namely person.
5. I think it's reasonable, before explaining pronominal syntax in detail, to
explain the characteristic features of these words [pronouns] in o r d e r that the
account we give of the syntax may be easier to c o m p r e h e n d .
6. T h e special two-part inflection has already been dealt with, but there is also a
twofold classification of deictic force, one absolute, as me in epaise me ("He hit
me")—and it's clear that this does point [to "me"], but without intensification or
contrastive force—and the other oppositional, as eme in eme auton epaise ("He hit me,
myself). For it's clear that the deixis is intensified in much the same way as [is done
by comparison] in [adjectives of] quality. So leukos "white" has absolute quality, but
leukoteros "whiter" is intensive, contrasted with some other [white] entity which we
fully expect to identify. In this way eme has intensive deixis with reference to that
with which it is contrasted. And why this distinction [between normal and empha
tic forms] is not maintained in all p r o n o u n s [being confined, in fact, to the singular
personal pronouns], will be discussed in the detailed treatment [beginning in
Chapter 57].
7. T h e p r o n o u n s [intensive, or emphatic, or contrastive] which are phonologi-
cally complete and have the acute accent are called 'orthotonic' [i.e. accented], a
name which also involves the synonymy of orthos ("erect") with hugies ("sound")
[[both adjectives are used to describe the initial or base form of a word, from which
other forms are derived by inflection; h e r e it is implied that the enclitic p r o n o u n s
are derived from the accented ones]], the other forms, which shift their accent like
a b u r d e n , from the fact that they 'incline' their weight u p o n something else, are 98
called enclitic. So in Iliad 4.63 soi men ego, su d' emoi ("[Let us yield to each other] I to
you, and you to me."), we have the full [emphatic] forms which keep their own
accents (soi andemoi), but in Odyssey 13.232 kai moi taut' agoreuson ("and tell me this")
we have the reduced form (moi) and with accent shifted [onto kai].
8. T h i r d person p r o n o u n s [in addition to their demonstrative force] also have
anaphoric uses, so that nouns mentioned earlier may be referred to pronominally,
as in Iliad 13.1 & 3:
2.1 Zeus d'epei oun Trōas te kai Hektora neusi pelassen,
. . . autos de palin trepen osse phaeino.
("After Zeus had allowed Hector and the Trojans to draw near the
ships . . .
but he himself turned his bright eyes away . . . ")
[[This autos is also contrastive, opposed to a tous men ("them," i.e. the Trojans) in
line 2.]]
9. (III) In the first book (Ch. 25) we have shown that such p r o n o u n s do not
function as the equivalent merely of bare [common] nouns, but have a force
equivalent to such nouns preceded by the definite article. [[In English "one" may
stand for a bare noun, as in "a small house next to a big one", but "he", "she", "it",
"they" must stand for article-plus-noun (or definite NP) as in "That's the house, I
know it", for instance.]] For they not only represent nouns, as is the case with other
90 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
pronouns as well, but they also refer [to definite entities], which is not a property of
bare nouns, but only of nouns with definite article [or of names, cf. 1.120 and 2.41].
10. One can also say, that insofar as a p r o n o u n is substituted for something
previously said, in that respect it exhibits anaphora, since the defining char
acteristic of anaphora is the second recognition of a previously mentioned entity,
which is the case with autos in the passage just quoted.
11. It will be shown later (Ch. 40-41) that deictics [this includes demonstratives
and personal pronouns] are not used as substitutes for nouns, but where it is 99
impossible to use nouns. T h u s p r o n o u n s are divided into demonstrative (or
deictic) and anaphoric, though both types fall u n d e r the general heading [of
pronouns]. For in both cases, in spite of the differences, a single feature unites
them—syntactic equivalence to nouns. For they are used either (a) where nouns
cannot be used, or (b) where they have already been said once, and cannot be used
again [i.e., 'pronominalization' is obligatory for A.D.]. For if in Iliad 13.3:
2.2 autos de palin trepen osse phaeino
("But he himself turned his bright eyes back . . . ")
one substitutes Zeus for autos ("he h i m s e l f ) , it will not unify the two sentences
concerning Zeus, but will make this the start of a new sentence. [[This is, perhaps,
the earliest statement of the discourse-linking function of anaphoric pronouns.]]
T h e same thing is true of all pronouns used in this way [i.e. anaphorically]. For
although it is grammatically possible to substitute the [antecedent] nouns for
anaphoric pronouns, [the effect is to] alter the structure of the discourse.
12. Whenever ekeinos ("that (one)") and houtos ("this (one)") do not point to
something actually visible, but refer back [anaphorically], one should bear in mind
that their pointing (deixis) is at something in the mind; so some deixis is visual, a n d
some is mental. In fact, [these two words] being inherently deictic, are never
preceded by the article, which is inconsistent with deixis.
13. (IV) In the third person, by exception, different words can be used to
distinguish different "third persons", though verbs are restricted to a single
ending for any number of persons—graphei Dionusios—("Dionysius is writing") or
[graphei] Truphon("Tryphon is writing") or anybody else "is writing" of those to
whom the action can be attributed. [[The same e n d i n g (ei) is used, regardless of
which "third person" is the subject.]] This is not the case with p r o n o u n s , since we
have autos anaphoric ["him, her it", unemphatic in all oblique cases], and ekeinos 100
and houtos besides ["that one", "he" emphatic] as demonstrative (deictic), differing
in the distance of deixis [houtos near, ekeinos far.] T h e same applies to hode ("this
one", near').
14. (V) And since, as we said above [Ch. 6], deictic p r o n o u n s [i.e. personal and
demonstrative pronouns] admit intensification of deixis, and the forms u n d e r
discussion [the demonstratives] are incapable, because of their unaccented final,
of taking variation of accent [for this purpose], as is the case with emou, emoi and its
class [namely the singular oblique personal p r o n o u n s , which have both enclitic
and emphatic forms], they have adopted the addition of a final longz after the case
ending, taking an acute accent on it for indication of greater intensity parallel to
B O O K II: P R O N O U N S 91
the above-mentioned first and second person p r o n o u n s : ekeinosi, houtosi, hodi ("that
one, this one", emphatic). As for autos, since it has no share in deixis, it will also
obviously have no share in the marker which is added to intensify deixis. And the
choice of i as the element to be added does not seem entirely arbitrary, since that is
the base form of the nominative of the third person. [[The ancient grammarians,
in need of a nominative to go with the oblique forms he, hou, hoi, "him, her, it,"
chose a form hi, which is hardly attested otherwise.]]
15. So it was foolish of H a b r o n to ridicule Aristarchus on the g r o u n d that he was
wrong to say that p r o n o u n s are a conjugate class marked bv person: "this defini
tion is shared by verbs, and belongs more properly to them. For the conjugation
(suzugia) of forms is a matter of variation of phonemes and of n u m b e r of syllables,
and also appears in quantity [of vowels] and accentuation. But no such bonds
apply to pronouns, and most of them are inconjugate, as is clear if you merely
decline them." [[This was obviously a dispute based on different definitions of the
technical term suzugos ("conjugate"), which for Aristarchus and A.D.—perhaps
u n d e r Stoic influence—meant some sort of semantic-syntactic similarity, constitut
ing a "form-class", whereas for Habron it seems to have meant formal inflection,
variation of suffixes and the like.]] Obviously Aristarchus did not define the
phonological form of pronouns, but the underlying basic meaning, as is p r o p e r 101
for all definitions. [[This dispute, between p r o p o n e n t s of definition according to
formal criteria and those of definitions by semantic criteria, continued on through
the Renaissance and into the 20th Century.]] For instance, they [Habron's side] say
that p r o n o u n s do not take the article, but as far asform is concerned, they certainly
do:
2.3 he ego monōs orthotoneitai
("The p r o n o u n "ego" is only accented on the ultima")
2.4 he egoge Attikē estin
("The p r o n o u n "egoge" is Attic")
[[I.e. T h e forms can be mentioned instead of used, and in that case they take the
article; for A.D. and Aristarchus they are not then p r o n o u n s , but for Habron,
A.D. is claiming, they would be.]]
16. So a p r o n o u n is that which fills the place of a noun with deixis or anaphora,
and which is not accompanied by the article. Note that this definition is valid for all
persons, including the third. For third persons are specified [by pronouns] either
by back-reference (anaphora), insofar as they refer to known persons, or by
pointing (deixis), insofar as the persons referred to are in sight. And that consti
tutes the conjugateness of p r o n o u n s [i.e. what makes them a distinct syntactic
category]. In this respect, verbs are non-conjugate, for, although they are defined
for first-person and second-person forms, they are not definite in the third
person, with the marginal exception of verbs like astraptei ("is lightening"), since
such activity is referred unambiguously to Zeus, so that it's not really the verb that
is defined, but the person who performs its action.
17. So it is clear that the reason for the use of a multiplicity of forms [of
third-person pronouns] is to avoid having different referents for the same form.
For [if there was only one third-person p r o n o u n ] the consequence would be that
92 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
p r o n o u n s , too, would be indefinite in the third person, since for a single word to
signify several things is a cause of indefiniteness in speech.
18. (VI) Besides, all other nominals presuppose a nominative, and form the
genitive and the remaining cases on the pattern of the nominative, not excepting
even the definite article, since we have shown that it is inflected according to a 102
p r o p e r pattern (Bk. 1.76), but the simple personal p r o n o u n s [i.e. not counting the
possessives emos ("my"), sos ("thy"), hos ("his"), etc.] are inflected [in part] by
suppletion, since there's no way by the regular pattern of case inflection that a
genitive emou ("of me") can be derived from [the nominative] ego ("I"), nor, in turn,
from emou can you get emoi ("to m e " dative) or erne ("me" accusative) [by an
otherwise attested declensional pattern]; dual and plural are also necessarily
suppletive [no "we two", hemeis "we"]. For it's impossible, if the genitive is not
properly derived from a nominative, that dual and plural should be properly
based on the singular. [Only] when it [the genitive] is derived [from its nominative]
can it give rise to the other n u m b e r and case forms.
19. For this reason, in the case of heteroclites (nouns with two or more different
stems) where the genitive generates the rest of the forms, it also implies another
nominative (i.e. one from the same stem as the genitive), as from the genitive
megalou ("large") [other cases are] m a d e [including nom. pl. m a s c ] megaloi, and
from this basis the [regular] nominative [singular masculine] megalos is attested.
[[Actually megalos is not attested in Greek prior to A.D., but it eventually becomes
the Modern Greek form, and so A.D. might have heard it used. T h e classical form
is megas. However, A.D. may in fact be saying that megalos is the true underlying
form of the nom. sg. m a s c , from which megas is derived by deletion of -lo-.]]
Similarly, in the case of hudatos ("water" gen. sg.) and also gunaikos ("woman" gen.
sg.) and pollou ("much", "many" gen. sg. m a s c ) . For polloi (n. pl. masc.) and hudata
(n. pl.) and gunaikes (n. pl.) are formed [from the stem of the gen. sg.], and
therefore nominative singular forms polios, *gunaix and *hudas coexisted. [[Again,
most likely he means as abstract underlying forms. T h e normal forms are polus,
gunē, and hudor. Polios, however, does exist as a dialect variant, occurring quite
often in H o m e r . A.D.'s imaginary *hudas is formed on the pattern of keras, keratos
("horn") rather than soma, somatos ("body") or hepar, hepatos ("liver"), though the
type of soma is far most n u m e r o u s (and regular)—mainly e n d i n g in -ma—and the
type of hepar at least has the final -r of hudor.]]
20. T h e r e f o r e emou ("of m e " gen.), since it is not properly declined [[there are
indeed forms like emoi, erne, emon, emous, etc., but they are possessive adjective
forms, from a nominative emos "my"]], but only inflected by stem changes [[there is
n o noun-declension with a dative -oi or an accusative -e to justify cutting emoi and
erne; and clearly no and hemeis involve different stems]], did not provide the basis
for the other cases and n u m b e r s , (VII) although ekeinos ("that (one)") and its class
[i.e. the third-person demonstratives hode, houtos, and the anaphoric autos] have a
regular genitive-to-nominative relationship, and also form the rest of their de
clension regularly.
2 1 . Now it's clear that people who persist in slandering [the personal pronouns]
as irregular and illogical in form are themselves talking nonsense, since they d r a g
in the analogy of n o u n declension, when we can't even concede that every n o u n 103
B O O K II: P R O N O U N S 93
can be analogized from every other. But on such a basis one could go on subdivid
ing indefinitely. Words are naturally compared to others of parallel inflection,
and if that's so, how can it help being futile to compare different parts of speech? If
you can do that, what's to stop us, since personal p r o n o u n s , like verbs, mark
person, from comparing them with verb forms and so, once again, excluding them
from regularity [since erne does not resemble -o or hēmas resemble -metha, etc.].
22. Perhaps someone may wonder about the reason (a) why p r o n o u n s reject the
[regular nominal] declensions, and (b) why only some of t h e m do [namely the true
personal p r o n o u n s , but not the demonstrative a n d anaphoric p r o n o u n s of the
t h i r d p e r s o n ] . N o u n s ( n a m e s ) w e r e d e v i s e d for t h e i n d i c a t i o n of
certain qualities, general or particular, e.g. anthrōpos ("man"—indicates the quality
of being a man), Platon ("Plato"—indicates the quality of being Plato) [cf. 1.78,
3.96, a n d Ch. 41 below.], a n d since they indicate without either deixis (pointing) or
anaphora (back-reference), their use becomes very general, so that the distin
guishing form of each word can assign the [distinguishing] quality of each thing.
T h e homonymous application, both of common and p r o p e r nouns, considerably
disturbs the assignment of qualities, so that the entity intended by the n o u n may be
deprived of exact determination. A n d for this reason the nominative, being
specially coined for each entity [the word or n a m e for anything is originally given
or "imposed" only as a nominative], requires that the oblique cases belong clearly
to the identity of the nominative [i.e. follow a clear, regular declensional pattern].
26. As we have just established, the personal p r o n o u n s , since they either refer
back or point, have no need of the extra features (parepomena) of a n o u n (i.e.
gender); but this is not true of ekeinos ("he", "that (one)") and autos ("he h i m s e l f in
nominative, "him, her, it, them" in oblique cases). ([I omit houtos] because it has a
derivative suffix, not a pronominal one, as is shown by temoutos ["next, then" a rare
poetic and dialect word], an adverb with the same ending.) [[This sentence is
peculiar, to say the least, and Buttmann wanted to delete part of it. T h e r e is no
good reason to exclude hode ("this") a n d houtos ("this", "that") from this
explanation—they, too, have gender and are deictic, even if they are derived, in
some way, from the definite article.]] And it's not surprising that these two [four, if
we count houtos and hode] should be exceptions when all the rest (personal pro
nouns) have been found to follow the required pattern [of no inflection for
gender]. And it is perhaps not unreasonable that these strengthened their form
[by adding the gender contrast; really "increased the n u m b e r of forms" is the
point] to-specify the referent more precisely. T h e separation [in space] implicit in
these [mainly in ekeinos, "that at a distance"] makes the pointing force less exact,
and therefore they adopted gender, not to represent the essence [of the re
ferents], but to discriminate the gender [apparently h e r e meaning sex].(The
nouns that are said to be of common gender, [name objects which] when observed
from a distance do not permit accurate distinction of gender, e.g. hippos ("horse"
or "mare", d e p e n d i n g on the article ho or he), anthropos ("man" or "woman"). But
when they are closer, and plainly visible, their gender can be marked by the
article.)[That holds for ekeinos, then.] Even more so in the case of autos. For
instances of autos refer anaphorically to third person items which are absent [out 106
of sight, not just far away], and so by their inflection, r e m i n d us of the sex of
referents.
27. T h e remaining question is why these p r o n o u n s (ekeinos and autos) end in -os
[i.e. are declined like kalos in the first and second d e c l e n s i o n s , - o s , -e, -on, etc.],
when there is so much variety of nominal endings. Perhaps it's because -os (i.e. the
I.—II. declension pattern] is the most general. This is the reason why the articles,
BOOK II: PRONOUNS 95
too, which are constructed with all nouns, follow this declension, and even the
p r o n o u n s applied to indicate possession of nouns, i.e. emos ("my") and its class [sos
"thy, hos "his, her, its", hemeteros "our", humeteros "your", spheteros "their", noiteros
"of us two", sphoiteros "of you two", and variants], and likewise almost all of the
nominal interrogatives which expect all nouns as possible answers—poios ("what
sort? what (+ N)?"),posos ("how much? how many?"),pēlikos("how big? how old?").
I say "almost all" because of the word tis ("who?", neuter "what?"), although even it
is not exempt from the -os declension, since in Aeolic it takes the form tios with
nominative in -os u n d e r the influence of the similar word hos, with which it is
combined into hosts, as in Callimachus Epigr. 23 (Mair):
2.5 hostis emon para sema [phereis poda]
("you who [set foot] by my tomb . . . ")
or in the proverb:
2.6 hostis epi deipnon opse kletheis erchetai . . .
("the guest who comes late to dinner . . . ")
28. (VIII) Now we must discuss the cases where we have what is an article in
form, but by transformation a [demonstrative or third-person] p r o n o u n , as in
Iliad 1.12:
2.7 ho gar ēlthe thoas epi neas Achaion
("For he went to the swift ships of the Greeks.")
with ho for houtos ("he", "this one"), or [the frequent expression (e.g. in Iliad 1.84)]:
2.8 ton d'apameibomenos . . .
("and in answer to him . . . ")
or as in Odyssey 13.88:
2.9 hos he rimpha theousa . . .
("so she, r u n n i n g lightly . . . ")
or in Odyssey 2.206:
2.10 heineka tes aretes eridainomen . . .
("We vie with each other for her virtue . . . ") 107
where tes means tautes ("of that woman"), and there must also be assumed an
ellipsis of the article [[since Attic requires an article with every definite possessed
n o u n ; here the ideal underlying form is tes tautes aretes,—genitive of he tautes
arete—where tes, which agrees with aretes, is then deleted by H o m e r , and tautes is a
possessive genitive modifying aretes, and is poetically replaced by tes: "the her
virtue."]], and so, in general, in all such cases [of ho for houtos, etc.].
29. Cases of nouns and p r o n o u n s , which are used in the place of nouns, take the
same structural position with verbs, e.g. Truphon didaskei ("Tryphon is teaching")
like houtos didaskei ("This one is teaching"). T h e oblique cases [of both] receive their
structure from the nominatives [i.e. subjects], and the intervening verb forms
show the relationship in each case:
96 THE SYNTAX OF APOLLONIUS DYSCOLUS
[[A.D. fails to see that this is merely the generic article which also occurs with
nouns a n d adjectives. H e has discussed this pattern in Bk. 1, Ch. 44 and 111.]] and
similarly with the postpositive article (relative p r o n o u n ) :
2.16 hos an paragenetai anagindsketo
("Whoever comes, let him read.")
as is quite logical. For the article can combine with any n o u n and refer to any n o u n
at all, though it is a single thing, and that is the mark of an indefinite. So it is p r o p e r
for articles to apply either to specific defined entities, equivalent to p r o n o u n s , or
to the most indefinite, equivalent to tis ("some(one)"). For a sentence like:
2.15 ho peripaton kineitai
("The walker moves.") 109
does not differ much from:
2.17 ei tis peripatei kineitai
("If anyone walks, he moves." See Bk. 1.9)
nor does:
2.18 (cf.2.16) hos an elthei
("whoever comes")
differ much from:
2.19 ei tis an elthei
("if anyone comes")
[[2.19 is abnormal for ean tis elthei; the a r r a n g e m e n t ei tis an is also found in the
N . T . (I Cor. 7.5) and in Plutarch (Tib. Gracch. 15.4)]] If you add a n a m e (or perhaps
"a noun") to one of the above constructions, that will shift the article [including the
relative p r o n o u n ] to its primary meaning [so it becomes definite and particular
instead of generic]:
2.20 ho Dionusios peripaton kineitai
("Dionysius moves when he walks")
[and, for the "postpositive article"]:
2.21 Truphon koimatai, hos menei me dianastas.
("Tryphon is asleep, who will wait for me when he gets up.")
partic. passive fem, of erb "love"), heimarmene ("fate", pf. pass. part. fern. of
meiromai "be given as due", "be allotted"), in which cases one need only point out
the fact of the changes. For the phonological forms of words do not weigh as much
in classification as what is meant by them. [[This seems to be Stoic doctrine.]]
34. Likewise houtos [[which is declined nom. sg. houtos, haute, touto, gen. sg. toutou,
tautes, toutou, etc., nom. pl. houtoi, hautai, tauta, gen. pl. toutōn, etc.; so (a) it begins
with h- a n d t- just where the article does; (b) the following d i p h t h o n g is -au-
wherever the article form contains -e or -a-, -ou- where the article has -o-, -b-, or
-ou-]] since it is derived from the article hos [[he doesn't really mean the relative
p r o n o u n , since it has h- throughout, but he does consider that hos underlies ho]] in
its pronominal sense, not its articular one, joins the p r o n o u n s , (and it is not
non-derived, as some people have thought). This is clear from its other features
(i.e. its declension).
35. In all other cases a masculine nominative singular ending in -os, whether the
word is primitive or derived, is the base for a feminine nominative singular in -e or
long a, with the exception of these derivatives [in -outos]. H e r e the derivative form
is always based on the corresponding primitive form, according to g e n d e r [[i.e. the HO
feminine of the derivative comes from the feminine of the base, the neuter from
the neuter of the base, etc.]] e.g. from telikos we get telikoutos ("so big", masculine n.
sg.), and from telike, telikaute ("so big" feminine n. sg., with -au-, not -ou-), a n d the
same for the neuter forms. T h e same thing is true of tosoutos and toioutos [i.e. their
feminines are tosaute a n d toiaute]. A n d similarly also for houtos, since it doesn't have
a feminine *houte, but r a t h e r haute, from he. And also in the neuter, from to,
beginning with t-, comes touto, also beginning with t-.
36. T h e same thing holds in inflecting for case, From [nom. pl. fern.] telikautai we
don't get a genitive [pl.] *telikauton, or from toiautai *toiautbn, as is otherwise the
case for all feminines [i.e. that the genitive plural can be derived from the
nominative plural]: e.g. from autai ["same, s e l f n. pl. fern] we say autōn (gen. pl.),
and likewise from the unsuffixed [i.e. without the -outos suffix] forms toiai toibn,
telikai telikon. It is clear that we derive telikautai [nom. pl fern.] from telikai, a n d
telikoutbn (gen. pl.) from telikon. T h e same rule applies in the case of hautai ("these"
nom. pl. f.) from hai (definite article) but (since it belongs to the same derivational
pattern) not *hautbn [in the gen. pl.], as is the case with the parallel form without
h--autai, autōn ["self, same"]. No, [instead of *hautbn, the genitive plural] is toutōn,
from the base form ton [definite article], with initial t- as well [as -ou- rather than
-au-] which never h a p p e n s in the nominative, b u t always in the genitive. For [the
base forms, the corresponding forms of the definite article, are] hai and ton.
37. T h e r e is another derivative parallel to the above, with -de, which is somewhat
m o r e poetical [not really so in 5th a n d 4th century Attic], so toiosde from toios,
tososde from tosos, etc., which can be changed [form for form] into the derivatives in
-outos, toionde to toiouton, tosonde to tosouton; a n d similarly hode can be changed into 111
houtos, being itself derived by c o m p o u n d i n g from ho. [[The forms of hode can all be
m a d e simply by adding de to the definite article; the forms of houtos require a little
m o r e in the way of special rules, as was r e m a r k e d above.]]
B O O K II: P R O N O U N S 99
38. We should not, then, agree with H a b r o n who says that houtos is derived from
the article in the same way as opsinos ("late" adj.) is derived from opse ("late" adv.) In
other derivations this is a matter of indifference, but in the present case it is
impossible for the base form to be changed to another part of speech, since we
have shown how it requires not only the same part of speech, but even the same
gender and the same case a n d the same n u m b e r . Necessarily, then, we say that the
derivation [of houtos and hode] is from the pronominal use of the article, and just as
telikos ("so big") and toios ("such") are poetical, but telikoutos and toioutos are in
everyday use, in the same way hos (pronominal, "he", "this one") is poetical, as in
Odyssey 17.172-3:
2.22 hos gar ra malista
hendane kerukon
("For he was the best liked of the heralds.")
where the more everyday form would be houtos gar. . . ("For he. . . ").
39. We must also add that the nominatives [sg. a n d pl., m. and f.] are derived by
ellipsis of t [e.g.hoi from toi, etc.] as we proved above [1.80], in the book on articles.
Besides, the Doric forms tautai and toutoi are not really pleonasms [i.e. are not
m a d e by adding an initial t to the "true" forms] but have simply reverted to the
original underlying forms, as we showed above (1.81). [[I.e. since the underlying
forms of the article in nom. pl. m. a n d f. are toi and tai, Doric keeps those, a n d also
derives the pronominal forms toutoi and tautai from t h e m by the rule, whereas
Attic deletes the t's and the substitutes h-.]] T h e oblique forms [of ho and houtos]
agree in beginning with t-, likewise the nominative neuter, and the other nomina
tive forms agree in h-. In general, when is the d i p h t h o n g ou ever preceded by h- [as
opposed to being absolutely initial ou- ? Generally speaking when ho- initial words 112
[are lengthened to ou-], they lose the h-,holos/oulos ("all") horos/ ouros ("boundary").
[[Such ou- forms are Ionic, and Ionic lost all initial h- very early.]] Only houneka is
an exception, and it takes h- because of touneka. So houtoi would not otherwise
begin with h- if it weren't derived by ellipsis of t-. [See above, 1.81 etc.]
40. (IX) [Personal] p r o n o u n s have acquired their use in place of names not as
some people suppose, out of ignorance of names; for what ignorance of names is
there in Iliad 4.63? [Hera addresses Zeus]:
2.23 soi men ego, su d'emoi
("[Let us yield to each other], I to you a n d you to me.")
But obviously it is because names cannot be used in the first and second person, as
we will show. [[In fact, baby-talk usages like "Does baby love Mommy?" and pidgin
talk to foreigners, like "Injun give white m a n buffalo-hide" were marginally
possible in Greek, as they are in English, but extremely rare.]]
4 1 . Someone may say, "Don't we use su and se ("you") even when we don't know
the name?" T o this we must answer that this is true by accident, since even if the
names are known the pronominal construction is still obligatory, and the p r o n o u n
is still quite as much substituted for the name. T h e p r o n o u n conveys the force of a
p r o p e r name—I don't mean the name considered as a phonological shape, but its
meaning, that is the unique character of the referent. [[I.e., personal p r o n o u n s in
100 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
context, like personal names in context, have unique referents, are unambigu
ous.]]
42. For this reason [first a n d second person] p r o n o u n s deprived of a speaker
("pointing") and an addressee ("pointed at") [i.e. outside a real speech act situ
ation] are useless; as, for instance, when written down they are utterly indefinite,
because they have been dislodged from their natural basis. T h e r e f o r e it seems
quite natural that, in epistolary writings (letters), if the names are not prefixed 113
[[ancient style began a letter, e.g. "Alexander to Aeschines, greeting"]], the refer
ences [of first a n d second person pronouns] cannot be established. For when the
p r o n o u n s are written down for the acts which are attributed by the person writing
to himself and to the one to whom the letter is directed, it is necessary to have
written down the assignment of names [to p r o n o u n s ] , so that the p r o n o u n s can
carry out their p r o p e r deictic function. For a sentence like:
2.24 ego soi kai palai egrapsa
("I also wrote you some time ago. . . ")
refers to the pre-placed nominative [i.e. "Alexander" in o u r example] with the
word ego ("I"), and with the word soi('to you") to the dative which follows it [e.g. "to
Aeschines" in o u r example]. Obviously the reference of the p r o n o u n s becomes
indeterminate when the names are removed. [[This chapter seems on the verge of
a proposal (such as Ross's, 1970) for an underlying explicit performative, but with
n a m e s r a t h e r t h a n p r o n o u n s — " I , A r i s t o t l e , h e r e b y d e c l a r e to y o u ,
T h e o p h r a s t u s . . . "—but stops short of this.]]
4 3 . (X) T h e reason why names ["nouns"] cannot be used in the first and second
person, is that the naming process cannot be applied in the first and second
person. For the application of a name for the sake of identification to those with
whom we speak [[the text here, which apparently refers to babies, is puzzling, and
perhaps corrupt, though Priscian seems to have had the same text before him]]
does not take place, since that is the function of the second person form; nor do we
apply names to ourselves, since that reference belongs to the first person. It's
absolutely necessary for all nouns in all cases but the vocative to belong to the third
person. T h e vocative primarily converts nouns from third person to second by its
function of pointing to the person who is called by name [and, hence, spoken to].
44. This, too, is clear, that when we have discourse [using pronouns] with 114
someone, we must distinguish the two participants [first a n d second person]. A n d
apparently it isn't possible to use nouns for this purpose, since they are inherently
third person, but talk by definition moves from first person to second. A n d
therefore the [first and second person] p r o n o u n s were introduced, to perform a
function which it was inherently impossible for nouns to perform, as when we say
egō soi egrapsa ("I wrote you"). We have already explained (Ch. 24) the reason why a
single p r o n o u n can take over the referential function of all nouns.
4 5 . I h o p e nobody will object that third person p r o n o u n s are of no use, since
nouns are already able to be used [in the third person] and go on to say "In fact, if
nouns h a d been available for use in the first and second person, p e r h a p s p r o n o u n s
would never have been invented." We can answer such a person as follows: once
[first and second person] p r o n o u n s had come into existence for replacing nouns,
B O O K II: P R O N O U N S 101
they automatically brought with them their special deictic (pointing) force. T h e
persons signified by them were simultaneously plainly visible, and thus this word-
type [pronouns] took over the exclusive function of deixis, which accompanied
them in the conjugational transition from first and second person to third. It is not
just because nouns are not applicable in three persons that p r o n o u n s arose, but
because nouns have no share in deixis, which is inherent in p r o n o u n s . Both [nouns
and pronouns] occur together in Iliad 3.229:
2.25 houtos dAias esti pelorios 115
("And that man is Ajax the gigantic")
and both are necessary; for the deixis [of houtos "that man"] is required because of
Ajax's being seen [by both Helen and Priam], but the n a m e Ajax is required
because of the individual quality about which [Priam] had asked. [["Individual
quality" is a Stoic technical term. Cf. 1.78, 3.96.]]
46. (XI) Since that is the case, it is clear that Apollonios grapho ("I, Apollonius, am
writing") and similar sentences are openly ungrammatical [[in fact, they're not
really rare, and are unambiguous]], for no other reason than failure to agree in
person, since the name is third person, but the verb is first. It's grammatical if both
are third: Apollonios graphei ("Apollonius is writing").
4 7 . But this rule is not quite absolute. For verbs signifying being (huparxis) or the
assertion of a characteristic individual quality (cf.1.34, 2.45) permit the a p p a r e n t
ungrammaticality: eimi Odusseus ("I am Odyssus"), Truphon onomazomai ("My n a m e
is Tryphon"), and others of this class. However, it is not possible for a verb of
individual predication toaccept[as predicatenominative]a generic word capable of
referring to anything, as if we could say ego onomazomai ("I am called T " ) , ego
kleizomai (same meaning). [[And if the ego is not predicate noun, but subject, then
the sentences are incomplete "My n a m e is. . . "]] But obviously one can say ego eimi
("It is I"—the stock answer of one knocking at a door), for here only existence, not
the application of a unique individual quality [e.g. a name], is signaled, but merely
a generic quality, for which ego is appropriate. [[In fact, the sentence d e p e n d s on
the encyclopedic fact that we know the voices of many people. You don't say "It's
m e " to a total stranger.]] T h e same a r g u m e n t [as for ego "I"] applies to the other
[first a n d second] person [pronouns]. So it is clear t h a t verbs which do not forbid
the use of predicate nominatives draw those n o u n s over from third to first person,
as when we say Truphon onomazomai ("I am called T r y p h o n " ) , but not *Truphon
biazomai ("I, T r y p h o n , am being overpowered"—not really ungrammatical, but
Truphon has to be an appositive of the subject, not a predicate noun). For 'being
overpowered' is not a unique attribute of T r y p h o n , as his name is; therefore it can 116
take the p r o n o u n which is in general applicable for all NP's [ego " I " ] .
48. (XII) Next we must discuss the construction of p r o n o u n s with verbs. Oblique
cases of p r o n o u n s can be used freely with verbs, from which construction one can
determine the implied relationship of the pronominal referent [to the act or state
n a m e d by the verb]. [[I.e. any verb which governs a case of a NP, takes the same
case of a pronoun.]] But this doesn't apply conversely. For not all verbs require
oblique cases of nouns or pronouns, since some verbs are self-sufficient [i.e.
inherently intransitive], while others are deficient [i.e. require some complement].
102 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
For instance, peripatei ("he's walking") and ploutei ("he's rich") and zēi ("he's alive")
and other such verbs do not require any oblique complement. We'll present a
complete account of such verbs later on in the appropriate section (Bk. 3.155 ff.).
49. I am not unaware that many grammarians claim that the verbal construction
is more complete if it adds pronominal subjects—ego egrapsa ('7 wrote"), ego lalō ('7
am talking") etc., and as proof of this they offer the fact that the following type of
sentence is impossible without p r o n o u n s :
2.26 ego men paregenomen, su d'ou
('7 arrived, but you didn't.")
[[i.e. sentences with contrastive emphasis on the p r o n o u n s ; the ungrammatical
equivalent would be *paregenomen men, paregenou d'ou "I arrived but you did not
arrive", approximately.]] "For without p r o n o u n s the sentence falls apart." But I
don't believe that this is universally the case [i.e. in non-contrastive sentences], a n d
I rely not merely on poetical citations (since poetic structure freely uses both
deletion and insertion [i.e. whichever was correct, the other might occur as a
figure of speech quite often]), but on common everyday usage, the practice of the 117
best prose-writers, and, most of all, on the force of theory which must be applied
even about constructions which are not in the slightest doubt. [I.e. all of g r a m m a r
requires justification, not just obscure points.]
questions either with 'yes' or 'no' or with an indicative verb, since affirmation is
implicit in indicatives [and 'yes' means "I affirm it"]. I ask someone grapheis ("are
you writing?") and the answer is nai ("yes") or grapho ("I am writing"). And, 118
conversely, the answer [may be] ou ("no") [which] negates only the indicative force
[i.e. means "I non-declare it], [or] ou grapho ("I'm not writing"). But there'll be
another chance to discuss these things (Bk. 3, Ch. 88 ff.). Implicit also [ingrapho ("I
write")] is the singular n u m b e r , and yet grapho does not require the word heis
("one"). Clearly also the nominative of the p r o n o u n is implicit; and yet if none of
the things listed above is elliptically missing, neither is the p r o n o u n egō elliptical in
grapho. [[In short, none of these things in Chapters 50-51 is supposed to be present
in the "correct" underlying surface form (grammatically normal) from which
actual surface forms—e.g. in H o m e r — a r e derived by ellipsis, pleonasm, re
ordering, substitution and other "figures". O n e must assume, then, that there is a
deeper abstract underlying form where they are present.]]
52. "But isn't it possible to say ego grapho without any ungrammaticality?" As I
said above (Ch. 49), not as a general rule [only, that is, if the subject has contrastive
emphasis]. And in the cases listed above (50-51), when there is some special
reason, the implicit words may be overtly expressed, as in Iliad 24.243:
2.29 reiteroi gar mallon Achaioisin
("[You will be] more easier for the Greeks [to kill, now that Hector is
dead].")
For when the intensification is double, one stage being i n h e r e n t in reiteroi
("easier"), t h e n necessarily the o t h e r must be achieved by a d d i n g [mallon
("more")]. [[He seems to imply a justification like this: (a) "you were always easier
for the Greeks to kill than Hector was", (b) now you are even more so. This version
follows the text of Ellebode 8c Portus.]] And we can also say [in response to "Are
you writing?"] nai grapho ("yes, I am writing"), strengthening the assurance of the
assent by the cooccurence of the two. And we may also say heis anthropos peripatei
("One m a n is walking") emphasizing the singular both in opposition to a possible
plural ["not several"] and also in opposition to a complete negation, as in:
2.30 oudeis peripatei, oudena heuron
("No one is walking, I found no one.")
Similarly, when we are making an absolute [non-contrastive] assertion, we will say
philologo ("I am a scholar") philologeis ("you are a scholar"); but if we wish to express 119
a contrast with someone else, we will attach the [subject] p r o n o u n , which has as its
special force contrast a m o n g persons. For [a subject p r o n o u n ] is clearly not
needed merely to identify the [subject] person; that m e a n i n g it has in common
with the [finite] verb. That's why infinitives invariably need such pronouns, since
they inherently lack person.
56. I n the third person, however, it is always necessary to add either one of the
pronouns or that for which they substitute, nouns [really NP's], not to show the
case, since, again, the nominative is implicit, nor indeed the person, for that is
marked as third [in the verb form], but to avoid indefmiteness (i.e. ambiguity).
[[The referents of first a n d second person p r o n o u n s in a live speech situation are
always unambiguous, but everybody else in the world may be a third person.]] A
single verb form has to suffice for the infinite n u m b e r of third persons. Clearly a
B O O K II: P R O N O U N S 105
single form applicable to an infinity signals Undefined person, causing the mind to 121
cast about a m o n g many possibilities. So p r o n o u n s which are capable of referring
to distinct individuals are construed [as subjects] with finite verbs, so that their use
may remove the ambiguity from the verbs. So we say:
2.33 graphei hodelhoutosl ekeinosl autos
("This one/ that one,/ yon man,/ he himself is writing.")
[[Even these pronouns are more ambiguous than " I " and "you", and are in part, at
least, anaphoric, d e p e n d i n g on context; and graphei by itself means "he is writing"
where "he" rather unambiguously refers to the last third person masculine singu
lar mentioned in the discussion, generally the subject of the last preceding main
verb. A.D.'s a r g u m e n t is insufficient.]]
57. (XIV) Now we have shown how the nominative case forms must always be
accented. Next we must move on to the other cases, which are characterized by
double possibilities of accentuation according to the construction in which they
occur, except in cases where the phonological characteristics of the word exclude
the two possibilities.[[ This is so for most of the dual a n d plural forms, which
cannot be enclitic. Perhaps he is also thinking of certain poetical forms, e.g. min
("him, her"), nin ("him, her, it, them") which are always enclitic.]]
58. T h e most general reason for accentuation [of oblique cases of pronouns] is
when there is emphatic contrast of one person with another; for instance, in Iliad
1.20:
2.34 paida d'emoi lusaite
("Set my d a u g h t e r free/or me.")
T h e r e is implied contrast of emoi with the humin in line 18:
2.35 humin men theoi doien
("may the gods grant you. . . ")
A n d it's clear that in humin men theoi doien the accentuation of the p r o n o u n humin
[[in H o m e r there are also forms humin a n d ummi for this dative plural, a n d A.D.
regards the first of these, at least, as enclitic]] prepares the way for the contrasted
person, emoi. [[Really it's the men which prepares the way most explicitly, as A.D.
says in 53.]] T h e aforesaid reason is often subdivided into several more, all of
which can be reduced to the same one.
59. (1) Every p r o n o u n , they say, which is coordinated by a conjunction must be
accented:
2.36 Dionusidi elalesen kai emoi
("He spoke to Dionysius and me.")
2.37 Dionusion timai kai erne
("He honors Dionysius and me.")
However, if the p r o n o u n stands before the conjunction, it isn't necessarily ac
cented: 122
2.38 echarisatb soi kai Dionusioi
("He was kind to you and Dionysius.")
106 THE SYNTAX OF APOLLONIUS DYSCOLUS
[[Here "us" is subdivided into "you" and "I". It's interesting that he doesn't specify
the deleted verb, since the first one, with su "you" would have to be imperative,
and the second with egō "I", though logically one would expect an aorist subjunc
tive, might perhaps be a future indicative.]] Similarly in Iliad 4.62-3:
2.46 (cf. I. 23) alletoi men taut'epieixomen alleloisin
soi men ego su d'emoi.
("But let's give in to each other on this, I to you and you to me.")
We will later e x p o u n d a very detailed explanation along these lines in the a p p r o
priate place [presumably in Bk. 4, in the lost portion]. So let's come back to our
subject.
62. If the first nominal [subject] is conjoined [i.e. is preceded by kai "both"], it can
never be this nominal which is shared with the second clause, but only the verb
[predicate]. A n d conversely, if the verb is conjoined it can't be that verb which is
shared, but only the accompanying subject nominal. For instance, in:
2.47 kai Truphon dielexato kai Apollōnios
("Both T r y p h o n and Apollonius conversed.")
it is not Truphon which is shared, but dielexato ("conversed"). A n d conversely, in:
2.48 kai dielexato Truphon kai anegnb
("Tryphon both conversed and read.")
it is Truphon which is shared. T h e same rule applies in the case of p r o n o u n s . [[Of 124
course the linked elements may also be objects, prepositional phrases, adverbs,
participial phrases, etc., in which case the shared portion will include both subject
and verb, a possibility which A.D. doesn't consider here.]]
6 3 . But if the first clause doesn't have any conjunction [i.e. an o p e n i n g kai
"both", . . ], then it is impossible to predict which element will be shared:
2.49 Truphon peripatei kai Apollbnios
("Tryphon and Apollonios are walking.")
2.50 Truphon peripatei kai dialegetai
("Tryphon is walking and conversing.")
[After Truphon peripatei kai] either element may be shared [i.e. the next item may
be either another subject or another verb]. But if a kai is p u t on the beginning, then
the conjoined clause must be according to the theoretical rule [[that if an NP
follows the first kai, another must follow the second, but if a VP occurs there, then
a second VP is adjoined]]:
2.51 kai Truphon peripatei kai Apollbnios
2.52 kai peripatei Truphon kai dialegetai.
For it's ungrammatical to say:
2.53 *kai peripatei Truphōn kai Apollbnios
("Tryphon both walks and Apollonius.")
[[This claim seems to be exaggerated, since such sentences can be found.]]
108 THE SYNTAX OF APOLLONIUS DYSCOLUS
64. So it is agreed that a conjoined nominal [i.e. one preceded by kai "also" or
"both". . . ] implies another one in the same case [preceded by a second kai "and"],
as is obvious already in Odyssey 1.10:
2.54 eipe kai hemin
("Tell us also. . . ")
where kai allois ("as well as others") is implied. Or else, following Aristarchus we
may reject the conjunction as not belonging [i.e. regard the kai as r e d u n d a n t ] so
that no second entity (prosopon) in the same case need be understood from it.
[[Either the underlying "normal" form is kai allois eipe kai hēmin ("tell both others
and us") with ellipsis of kai allois, or else it is merely eipe hemin ("tell us") with
pleonastic insertion of kai.]]
65. So conjoined p r o n o u n s are accented because of the conjoined and con- 125
trasted nominal. Hence in Iliad 9.160 and 481 quoted above (ch. 59):
2.41 kai moi hupostetō
("and let him yield to me")
2.40 kai m'ephilese
("and he greeted me")
it is not the p r o n o u n s that are linked by kai, but the verbs, and the word o r d e r has
been changed [i.e. the underlying forms are kai hupostētō moi and kai ephilese me,
with kai immediately before the verbs]. For the conjoined contexts do not intro
duce [or imply] any n o u n phrase (prosopon) in the same case [as moi, me] to be
contrasted with the p r o n o u n s . T h e r e f o r e , since there is no contrast, the p r o n o u n s
reject their inherent accent [and become enclitic], since they are not in construc
tion with the conjunctions.
66. T h e same rule applies also to conjoining with disjunctive words [i.e. e "or"],
and in general to any word capable of taking a nominal complement construction.
I put it this way because a conjunction like hoti ("that", or "because") applies to
verbs [we would say it subordinates or nominalizes whole sentences], for which
reason it has no effect on p r o n o u n s , which are accented, therefore, only if the
context requires. T h u s we can say:
2.55 hōti se etimesa hubristhen [with se enclitic]
("Because I h o n o r e d you, I was injured.")
70. XVI) Oblique forms of pronouns, when construed with verbs [as objects or
complements], then indicate or imply two NP's, one the subject or actor (diatitheis),
the other the object or complement (diatithemenos) [signified directly by the pro
n o u n itself], and if they are put in the accented [non-enclitic] form, they then
presuppose still a third NP [one contrastively opposed to the object, "not X" or
"not anyone else"], as is shown by the difference between tupto se ("I'm beating
you", unemphatic) and se tupto ("I'm beating you" not anyone else). A n d in this way
the construction is subdivided [into the absolute and the contrastive patterns]; the
absolute pattern demands a postposed p r o n o u n form with enclitic accentuation,
while the contrastive pattern places the p r o n o u n in front, where it takes the long 128
form and the associated full accentuation, as in dos mot ("give me. . . ") and emoi dos
("give me. . . ").
7 1 . T h e following rule is also valid, that prepositions may be c o m p o u n d e d with
[nouns in the] nominative and with verbs, but form prepositional phrases only
with oblique cases [as objects]; this is obvious with p r o n o u n s [with which there are
no prepositional compounds], but with nouns it is sometimes possible to be misled.
For we say Huperbolon, Epikouron [these are accusatives of c o m p o u n d names
Hyperbolus and Epicurus, but might be misinterpreted as huper bolon "beyond a
net", epi kouron "against a lad"]. But if you examine them carefully, you will see that
they are not phrases with an oblique case, but oblique inflection from a c o m p o u n d
nominative. [[A.D. always regards the nominative as the base form of a noun, and
the nom. sg. masc. as the base form of any adjective. Only nominatives, then, are
directly created by compounding.]] Prepositions which are restricted to occur
rence with particular oblique cases [i.e. not all three cases, as with epi] make the
occurrence of a prepositional phrase clear, as in kata Ktesiphontos ("against
Ctesiphon", but never *kata Ktesiphonti), pro Aristarchou ("For Aristarchus" but
B O O K II: P R O N O U N S 111
never *pro Aristarchon). A n d the reasons for this will be explained in the section
dealing with their construction [namely that of prepositions; see Bk. 4, 18-20].
72. W h e n prepositional phrases occur with p r o n o u n s , they must take pre-verbal
position. [[I.e. in underlying structure; on the surface they may occur anywhere,
by "hyperbaton", re-ordering.]] T h e prepositions and the accompanying oblique
cases [of pronouns] which they govern naturally are pre-posed together. [[Pied
piping.]] A n d thus since the verbs are in second position after the p r o n o u n s ,
necessarily the p r o n o u n s acquire [contrastive] accent, since we have shown that it
is impossible for an enclitic to occur before a verb with which it is directly
construed. For we can say:
2.66 egrapsa soi lalesai Apollōnioi 129
("We wrote you to talk to Apollonius", where soi "you" is enclitic
and precedes lalesai "talk")
but the p r o n o u n is not governed by lalesai but by egrapsa, u p o n which it rests its
accent [the second accent on egrapsa is considered to have been thrown back by the
enclitic].
73. (XVII) T h e r e f o r e I believe that the placing of accented (non-enclitic) pro
nouns after verbs is a transposition, e.g. if you say elalēsas emoi ("You spoke to me");
the correct o r d e r is emoi elalesas, as in Iliad 17.556-7 [with some mixture of
16.498-9]:
2.67 soi men de Menelae katepheie kai oneidos
essetai emata panta
("You, Menelaus, will have shame a n d sorrow all of your days.")
For if the non-contrastive p r o n o u n cannot take the position of the contrastive one,
that is if the enclitic form can never occur before the verb, then conversely the
emphatic form should never be placed after the verb, since when postpositive
pronouns are shifted to the beginning, they lose their inherent phonological
character [of being enclitic] and change to take full accent. Cf. Odyssey 11.471:
2.68 egno de psuche me podokeos Aiakidao
("The spirit of swift-footed [Achilles], descendant of Aeacus, rec
ognized me.")
with 11.91:
2.69 erne d'egno kai proseeipe
("[The spirit of Teiresias] recognized me a n d addressed me.")
For h e r e [in 11.91] the accented form is not d u e to the sense [which is not emphatic
or contrastive], but merely d u e to the shift of position [to the beginning]. [[This
ordering rule is reversed in M o d e r n Greek.]]
74. T h e evidence of the phonology supports this. In o r d e r to make the syntactic
structure easy to grasp, the non-contrastive, postpositive p r o n o u n [of the first
person singular—emoulmou etc.] loses its initial e-, a n d since this—syllabic loss—is
impossible for the monosyllabic p r o n o u n s [of the second a n d third person singu
lar — sou etc. and hou etc.], they [all three] threw back their accents, a n d thus 130
became enclitics.
112 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
75. Someone may suppose that the plural p r o n o u n s violate this rule, since when
encliticized they do not lose their vowel; ekousen hembn Dionusios ("Dionysius h e a r d
us"). [[A.D. maintains that the first a n d second person plural p r o n o u n s distin
guished emphatic from unemphatic forms by the position of the accent: emphatic
hemon, hemin, hemos, ("us"), humon, humin, humas ("you-all"), with circumflex on the
ultima, and unemphatic - A.D. says 'enclitic' - hemōn, hemin, hemas, humon, humin,
humas, with acute on the penult. This view is hardly s u p p o r t e d by any other
evidence, and is rarely reflected in o u r texts, where the forms with circumflex on
ultima occur in all positions. Some grammarians and a few editors do follow A.D.'s
rule.]] But we said in the beginning that sometimes a characteristic phonological
defect makes a general rule seem to have exceptions, as for instance in the case of
three-gender adjectives which are prevented by phonological characteristics from
exhibiting three terminations. [[Nobody is sure what A.D. had in mind here. T h e
scholiast proposes gerōn "Old", an -nt- stem, but (a) it is inflected in all three
genders, and (b) if the neuter is rare, it is not for any phonological reason.
Buttmann proposes c o m p o u n d adjectives like exairetos "choice", for which a dis
tinct feminine form is excluded (more or less) by a rule " C o m p o u n d adjectives in
-os do not form a distinct feminine". This comes closer to being a phonological
rule, and most adjectives in -os are three-termination adjectives.]] I n this case [i.e.
plural pronouns], the r o u g h breathing [initial h-] blocks elision. A n d we can show
that plural [pronouns] necessarily begin with r o u g h breathings (h). But if the e (of
hembn etc) were shortened, then inevitably the vowel would become smooth [lose
the h-]. For p r o n o u n s which begin with short vowels never have initial h-.
[[Remember that, for A.D., the r o u g h breathing, which we transcribe as h-, was a
mere graphic device, corresponding to nothing in pronunciation a n d with only a
little m o r p h o p h o n e m i c effect.]] That's why the Aeolians shortened the initial
vowel to a in ammbn[since Aeolic drops all h's]. A n d t h r o u g h o u t it is initial e- which
is added or deleted in finite verb forms and p r o n o u n s , e.g. hebrbn [or hbrbn, "I
saw"], heeha [or heka, "I put"], eeipa [or eipa, "I said"], and likewise ebe or be ("he
went") ephe or phē ("he said"), and, a m o n g p r o n o u n s , heoi [or hoi "to him"], heou [or
hou "of him"]. A n d that's why it is that p r o n o u n s beginning with e necessarily lose
the e [when enclitic]. In the plural p r o n o u n s [of the first a n d second person] then,
it's enough to indicate encliticization by shifting the accent: ekous' hembn ("He
heard us") in the non-contrastive use, putting the accent on the first syllable. It's
impossible to throw it back to the preceding word [as a real enclitic does] because
of the three-mora rule. [[An enclitic cannot be more than three-moras in length,
i.e. three shorts, a long and a short, or a short a n d a long. Within a single word the
most that can follow the accent is a long and a short.]] A n d that's why only forms
like spheas("them") as in Iliad 4.284 etc. [along with sphebn ("of them"), sphi ("to
them"),sphea ("them" neuter, Ionic) and sphe ("them"), but not sphisi ("to them")
even though it is only two moras long] of the whole g r o u p of [plural] p r o n o u n s
throw the accent back [become enclitic], and not forms like hemas. [[All the hem- 131
and hum- forms are at least three moras long.]] In our book "On Accent" we treat
this in greater detail.
76. (XVIII) We must also explain that the fully accented, preposable p r o n o u n
forms may also be shifted from their normal position, since they are fully free in
their accentuation and spelling, and so do not block transposition. But the encli-
B O O K II: P R O N O U N S 113
tics, on the other hand, since they must lean u p o n a preceding word, to which they
shift their accent, necessarily cannot allow movement [to the beginning of a
sentence.] For that would deprive them of their characteristic features, since they
would have nowhere to place their accent. Hence certain forms are always enclitic:
sphin ("them" dat.), sphe ("them" acc.), min ("him, her, it" acc.), toi with initial t-
("you" dat. sg., Epic and other dialects), and Doric tu used as an accusative [not the
nominative], and the third person dual forms [sphoe, sphoin]. T h e s e never precede
verbs.
77. It is very silly for anyone to think that this basic theory [enclitic always after,
full forms always before the verb] is overthrown by counter-examples, on the
g r o u n d that enclitics d o occur before verbs a n d accented p r p n o u n s with preposi
tions are placed after verbs, e.g.:
2.70 semeron elalesa kata sou
("Today I spoke about you.")
and:
2.71 semeron se etheasamen
("Today I watched you.")
and, from poetry:
2.72 kai tu philippon etheken
("And he made you a horse-lover." - source unknown).
T h e reason for establishing this rule is to establish as normal the cases of reorder
ing ("hyperbaton") which have been overlooked because they occur in everyday 132
speech, and to keep us from thinking that this r e o r d e r i n g is something excep
tional occurring only in the poets. [[I.e. the theory of (76) is designed to establish
the underlying order; all apparent surface violations are then d u e to an optional
r e o r d e r i n g transformation. A.D.'s observation that enclitics do not begin sen
tences is correct, of course, but his reinterpretation of that as "position before the
verb" seems to d e p e n d on a notion of sentence o r d e r in which verbs are initial or
second all the time. This, of course, is not the case on the surface. J. Wackernagel
(1892) established the correct rule for H o m e r ; enclitic forms like moi, min, toi, etc.,
must always come after the first word of a sentence, separated at most by another
postpositive particle.]] T h e correct [underlying] o r d e r (katastasis) is:
2.73 semeron etheasamen se
("Today I watched you.")
with the adverb preceding the verb and the p r o n o u n in the correct position for
enclitics; and similarly the correct o r d e r for accented p r o n o u n s is:
2.74 kata sou elalesa
("I spoke of you.")
a n d for the poetic line:
2.75 kai tu philippon etheke
("And he made you horse-loving.")
114 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
78. Accented pronouns are also used whenever they express reaction to unex
pected unfriendly behavior after pre-existing friendship:
2.78 emoi etolmesen tauta poiesai
("He dared to do this to me!")
2.79 erne hupemeinen hubrisai
("He h a d the nerve to insult me!")
Clearly this agrees with what we said earlier (Ch. 58). T h e situation is like this: "he
ought to have injured someone else, not me, who deserved his respect". T h e same
a r g u m e n t applies in cases where there has been a preexisting injury: 133
2.80 dei emou horontos ekeinon aikizesthai
("He should torture him in my presence.")
i.e. not in the presence of anyone but me, the previously injured party. [[This type
of contrast is exclusive, with everyone else, whereas the examples in 58 and
following chapters are mostly contrastive in the narrow sense, opposing one
person to one other.]]
79. It is also said that the pleonastically lengthened forms are always accented —
that is emeio ("of me") and the rest [seio "of you",heio "of him"] — and this is natural,
first because of the contrastive force present in them, e.g. Iliad 6.356:
2.81 (cf. 2.57) heinek' emeio kunos kai Alexandrou henek' arches
("For my sake, bitch that I am, and for Paris's power.")
I am not claiming that the corresponding forms without the extra i [i.e. emeo, seo,
heo] can never be accented, for instance as in Iliad 6.328:
2.82 seo d'heinek' ante te polemos te
("Because of you [there is] war and battle.")
B O O K II: P R O N O U N S 115
This is clear both from the [contrastive] meaning and the conjunction heneka [since
conjunctions and prepositions require accented forms], as we have shown (Ch.
67). In the second place one should not assume either pleonasm [i.e. insertion of a
letter] or ellipsis [deletion of a letter] unless the complete form is used as a base [i.e.
emeo is the base, not meo], or else the change that is applied, insertion or deletion,
might be canceled by the other deficiency [starting from less than a complete
form]. There's no point in giving examples, since the facts are clear. A pleonasm
[letter-insertion] cannot be performed on p r o n o u n s unless they are complete to
begin with. A n d the complete form is necessarily the accented form. Naturally,
then, the resulting form after the phonological change must agree in type with the
base from which it is formed. T h e base forms are those with e [emeo, seo], and so the
pleonastic forms [emeio, seio] are always accented. I n the third place, no enclitic can
be longer than three moras, so how could emeio be enclitic? [It has four moras.]
Neither can seio, even t h o u g h it is three moras in length. For if the enclitic forms 134
have a smaller n u m b e r of moras, how could p r o n o u n s to which a mora has been
redundantly added reverse themselves and become enclitic?
80. Pronouns which have a pleonastic [redundant, inserted; r e m e m b e r , the h-
does not count] e- prefix, e.g. hee (for he "him" acc.), heoi (for hoi "to him" dat.) etc.
[the "etc." perhaps includes Boeotian heous gen., late Epicheoio or heeio gen., but hee
a n d heoi are the only Homeric forms] have far m o r e reason to require accentua
tion. For if p r o n o u n s which inherently begin with e- [emou, emoi, erne "me"], when
they become enclitic, d r o p the e-, as in dos moi ("give me"), akouson mou ("listen to
me"), obviously those which take a pleonastic e- will, conversely, reject encliticiza-
tion, as we have shown above (Ch. 79).
8 1 . T h e dual forms of the first and second person [no, ndi "we two" nom. or a c c ,
noin, nōin "us two" gen. or dat.,sphō, sphdin "you two" n o m . -acc.,sphdin, sphdin "you
two" gen. -dat.] are always accented [never enclitic] because of their recessive
accentuation [of course nō and sphd have only a single vowel]. For it has been
established that enclitic words always have their accent [when they take an accent]
as an acute on the ultima, either in fact or potentially (by "potentially" I mean cases
of circumflex on the ultima), so that a throwing back of the accent will put an acute
on the preceding word and delete it from the enclitic. But in the case of noin and
sphdin this is impossible. [ [ T h e r e are m a n y a p p a r e n t exceptions to A.D.'s
generalization—seo, heo, sphedn, sphisin, spheas,etc., all enclitic,—but we have no
clue as to how he got a r o u n d this fact. Possibly he assumed underlying forms seo,
sphedn, sphisin, etc., which obligatorily took their accent on the first vowel when
they were not enclitic.]]
82. T h e third person dual forms are enclitic, and [when accented] oxytone: sphde
("they/them two" nom.-acc.) sphdin ("them two" dat.-gen.). And what's the reason
why they are not recessively accented? T h e same as for eimi("I am") andphe~mi("I
say"), because they are inherently enclitic, although otherwise verbs are invariably
recessively accented. Interrogatives are never enclitic because of their initial
accent, whereas their corresponding indefinites are enclitic because of their acute
on the ultima. [[Interrogatives and indefinites are all paired in Greek: tines "who?"
(pl.), tines "some people", pds "how?", pos "somehow," pote "when", pote "once",
"sometime" etc.]] I do not hereby mean to claim that all indefinites are enclitic, but 135
116 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
a n d Iliad 1.2:
2.84 he muri' Achaiois alge' etheken
("which caused many troubles for the Greeks")
where the punctuation which immediately precedes the relative testifies to its 137
construction-initial position. [[In both cases an attributive adjective —polutropon
"wily" a n d oulomenen "deadly" — stands between the relative and its antecedent —
andra "man" a n d menin "anger"; hence the comma.]]
86. Personal p r o n o u n s are always accented when combined with appositional
autos ("self," intensive or refexive in meaning); [as in Ilian 16.12]:
2.85 ēe ti Murmidonessi piphaskeai e emoi autōi
("Are you revealing something to the Myrmidons, or me, myself?")
and 15.231:
2.86 soi d' autoi meleto
("You, yourself, must look after [Hector].")
and 14. 327:
2.87 oude seu antes
("nor of you, yourself)
a n d Odyssey 5.190-1:
2.88 oud' emoi autoi
thumos eni stethessi sidereos
("Nor is the spirit in my own breast m a d e of iron.")
T h e situation in the third person is not the same [according to some who say] that
accented third person p r o n o u n s are invariably used as reflexives [whether or not
they are combined with autos]. We will show later (Ch. 95-9) that this claim is false,
a n d that usage is not uniform, and that it is not the change of accentuation that
triggers the shift to reflexive meaning. So the form is enclitic [according to them]
at Odyssey 4.667-8:
2.89 alla hoi autoi
Zeus oleseie biēn prin hēmin pēma genes thai
("But may Zeus destroy his life before he becomes a problem to us.")
a n d at Odyssey 8.396:
2.90 Eurualos de he auton aressastho epeesi
("And let Euryalus pleas him with words.");
[and in both of these it is non-reflexive,] and it is accented [they say] at lliad 4.497:
2.91 amphi he paptenas
("brandishing a r o u n d him [self]" This is used reflexively, but lacks
an autos.)
118 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
poses the full word. [[And in all Attic writers, h u n d r e d s of years later, the two
occur side by side.]]
94. J u s t as, in words, excess (insertion,pleonazon) or deficiency (deletion, endeon)
[of letters] is shown by the patterning of syllable structure, so it is not arbitrary or 143
unnatural to determine excess or deficiency (leipon) in sentences, which are built of
words, by considering the syntactic patterning of the words. We showed above
[Bk. 1, Ch. 127] that, in the case of articles, one could supply the missing articles by
observation of the sentence. It is not strange, then, for the sentences we've been
considering to allow ellipsis (deletion) of [personal] p r o n o u n s , [since they can be
easily supplied].
95. Since first and second person accented p r o n o u n s are used as reflexives when
accompanying a verb in the same person, and also are used as simple personal
p r o n o u n s when the verb is in a different person, as in Iliad 9.335-6:
2.116 emeu d' apo mounou Achaionlheileto
("He took my prize from me alone of all the Greeks.")
and Iliad 24.754:
2.117 seu d' epei exeleto psuchen
("Since he took your life away from you.")
it is obviously necessary for the same thing to be possible in the third person as
well, and actual usage provides confirmation. F u r t h e r m o r e , when the accented
form [hoi] is read in Odyssey 22.214:
2.118 mnesteressi machesthai, amunemenai de hoi autoi
("to fight with [us] suitors and to defend him")
since the person defending is Athena disguised as Mentor, and not Odysseus, so
that the sentence means "Don't let Odysseus persuade you to fight against us and
defend him" it is quite reasonable that the p r o n o u n is accented to contrast with
"us". [[This last sentence, from " F u r t h e r m o r e . . . ." on, may be out of place here,
but sounds like genuine A.D.]]
96. I n the sentence of Iliad 5.800:
2.119 e oligon hoi paida eoikota geinato Tudeus
("Surely the son Tydeus begot is not much like him.")
hoi ("him") must be taken as reflexive, since the subject of geinato ("begot") is the 144
same person as the p r o n o u n . [I.e. has the same referent.] " H e begot a son not even
a little bit like himself." For obviously if it were the simple (non-reflexive) p r o n o u n
it would have to refer to a different person than Tydeus. Conversely, in Odyssey
17.387:
2.120 ptochon d' ouk an tis kaleoi truxanta he auton
("No one would invite [from abroad] a beggar who would ruin
him.")
99. "But", they go on, "the accompanying verb [egenonto "became"] is plural,
and the pattern h e r e is emoi egenonto, soi egenonto, hoi egenonto [they became for me,
they became for you, they became for him]." [So there would be no possibility of
"for h i m s e l f referring to the subject "they" of egenonto.] For this reason they
sometimes e m e n d hai. . . egenonto ("which [ships] became. . . ") to has. . . eteuxen
("which [ships]. . . he made"), in o r d e r to get a singular subject to agree with the
singular p r o n o u n hoi autoi, the whole passage then going:
2.127 has pasi kakon Troessin eteuxen kai heautōi
("which he m a d e as a bane to all the Trojans and to h i m s e l f )
or else, like Comanus, they assume hyperbaton [a r e a r r a n g e m e n t of word o r d e r ] ,
so that the underlying sentence r u n s :
2.128 Alexandroi etektenato neas kai heautōi
hai pasi kakon Troessi genonto
("He had built the ships for Paris and for himself, ships which
became a bane to all the Trojans.")
as if some natural necessity had established a law that accented forms of the third
person p r o n o u n s invariably required reflexive interpretation. If that were so, how
could they handle the so-called possessive p r o n o u n s [heos, hos, etc. "his"] which
cannot ever have enclitic application? We will show [in Ch. 102-3] with r e g a r d to
them that, remaining accented in every case, they admit both the simple ("his",
"her", etc.) and the reflexive ("his own", "her own", etc.) interpretation, the
accompanying verb being sufficient to determine the distinction, How, then, is the
reading in Odyssey 4.667:
2.129 alia hoi autoi
("But for him [may Zeus bring death].")
with enclitic hoi not an obvious error, especially since the contrastive person
appears immediately prin hēmin ("before [he causes trouble] for us")? T h e above
examples should be e n o u g h for a clear account of the situation. [[I.e., the rule is
that accented personal p r o n o u n s are regularly contrastive, especially if initial or
accompanying a conjunction, but need not be reflexive.]]
100. (XX) It is not a waste of time to consider the advantages of such a principle
in the third person, that accentuation always means reflexivity, and if the p r o n o u n 147
is not accented, it has to be the simple (non-reflexive) form. T h e constructions
with first and second person p r o n o u n s obviously are unambiguous, since you can
determine if subject and object are different persons or not right away; as in
Alcman fr. 30 E d m o n d s :
2.99 se gar hazomai
("For I respect you.")
compared with Odyssey 13.313:
2.100 se gar auten panti eiskeis
("You disguise yourself as anyone.")
BOOK II: PRONOUNS 125
a n d Iliad 11.828:
2.130 all' eme men su saoson
("But you must save me.")
compared with 10.378:
2.131 erne lusomai
("I will ransom myself.")
etc. [[He could also have added examples with objects in first and second and
verbs—i.e. subjects—in third, and vice versa.]] But in the third person situation
[where both object and subject are third person] it is often very difficult to
determine. Since there are many different "third persons", any one of which may
be indicated by a single word, whether p r o n o u n [[he, hou, hoi, min, etc. all may
m e a n "him", "her", "it", "himself, " h e r s e l f , etc. in Homer]] or verb form [[third
person singular endings -ei,-si, -e, -tai, etc. may refer to any "he", "she", or "it"]],
necessarily, because of this coincidence (sunemptosis) of forms, the patterns have to
be exactly the same whether there is a "transitive" construction [subject and object
different] or a "reflexive" one [subject a n d object the same.]
101. T h e p r o n o u n heo ("him", "her", etc.) a n d members of its class [including
perhaps minlnin "him", "her" and spheas etc. "them"] in the oblique cases are all
third person, and the verb kedetai ("he cares for") and all similar finite verb-forms
are also third person, and if you p u t t h e m together [heo kedetai] you've got two 148
third persons, one in the p r o n o u n and one in the verb. T h e result may be heo
kedetai with two different arguments (prosopa, actants) meaning autou kedetai ("He
cares for him") expressing the notion that someone cares for someone else. But it's
also possible for someone to care for himself, and it's clear that this won't be
expressed by any other construction than this same one; again we get heo kedetai
("He cares for h i m s e l f or "she cares for h e r s e l f ) . This makes it very difficult to
determine whether A cares for B or A cares for himself. Suppose that Eumaeus
cares for himself; then this won't be expressed any other way than (Odyssey
14.461):
2.131 epei heo kedeto lien
("since he cared very much for h i m s e l f )
or, if he cares for Telemachus, once again that will give:
2.132 epei heo kedeto lien
("since he cared very much for him").
It's for this reason that it would be a useful invention if accented [oblique]
p r o n o u n s were always intransitive, i.e. reflexive, and enclitic p r o n o u n s were
always simple, i.e. indicating a different third person.
102. That's the clearest account we can give of the theory [i.e. that while enclitics
may not often be reflexive, accented forms may be either emphatic or reflexive];
so we must always determine the sense not by studying the accent, but from the
sentence context, just as, in other types of indefinite ambiguity, distinctions are
m a d e from the sentence context, not from the presence of enclisis or accent. As we
said above, even if one should grant that the distinction of simple from reflexive
126 THE SYNTAX OF APOLLONIUS DYSCOLUS
of] a simple [non-reflexive] p r o n o u n , as, for instance, emos dialegetai philos ("a
friend of mine is talking"), sos hubrizetai doulos ("a slave of yours is being mis
treated:), or, from Odyssey 4.618 (=15.118):
2.135 hoth' heos domos amphekalupsen
("when his palace sheltered [me]")
which means "when his {autou) palace sheltered me", [i.e. with autou the simple
third person p r o n o u n , not hautou, the reflexive]:
keise me nostesanta
("when I went back there", a continuation of the same quotation),
or Odyssey 4.512:
2.136 sos de pou ekphuge hēras adelpheos
("Your brother escaped d o o m " )
which means ho adelphos sou exephuge ["your brother escaped" with sou, second
person genitive, non-reflexive], or Iliad 9.57:
2.137 emos de ke kai pais eiēs
("You could be my son, even.")
or Odyssey 2.131:
2.138 pater d' emos allothi gaies
("My father [is alive or dead] somewhere else in the world.") 151
a n d the same thing [non-reflexive] applies to all examples of this type [type 1,
where the subject is the possessed].
105. It is not a counter example that a verb may be construed with a possessed
n o u n [in the nominative of the same person as the possessor] if it is a verb of being
(existence, huparxis). But if it's any other verb, the construction is ungrammatical.
For instance, you can say emos eimi oiketes, which is analyzed as containing a
reflexive, emautou eimi oiketes, "I am my own servant." But it is impossible to say
*emos therapon grapho (*I, who am my servant, am writing''). H e r e you could change
to graphei, with the verb agreeing with the third person nominative [i.e. "A servant
of mine is writing"] O r else you could make it second person, provided you insert
the participle of "to be", emos on therapon trecheis ["Being my servant, you are
running"]. [[This participle is needed only for A.D.'s d e e p structure; it need not be
present on the surface.]]
106. F r o m this we can understand that whenever a verb has as subject the person
who is the possessor and as complement the thing possessed, then the possessive
p r o n o u n requires reflexive interpretation, if there is also some contrast, as in ton
emon huion didasko ("I'm teaching my son"), ton son huion epaideusas ("You taught
your son"), meaning necessarily "my own son" {emautou), "your own son" (sautou).
[[emautou "of m y s e l f a n d sautou"of y o u r s e l f are genitives of the reflexive pro
nouns.]] For the verb concerns not only the subject contained in the personal
ending, but affects also the person in the p r o n o u n . We have already observed that
if a finite verb is construed with an oblique p r o n o u n of the same person, it
necessarily requires reflexive interpretation, as in the eme phemi case [discussed in
128 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
Ch. 89-90) and the like. But it has also been shown that possessive p r o n o u n s
contain [a personal p r o n o u n in] the genitive case alone, and can be transformed 152
into such genitive p r o n o u n s [[in Greek this can always be d o n e ; in time the
possessives went out of use, and in Modern Greek only the genitives are used ]],
and that the actual case of possessive p r o n o u n s [i.e. not the i n h e r e n t genitive] is
the case of the possessions. I n effect one who says ton emon doulon epaisa ("I struck
my slave") is saying Hon emou doulon epaisa which must be transformed into ton
emautou [doulon epaisa]. [[The starred sentence is an abstract underlying form.]]
107. However, in case the person is not contrasted, the possessive may be
transformed into a simple enclitic personal p r o n o u n . If it's contrastive, it must be
reflexive. That's why only the reflexives are accented [i.e. perhaps, why Hon emou
doulon must become ton emautou doulon], for in their case contrast is obligatory. For
instance, one can say:
2.139 (a) pros tais emais thurais hesteka.
("I'm standing by my front door.")
or change it into:
(b) pros tais thurais mou hesteka
(same sense)
but what Menander (fr. 830 K) actually says is:
(c) pros tais emautou nun thurais hestek' egō
("I'm standing now by my own front door.")
and quite properly he uses the reflexive which goes with the occurrence of the
emphatic ego ("I") in hestek' ego; the implied contrast is "not by someone else's
[front door]."
108. This rule also holds good for third person possessives, a n d the same point
must be observed here that we said above should be noted [in the case of first and
second person pronouns]. Either the complement of the verb refers to the same
person as the subject, or it is directed at a different person, since, again, the same
word may apply to several persons, as in Iliad 20.234-5:
2.140 hon kai anereipsanto theoi Dii oinochoeuein 153
kalleos heineka hoio
("Whom the gods snatched u p to p o u r wine for Zeus, because of
his beauty.")
For the sense is *heneka tou autou kallous [["because of his beauty"—normal Greek
must say tou kallous autou, but A.D. several times uses this peculiar order, which, in
ordinary Greek, would be taken to mean "the same beauty." Most probably he puts
the genitive autou in this position as the correct underlying or deep position—as it is
for all nouns and non-personal p r o n o u n s the surface position—from which it is
shifted by a late rule.]], as the verb anereipsanto has theoi as its subject—anerpasan hoi
theoi Ganumede tou *autou heneka kallous ("The gods snatched u p Ganymede be
cause of his beauty"). If the accusative of Ganymede's name is transformed
(metalambanetai) to a nominative which will be subject of the verb [i.e. if the object is
made subject of a passive verb], then the transformation will make the c o m p o u n d
[reflexive] interpretation necessary, herpagē ho Ganumedes heneka tou heautou kallous
B O O K II: P R O N O U N S 129
("Ganymede was kidnapped because of his own beauty"), as is the case in Iliad
3.332-3:
2.141 deuteron au thoreka peri stethessin edunen
hoio kasignetoio Lukaonos
("Next he put a r o u n d his chest the breastplate of his brother
Lycaon")
"he" meaning Paris, of course; and since the action of the verb proceeds from him,
a n d the thing possessed is his brother, therefore hoio kasignetoio ("his brother") will
be transformed to heautou ("his own").
109. F u r t h e r m o r e , in Iliad 16.753
2.142 hee te min olesen alke
("And his courage destroyed him [the lion].")
the subject of the verb is the thing possessed, which requires interpretation of the
possessor as a simple p r o n o u n , he *autou alkē auton apōlesen ("his courage destroyed
him"). Now if the lion were the subject, and were acting u p o n the possessed thing,
courage, the sentence would go heen olesen alken ("he lost his courage"), meaning
ten heautou ("his own"). But if a different person were understood as subject of the
verb, e.g. if one were speaking of Heracles winning his struggle with the lion and 154
said heen alken olesen ("he destroyed his courage"), the possessive is again a simple
genitive p r o n o u n , ten alken olesen autou, i.e. tou leontos ("He destroyed his—the
lion's— courage").
110. F u r t h e r m o r e , in Odyssey 4.191-2
2.143 Nestor phask' ho gerōn hot' epimnesaimetha seio
hoisin eni megaroisin
("The old man Nestor used to say [that you were wise] whenever
we mentioned you in his palace")
there are two verbs with which hoisin eni megaroisin ("in his palace") can be
construed, and if it is the singular verb [phaske "said"] that has as subject the
singular Nestor, who is also the possessor of the palace, then the structure of the
sentence d e m a n d s a reflexive p r o n o u n , i.e. ephasken en tois heautou ("he used to say
in his own [palace]"). But if it goes with epimnesaimetha ("we mentioned") which has
a different subject in the first person [plural], in which the speaker includes
himself along with Nestor, the resolution must necessarily be to a simple p r o n o u n :
"when we mentioned you" hoisin eni megaroisin, that is en tois *autou,"in his
[palace]".
111. Similarly in Iliad 16.799-800
2.144 tote de Zeus Hektori doken
hei kephalei phoreein
("And then Zeus gave [it] to Hector to wear on his head")
H e r e there are three referents [hupokeimena, NP's, arguments, actants], (1) Zeus,
who gave something [actually a helmet] to (2) Hector for him to wear on (3) his
head. So if the dative of "Hector" is deleted, then Zeus becomes the possessor of
the head: "Zeus gave [it] to his head to wear", i.e. tei heautou "his own [head]".
130 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
115. In Iliad 23.387, where the possessive p r o n o u n can be read (hoi, accented):
2.150 hoi de hoi eblaphthesan
("But his [sc. horses] were hindered")
it was rejected because of the wrongly believed rule [that the possessive p r o n o u n
was always reflexive, which of courses is impossible in the nominative], though the
p a t t e r n is normal—hoi emoi eblabesan("Mine were h i n d e r e d " ) , hoi soi eblabe-
san("Yours were hindered"), hoi hoi eblabesan("His were hindered"). Evidently a
fact of Homeric style favored the other reading [hoi, enclitic, "for him"], the fact
that he very frequently uses datives for genitives [as we would put it, uses datives of
interest, or of advantage or disadvantage]. If it weren't for this enallage [use of one
form in place of another], they might have even rejected the verse on the g r o u n d
that the p r o n o u n was ungrammatical. In the same way, inlliad 3.243-4, we must
interpret:
118. I am quite well aware that in other dialects a distinction is m a d e in the form
[between the genitive of the personal p r o n o u n and that of the possessive pro
noun]. So I suppose that H o m e r , being aware of the ambiguity of the form, chose
to use nearly always a form from some other dialect for the personal p r o n o u n — 159
emethen or emeo, or emeio, or emeu, but never emou (and this can't be because he was
ignorant of the -ou ending, since he frequently uses all the cases derived from it),
whereas the genitive of the possessive he inflects like any other word in -os, always
using either the form in -ou (emou, sou, hou) or the Thessalian variant in -oio (emoio,
soio, hoio), like kaloio ("of the fair one"), as in Iliad, 14.118 (Odyssey 19.180):
2.153 patros emoio pater
("father of my father")
B O O K II: P R O N O U N S 133
case with their heads: Hektoreios chiton ("Hector's shirt" nom. sg. m a s c ) , Hektoreiou
chitonos ("of Hector's shirt" gen. sg. m a s c ) , emos philos ("A friend of mine" nom. sg.
m a s c ) , emouphilou ("of a friend of mine" gen. sg. m a s c ) . T h e obligatory concord
(akolouthia) is clear. Similarly in H o m e r , Iliad 9.543:
2.160 pater d'emos autik' o'istheis
("As soon as my father learned. . . ")
and Odyssey, 3.83:
2.161 patros emou kleos euru meterchomai
("I'm seeking the wide fame of my father.")
and also Odyssey 4.71:
2.58 tōmōi kecharismene thumōi
("O man dear to my heart!" see Ch. 68 above)
121. Maybe you'll say "Can't you have the genitive of a n o u n construed with the
genitive of the accented personal p r o n o u n in such a way that the form and
structure are ambiguous? After all, any declinable word can be p u t in the genitive 161
next to a n o u n in this way. We say:
2.162 Aristarchou agrou ho karpos estin
("The produce is from a field of Aristarchus'.")
and nothing stops us from constructing the phrase emou agrou karpos ("produce of
a field of mine") just like Aristarchou agrou karpos ("produce of a field of Aristar
chus' ")." But the minute you put a form like emou into that construction it becomes
unambiguously the genitive singular masculine (or neuter) of the possessive
adjective emos. For the genitive of the personal p r o n o u n , such as ego, cannot be
used as a possessive genitive unless it is the enclitic form, which is a characteristic
distinguishing feature setting it off from the possessives, which are always ac
cented. With the personal p r o n o u n the phrase has to be tou agrou mou ho karpos
estin("lt's the produce of my farm").
122. Perhaps someone will think that the proof is not complete, and that the
proposed rule is still dubious. T o him we can reply that if [emou] were the accented
genitive of the personal p r o n o u n [in these examples], its other characteristic
features would have to be present also. We noted (Ch. 120) that possessive
genitives are indifferent to [the case, g e n d e r and n u m b e r of] the possessed head
nouns, but that possessive pronouns must always agree grammatically with the
head noun. So we cannot say Hon emou agron ho karpos ("The produce of the fields
of me") or Hoi emou agroi sunebe eskaphthai ("It t u r n e d out that the field of me had
been spaded") or *tēs emou oikias ("of the house of me") nor any other such phrase
where the ending of the n o u n is different from that of emou, but we must alter the
forms to possessive p r o n o u n s agreeing with their heads: ton emon agron karpos
("produce of my fields"), toi emoi agroi sunebe eskaphthai ("My field t u r n e d out to
have been spaded") This shows unambiguously that when we say tou emou agrou ho 162
karpos ("the produce of my field"), the phrase is to be analyzed as containing the
possessive p r o n o u n , not the genitive of the accented personal p r o n o u n . It is also
possible to change the personal p r o n o u n to enclitic and rephrase the possessive
B O O K II: P R O N O U N S 135
appears that Zenodotus was not wrong about the p r o n o u n s , but about Homeric
usage. A n d for this reason we must prefer Aristarchus' reading, apo heo cheiri
pacheiei [with heo, the correct Homeric form, not hou],
126. T h e following readings are also attributed to Zenodotus, in Iliad 24.486:
2.165 mnēsai patros seio
("Remember your father.")
and 14.118 (Odyssey 19.180):
2.153 patros emeio pater
("father of my father")
where, contrary to the rule we have established, he uses genitives of [accented]
personal pronouns r a t h e r than possessives. (XXII) Now it is possible to defend his
readings here on the g r o u n d that it is not u n h e a r d of to use the genitives for
possession, since, as we said, possessives can be transformed into genitives. A n d
besides, there are other passages of H o m e r where the same construction occurs,
e.g. Iliad 4.343:
2.166 proto gar kai daitos akouazesthon emeio
("You two will first be invited to my dinner.")
and 4.174:
2.167 seo d'ostea pusei aroura
("And the earth will rot your bones")
and Odyssey 9.405:
2.168 ei me tis seu mela
("Unless someone [drives off]) your flocks [against your will].")
and Iliad 19.105:
2.169 hoi th' haimatos ex emeu eisin
("who are of my blood")
so that the readings are not totally against Homeric usage. For it's obvious that, if
these Zenodotean readings [in 24.486 and 14.118] are changed to possessives, the
four passages just cited will disagree. But if those four are acceptable, then the
Zenodotean pair should not be rejected.
127. One must object, however, to those readings that it is incorrect to apply an
accented [genitive of a personal] p r o n o u n in the possessive construction, but 165
rather either an enclitic or a possessive, as we have shown. A n d it's obvious that the
reading is all right in 2.168, ei metis seu mela broto~n,and similar passages, because
the enclitic p r o n o u n is used. A n d as for 2.166 proto gar kai daitos akouazesthon emeio
this has been incorrectly interpreted as a possessive construction. It is better to
take emeio ("from me") with the verb, u n d e r s t a n d i n g Homeric ellipsis of the
preposition peri ("about") [with daitos], as in Iliad 2.356:
2.170 tisasthai d'Helenes hormemata te stonachas te
("to get vengeance for the worries a n d sorrows we have suffered
about Helen")
B O O K II: P R O N O U N S 137
a n d Odyssey 1.40:
2.171 ek gar Orestao tisis essetai Atreidao
("For vengeance for A g a m e m n o n will come from Orestes.")
and in Odyssey 1.69:
2.172 Kuklōpos kecholotai
("He is angry about the Cyclops.")
and in many other places, so that the correct underlying form (logos) [of 4.343] is
emou akouazesthon proto peri daitos ("You two shall first hear from me about din
ner"), where the accented form is required by the insertion of i [to make emeio from
emeo].
128. "But" [you may say] "what about the accented forms in Iliad 4.174:
2.167 seo d'ostea pusei aroura
("And the earth will rot your bones.")
a n d in 19.105:
2.169 hoi th' haimatos ex emeu eisin
("and who are of my blood")?"
I n reply to this we can say that the only place where one can j u d g e a p r o n o u n
wrongly accented is where it has the possibility of being accented either way [i.e.
either emphatic or enclitic] and the reading gives it the wrong way. For instance, in
Iliad 1.396:
2.173 pollaki gar seo patros eni megaroisin akousa 166
("I often h e a r d your father in the palace")
an accented reading of seo is wrong, and in 14.328:
2.174 ho's seo nun eramai
("So I now love you")
it is wrong not to accent, because of the explicit contrast [between "you" a n d other
persons mentioned in the preceding context]. Likewise there is no violation of the
rule in 1.336:
2.175 hos sphoi proiei Brisēidos heineka koures
("who sent you two for the girl Briseis")
where the accented form occurs in a context requiring the unemphatic p r o n o u n s ,
equivalent to hos exepempsen humas ["who sent you pl." with humas, not humas, since
A.D. counts humas as enclitic]. H e r e it is a characteristic of the form [sphoi] that it
cannot be made enclitic, so no one can object to this reading. It would be consid
ered otiose to pile u p such examples, since we have already discussed (Ch. 55, Ch.
76) those p r o n o u n s which are necessarily enclitic or necessarily accented. We
showed (Ch. 73, Ch. 98) also that p r o n o u n s placed initially [in the sentence or
clause] are naturally accented, so that ex. 2.167 seo d'ostea pusei aroura cannot be a
bad reading. For is it capable of enclisis [in this position] and yet not enclitic? [No.]
Similarly we showed (Ch. 69) that prepositions require accented p r o n o u n s , so no
138 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
other accentuation [i.e. no enclisis] is possible in ex emeu {Iliad 19.105, as ex. 2.169),
but only the accented form.
129. But there is no such excuse for the Zenodotean reading emeio [in ex. 2.153
patros emeio pater] since a more appropriate form, namely the possessive [emoio
"my" gen. sg. m a s c ] is grammatical and metrical in this position; so it would be 167
foolish to change it. I n d e e d in general what causes ungrammaticality is nothing
else but deviation from the normal (kuriōteros) form which is called for by the
context.
130. While we're on the subject of Zenodotus' p r o n o u n reading, it's not out of
place to mention Iliad 1.8:
2.176 tis tar sphoi
("Who then [set] the two of them [to fight].")
Seleukos also prefers this reading [rather than sphoe "the two of them"] a n d so do
many other critics, citing as evidence various editions of H o m e r . T h e a r g u m e n t
they use [for preferring sphoi, which regularly means "you two", and is then
accented sphoi] runs like this. "Since the genitive-dative of the second person dual
p r o n o u n [sphoin] becomes the third person dual form merely by changing its
accent and becoming enclitic [sphoin], as we can see by comparing Iliad 8.416:
2.177 guiosei men sphoin huph' harmasin okeas hippous
("He will lame the swift horses that draw the chariots of you two.")
with 8.402
2.178 guiōsō men sphoin huph' harmasi. . .
("I will lame [the horses] that draw the chariots of the two of t h e m
[i.e. H e r a a n d Athena].")
therefore it is also necessary in the case of sphoi, that when accented it is second
person dual, but by changing the accent while keeping the spelling constant, it
should become third person." A m o r e general proposition is this, that wherever
first and second person p r o n o u n s have different endings, then always the third
person, too, is different. T h e mention of first and second person is a d d e d h e r e
because of [the nominative singular forms] ego ("I"), su ("you"), hi ("he/she").
[[Actually these forms are the only ones in which first person differs from second
in inflexion, whereas differences of the third person a p p e a r also in the dative
plural sphisi(n)—as opposed to hemin, humin or ammi, ummi—and the optional form
min (acc. sg. m. or f ) , as well as this accusative dual form, third person sphoe,
enclitic, vs. first and second nōi, sphoi or no, spho, and the gen.-dat. dual forms
which A.D. discussed earlier, sphoin second dual, sphoin, enclitic third dual.]] For
here, in just the way the second person [su] is unlike the first [ego], so the third [hi] is 168
unlike the second [su]. A n d by this a r g u m e n t one would expect sphoi with -i [not
sphoe.
131. But it is also possible to offer the following arguments in defense of the
traditional reading [sphoe]: (1) that it is more correct, to avoid the change of person
being accomplished merely by a change of accent [from sphoi to sphoi] while leaving
the segmental phonology unchanged, a p h e n o m e n o n which is otherwise not a
B O O K II: P R O N O U N S 139
regular feature of p r o n o u n s [except for the case of the dat.-gen. dual sphoin
discussed in Ch. 130]. For the sole function [of this accent variation otherwise] is to
emphasize and deemphasize the same person, not to change the person. (2) T h e
third person forms which have the same endings [as first and second person
forms] otherwise never have the same beginnings [i.e. stems], e.g. hemōn, humon,
sphon ("of us, of you, of them"), erne, se, he ("me, thee, him/her"), etc. So it is
implausible that a third person p r o n o u n should agree both in stem and in e n d i n g
with a second person form, and obviously, if forms with the same ending should
have a different stem, then forms with the same stem should not get the same
ending. Therefore sphoe is correct. (3) We show elsewhere [lost work; but cf. Ch. 3
above] that in verbs there is a kinship between second and third person forms [e.g.
legeis, legei or deiknus, deiknusi or lelusai, lelutai where the two singular forms differ
by a single letter, or the active and middle dual forms, though the plurals are quite
different—legete, legousi or legesthe, legontai, etc.], a n d it would be r e d u n d a n t to
document this here. But in p r o n o u n s it is rather the first and second persons
which exhibit kinship [me, se, mou, sou, etc., hemeis, humeis, etc., no, spho, etc.],
whereas the third person forms are either deficient or r e d u n d a n t , showing differ
ent particular stems, like ekeinos, houtos [both of which are demonstrative as well,]
min [a unique form], and many more. Although both first and second person
p r o n o u n s have a dual nominative [emphatic subject p r o n o u n ] , there simply is no
such form to match them in the third person. [In H o m e r , there isn't any third
person nominative at all, except for demonstratives or autos.] A n d while third
person plural reflexives are c o m p o u n d forms [heauton etc.], there are no parallel
compounds for first and second person plural, [but phrases hemon auton, humon 169
auton]. A n d the proofs of this point can be multiplied. T h e r e f o r e , there is no
necessity either for a first person dual to e n d in -e because of sphoe, or for a third
person to end in -i because of not.
132. Obviously the same arguments that support the form sphoe are arguments
against the third person form sphoin. H e r e we can say that it would be implausible
for this dative form to lack the characteristic feature of datives, the suffix -i,
followed by -n, since it's a dual dative. [[All gen.-dat. duals in Greek e n d in -in.]]
A n d if one considers the matter carefully, it will be realized that [originally] the
second person dual was lacking, and forms of the third were naturally transferred
to function as second person. For every p r o n o u n beginning with sph- [otherwise]
belongs to the third person: spheas ("them" acc.. pl.), sphisi ("for t h e m " dat. pl.),
spheteros ("their(s)" poss. adj.), sphas ("their(s)" poss. adj.). [[In fact, both spheteros
and sphos are occasionally used of other persons, but A.D. would n o doubt r e g a r d
those cases as enallage.]] A n d then the matter of accent was so arranged that the
accented form [sphoin] was transferred to the second person—quite reasonably,
since the second person contains deixis, for which accentuation is appropriate—
and the third person, which has no inherent deixis, was properly reserved for the
enclitic [sphoin.] This [lack of deixis] also explains why other unique third person
forms are obligatory enclitics—sphe ("them" acc.), a n d min ("him/her" acc.).
Moreover, it is not u n n a t u r a l that the p r o n o u n sphoe ends in -e, since the accusative
singular form shares that ending [i.e. he; and he might have also mentioned the
form sphe].
140 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
133. (XXIII) Next we must discuss emautou ("my own", "of m y s e l f ) which,
according to T r y p h o n , is c o m p o u n d e d with the same ambiguity as emou, either
from the possessive or from the genitive of the personal p r o n o u n . H e thinks that 170
when it occurs without the article, it is c o m p o u n d e d with the personal p r o n o u n -
emautou akouo ("I hear m y s e l f ) , sautou pheidei ("You are sparing y o u r s e l f ) — b u t
when the article is before it, it is c o m p o u n d e d with the articular possessive [i.e. tou
emou "of my X" masc. or neut. sg.], e.g. tou emautou philou ekousa ("I h e a r d my own
friend"), tou emautou oikou despozo ("I am master of my own household"). What
misled him, evidently, is the possessive structure and the presence of the article,
neither of which is the exclusive feature of the possessive p r o n o u n [but can occur
with any possessive genitive]. It is easy to show from his own examples that the
possessive p r o n o u n [emos, erne, emou] is not present in this c o m p o u n d [reflexive]
form [emautou], first of all from the articles themselves.
134. It has been established that each of the possessive p r o n o u n s [or "adjec
tives"] admits of an article in the same gender, n u m b e r and case [i.e. ho emos, hoi
emoi, toi emoi, ta ema, etc.]. How, then, can it be the possessive which is present in the
reflexive [emautou] which may be preceded indifferently by all articles—ton
emautou ("my own" acc. sg. masc), tous emautou ("my own" acc. pl. m a s c ) , tas
emautou ("my own" acc. pl. fem.), in essentially every gender, n u m b e r [and case]?
But since possessive genitives are acceptable with all different head nouns [things
possessed], obviously it must be the genitive [emou "of me"] of the p r o n o u n which
is present [in emautou], not the possessive p r o n o u n .
135. But if the mere presence of an article makes a form become "articular",
then nothing prevents the enclitic genitive from also being called articular—tous
philous mou ("my friends" acc pl.), tou philou mou ("of my friend" gen. sg.). Since
that is a foolish conclusion, so also is it foolish to suppose that the reflexive
p r o n o u n becomes "articular" by the mere presence of the article. For the article 171
which appears in the construction does not go with the p r o n o u n , b u t with the
object possessed [present or understood], with which it also agrees.
agathos daimōn ("good spirit") and the like which contain no possessive mark,
require complete inflection of both parts [i.e. the genitives are Neas poleos and
agathou daimonos, etc.; the first element in each case is an adjective which agrees
with the second]. But idioms like Korakos petra ("raven's rock") or huos kuamos
("pig's bean", the plant called "henbane") and others like them are declined only in
[the second part, here given in] the nominative. For if [the first part], the genitive,
were changed from its possessive pattern [i.e. p u t in another case], the phrases (ta
tou hupokeimenou) would lose their meaning.
137. Besides, if the possessive [emos, emou] were really present in emautou, the 172
phrase would involve two possessive relations, one that is inherent in the possess
ive form, just as in emos or emon, and the second due to the genitive case. This can
also be seen in possessives derived from names—Aristarcheiou agrouprosodos ("[the]
access road of Aristarchus' farm"), where both the farm and the road are pos
sessed. Simple possesion would be with the genitive—Aristarchou agros ("Aristar
chus' farm"), Aristarchou prosodos ("Aristarchus' road"). So also with the possessive
p r o n o u n : emos agros ("a field of mine"), tou emou agrou prosodos ("an access road of
my farm"). Both the field a n d the road are possessed. But in the case of emautou
there is no double possession, but only a single one. [[An English double possessive
of this type would be mine's—"Your book's cover is red, mine's is blue."]]
138. This was proved with regard to the genitive [of the reflexive p r o n o u n ] :
emautou agron eskapsa ("I d u g my own field."), emautou philoi anethemen ("I comuni-
cated [it] to my own friend"). [[But here and in Ch. 137 definite articles would be
normal: ho emos agros, he tou emou agrou prosodos, ton emautou agron, tōi emautou philoi.
In fact, A.D. established this as a rule in Bk. I: 4 3 , 102. Possibly he is here giving an
underlying form, to which articles will be a d d e d by a later rule.]] (XXIV) A
nominative corresponding to this genitive is impossible to find, either in grammat
ical theory or in actual use; this is not like other forms which are merely rare (some
who are ignorant of the reason for its nonexistence may accept such a nominative
as rare). I n d e e d it cannot occur in construction with a nominative; not only is
*emautos or *sautos impossible, but so are phrases like *emautou doulos etupsen (?"My 173
own slave struck [a blow]"). or *emautou doulos etupsa ("I, my own slave, struck");
h e r e the simple genitive of the personal p r o n o u n must be used with a nominative, 173
like any genitive [except the reflexive], e.g. ho doulos mou etupsen ("My slave
struck"). [[Why he uses the normally transitive tupto "strike" in these examples is
puzzling, since it introduces an extraneous anomaly. T h e point could be m a d e
more clearly with an intransitive verb like paregeneto "arrived". Probably he de
rived these sentences by t u r n i n g an accusative direct object "I beat my own slave"
into a nominative. But he continues to use tuptō also in the following chapters.]]
139. It's clear e n o u g h that it isn't the case [of emautou] that blocks this construc
tion [in emautou doulos etupsen or etupsa], but the presence of the c o m p o u n d
[reflexive] in it, whose function is to distinguish an action which relates one entity
[the subject NP, actant] to the same entity [as non-subject]. Anyone who says
emauton etupsa "I hit m y s e l f affirms a relationship coming from himself to him
self, as if he were saying ego erne auton etupsa ("I myself hit me m y s e l f ) . [[Here the
ego " I " and the auton " s e l f are contrastive and emphatic. A.D. does not assert, but
certainly suggests that such sentences underlie all reflexive sentences, via a "Re-
142 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
flexivization Rule".]] Since the verb which governs the oblique [reflexive] is
necessarily in the same person [i.e. has a subject of identical reference], it's
impossible to add another entity [NP, actant] in the nominative [since there is a
subject already present]. But now [as of Ch. 138] we're not talking about accusa
tives or datives, which do not require another entity [NP, actant] to modify, but
only about the possessive genitive, which is in the person of the possessor, and
modifies the possessed [in the nominative]. T h u s it is incapable of co-occurring
with a nominative [referring to] that person [the possessor], since the possessed,
being nominative, compels the verb to agree with it, as is clear in the case of the
other constructions, Aristarchou gnōrimos anegno ("An acquaintance of Aristarchus
read"), phils mou dialegetai ("A friend of mine is talking"). T h e verb that co-occurs
with the two [possessed and possessor] cannot simultaneously agree with both,
since they are different persons, e.g., emautou is first person, and with it [as it is
reflexive] the co-occurring verb must agree in person: emautou etupsa paida("I hit
my own slave"). But if the accusative in this construction is changed to nominative,
it won't allow any other verb than a third-person form—pais etupse ("A slave 174
struck"). A n d therefore neither *emautou etupsa pais ("Of myself, I hit, I the slave")
nor *pais etupsen emautou ("A slave hit of m y s e l f ) is meaningful.
140. T h a t the difference of reference [of the reflexive and the subject] is the
cause of the ungrammaticality becomes clearer if the nominative [i.e. subject]
which was previously third person is p u t into construction with a verb of being
(existence) and shifted into the first person: eimi doulos ("I am a slave": of course
this makes it a predicate nominative, not a subject, but A.D. has no words for these
categories). A n d if we p u t emautou ("of m y s e l f ) with this, the structure is
grammatical, as the verb is no longer one marking a relation between two
referentially distinct entities. Similarly in the second person sautou ei huperetes
("You are your own servant") and the third heautou estin epiboulos ("He's his own
secret enemy"/ or, reading epebolos, "He's in control of h i m s e l f [Uhlig emends to
epibolos, otherwise unknown.]) In this structure, if you replace the reflexive by a
simple personal p r o n o u n , the nominative may apply to a different person, with
which the verb will agree: epiboulos/ epebolos autou esti ("He's his [i.e. someone else's]
secret enemy", " H e ' s1 in his2 possession"), doulos mou ei ("You are my slave"). So
much for this structure. [[A.D. has not yet stated a rule, but the implicit one is "A
reflexive may be in any oblique case, but must be coreferential with the subject of
its clause." H e does not say that reflexivization is obligatory.]]
simple p r o n o u n s , which accept effects from another entity [i.e. NP], and (2) the
c o m p o u n d [i.e. reflexive] p r o n o u n s , which t u r n the influence from the same
entity back u p o n the same entity, the influencing a n d the influenced entities
[being identical]. For this reason they are also called "reflexive" p r o n o u n s ,
metaphorically from things which reflect back to themselves, or else "self-
influenced" (autopathes). T h e simple p r o n o u n s , conversely, are shown to be
"other-influenced" (allopathes), i.e. affected by another entity. T h e [compound
forms of the reflexives] are very a p p r o p r i a t e to the union of the two things
simultaneously included. For just as the influencer (diatithen) and the influenced
(diatithemenon) a r e a single e n t i t y , single also is t h e f o r m p r o d u c e d by
compounding—emauton epaisa("I struck m y s e l f ) , hup' emautou eplegen("I was
struck by myself"—note that A.D. has no objection—pace Postal (1971)—to such
sentences). [[The sentence [earlier in the chapter] i n t r o d u c i n g the passive
transformation is a little sketchy. T h e topic comes u p again Bk. 3, Ch. 157.]]
142. "If the oblique cases require a nominative, so that the influence can be made
clear as coming from someone," someone may say, "how can the oblique case
present in sauton ("yourself [acc.]") help requiring a nominative as source of the
influence on it?" [[The mss. contain a line or two in the middle of this that is
rejected by Uhlig: "How will sautou become sautos? W h e n it does, it will require an
oblique case for the structure to rest on." It may belong somewhere else.]] T o him
we can reply that the nominative is inherent in the verb, sauton epaisas, that is, su
("You hit y o u r s e l f ) . And if you change it to *sautos epaisas you will have two 176
nominatives [i.e. subjects], one from the p r o n o u n and the other from the verb, for
which reason the structure is impossible unless one of them is changed to an
oblique form. T h e one present in the verb has no case-ending, since verbs are not
declined, but the one represented by a p r o n o u n can be declined [so it has to be the
one to become oblique].
143. Someone may then go on to argue, "if the nominative is inherent in the
verb, so also in *emautos or *sautos the oblique case may be present in the initial part
of the p r o n o u n s , *emautos epaisa representing an underlying erne autos epaisa ("I
myself hit me") [and exactly this pattern occurs in Latin—me ipse perctio], since we
can certainly say it without the intensive p r o n o u n autos, [as in Iliad 13.269]
me phemi. Therefore there's nothing to stop constructing the sentence *emautos
epaisa." But in reply to this we can say that the understood oblique case is not just
one, but three. Why is it any more erne (acc.) that is present in *emautos than emou
(gen.) or emoi (dat.)? A n d all these cases are necessary to make the grammatical
structure clear, but their identity disappears in composition. It would be too much
to prefer the ambiguous to what is both clear and ordinary [in usage].
144. But perhaps someone will object that "ambiguities are all solved by the
context; since verbs that appear with nominal complements cannot be combined
with every oblique case, but some take the genitive, others the accusative, a n d still
others the dative, the uncertainty about case [whether erne, emou, or emoi in
*emautos] will always be cleared u p by [knowledge of] the constructions (suntaxis) of 177
the verb. For instance, given *emautos akouo, akouo [since it governs only a genitive]
testifies that the implicit case [in *emautos] is genitive [so it means "I hear m y s e l f ]
O r take *emautos tupto; because of tupto it is clear that *emautos contains the
144 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
accusative, erne ["I hit myself']". But the following a r g u m e n t will now raise its
head: some verbs may take two different cases: temnō se ("I am cutting you."), temno
soi ("I'm making a truce with you"), gumnazo se ("I p r e p a r e you"), gumnazo soi ("I
p r e p a r e for you"). And in such environments how can *emautos distinguish be
tween cases? You can't help things by using articles, either. But the traditional use
of the c o m p o u n d [reflexive] pronouns will distinguish emauton temno ("I'm cutting
m y s e l f from emautoi temno ("I'm making peace with m y s e l f ) .
145. T h e clinching argument, even if you reject all these proofs, is that the whole
investigation comes to nothing, because even if emautos were possible, it would be
indeclinable. For, according to the rules of c o m p o u n d i n g it cannot be inflected in
the middle of the word after thee m - ,since once a nominal stem is c o m p o u n d e d [as
first member] it becomes indeclinable. But neither can it be inflected at the end,
for the nominative is always required as subject of the verb, as in ego akouo sou ("I
hear you"), ego etimon se ("I h o n o r e d you"), ego soi charizomai ("I do you a favor").
Therefore, won't *emautos be indeclinable, at the end because of the verb, at the
beginning because of composition? A n d if that is correct, how can it help being
futile to look for a nominative oi emautou? If it could be, then the nominative 178
couldn't have any oblique cases, and it is also clear that the actual cases, which are
oblique, can't have any nominative.
146. (XXVI) Next we must add the reason why the c o m p o u n d [reflexive] pro
n o u n can never have a nominative. Suppose you have a construction ego autos
hubrisa ("I myself injured. . . "), and, in the passive, ego autos hubristhen ("I myself
was injured"), in which constructions it is agreed that the relationship requires two
persons, one to do the injury and one to be injured. (For if the relationship were
singulary, from one person to the same person, so that you might say ego autos
hubrisa erne ("I myself injured me"), it would turn the construction into a com
p o u n d oblique [i.e. reflexive], emauton hubrisa ("I injured m y s e l f ) . A n d it's also
clear that c o m p o u n d words in general were devised to mean something different
from the same elements forming a phrase. For if c o m p o u n d s had the same
meaning as the corresponding phrases, then the c o m p o u n d i n g construction
would be totally unnecessary. There's a difference between chēōn kai alopex ("goose
and fox") and chenalopex ("fox-goose", "Egyptian goose"), b e t e e n mison gunaika
("who hates a woman") a n d misogunes ("woman-hater"), and so on. If then ego autos
c o m p o u n d e d could mean something m o r e than the phrase ("I myself, emphatic),
then compounding would be necessary. But in fact that is not so. Wherever there
could arise some difference by compounding, there c o m p o u n d i n g was appropri
ate. T h e oblique cases of the simple personal p r o n o u n are always used in con
struction to mark a person different [from the subject], whereas c o m p o u n d e d 179
p r o n o u n s [i.e. reflexives] abjure this change of person [and refer only to the same
person as the subject]. Which is not possible, as we have said, in the nominative
case [which is the subject]. Necessarily, then, the nominative was excluded from
composition [i.e. reflexives have no nominative]. [[This whole discussion seems so
vague a n d fumbling to us that we want to come to A.D.'s help. I n this Chapter he
has overlooked the fact that the reflexives of the first, second, a n d optionally, the
third person plural are phrases, not compounds—hemon auton ("of ourselves"),
humon auton ("of yourselves"), sphon auton or heautōn ("of themselves").]]
B O O K II: P R O N O U N S 145
147. (XXVII) Not only do the [reflexive] p r o n o u n s take the first p a r t of their
c o m p o u n d form from the oblique cases, but also the word allelon ("each other",
"one another" gen. pl.; we would call it a reciprocal p r o n o u n , but for A.D. it
belongs among the nouns, since it is neither deictic nor personal, t h o u g h perhaps
anaphoric). For it, too, cannot form a nominative, [and hence uses its genitive, the
next case in order, as base form]. T h e construction with a verb makes clear the
passage of an effect between two persons. T h e sentence allelon ekousan ("They
heard one another") means, essentially, alloi allōn ekousan ("Different ones h e a r d
different ones", "Some h e a r d some, others h e a r d others", etc.), or allelous epaisan
("They hit each other") means alloi allous [epaisan]. [[Of course this isn't strictly true
in classical Greek; but it is etymologically true, in a sense.]] If we constructed a
nominative *alleloi, it would mean alloi alloi ("others [m. nom. pl.] others [m. nom.
pi.]") which is meaninglesss. For the sense is left incomplete because of the
simultaneous presence of two subjects [and no object]. [[ Conceivably it might be
possible to construct a scenario such as we can make for English: Groucho Marx
came to a party disguised as H a r p o , and H a r p o came disguised as Groucho, so, for
this party "Groucho and H a r p o became each other"—alleloi egenonto. But no such
sentences are attested. A.D. also overlooks the fact that reciprocal sentences do
have overt subjects, so he's going to get two nominatives anyway. T h e notion of
apposition does not seem to have occurred to him yet, and this is partitive
apposition, as in Bk. 1, Ch. 57 and 156, which he explains by enallage—nominative
for genitive.]]
148. T h e meaning is not the same as that of the pronominal c o m p o u n d [i.e. the
reflexive]. For that signifies self-influence—heautous etupsan ("They beat them
selves") whereas the other marks the passage of influence from one set of persons
to another. For instance, the phrase in Odyssey 16.293 (=19.12):
2.179 allelous trosete
("[for fear] you may wound each other")
is not the same as heautous trosete ("you may wound yourselves"), and, similarly, the
passage of Pindar (fr. 163 Schr.) was m a r k e d as exceptional by T r y p h o n — t h e
passage dealing with Otus and Ephialtes, where T r y p h o n accepts as normal:
2.180 allalophonous epaxanto lonchas
("They planted their internecine spears.")
but not the next bit:
2.181 eni sphisin autois
("in themselves", i.e. "in each other")
because they didn't fix their spears in themselves, but in each other. [[This usage,
reflexive with reciprocal sense, is not unusual in Greek, and is c o m m o n in other
languages as well.]]
149. It's clear that the construction [with allelon] which requires both an oblique 180
case [complement] and a nominative [subject], necessarily c o m p o u n d e d the
nominative on the beginning so that it would remain unaltered, according to the
phonological rule of composition and because of the accompanying verb [which
always requires a subject]; but the end is inflected according to the requirements
146 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
of the construction. I n a sentence like allelous etupsan ("They beat each other"),
etupsan is linked to the beginning of the c o m p o u n d [which is its subject], alloi
etupsan allous, and since the different oblique cases are needed to p u t the verbal
complements in their grammatical form, they are marked on the inflected part,
the end part, so that, as required, with ekousan("they heard"), a genitive [allelōn] is
grammatical and with verbs like etupsan ("they hit") an accusative, allelous ("each
other"), and finally, with echarisanto ("they did favors") [dative], allelois ("for each
other"). It's obvious again [as in the case of the reflexive] that if the o r d e r of the
parts of the compound were reversed, the word would then be indeclinable, since,
once again, neither can the first member be inflected, nor the subject of the verb
be put in any other case. [[This is true of finite verbs, but with infinitives and
participles subjects do a p p e a r in all cases.]]
150. (XXVIII) Since the compound [reflexive] p r o n o u n of the third person also
has a plural form [made from the singular], but the first a n d second persons do
not [but use instead the phrases hemon auton, humon auton], we must also say
something about this fact. It is said that Aristarchus objected to the use of the
c o m p o u n d third person forms, arguing that it was i m p r o p e r for an inherently
singular form [i.e. the first member, he-"him/her"] to be transposed into the
plural, giving you singular heauton ("himself) and plural heautous ("themselves"
but looking like "*himselves"), supporting his position by the usage in the first and 181
second person, where the nonexistence of these pluralized c o m p o u n d forms [i.e.
*emautous, *seautous], shows that they shouldn't be allowed in the third, and also
citing Homeric usage. For parallel to hemeas autous ("ourselves" acc.) and humeas
autous ("yourselves" acc.) [Homer has] in the third person spheas autous. [[Such
forms also occur in Attic prose—sphon autōn, sphas autous, etc.]]
151. But Habron tries to refute these arguments with evidence from usage,
including quotations from Plato; and to show that it is acceptable to form plurals
from compounds containing a singular first element he cites the c o m p o u n d
hendekaton ("eleventh" acc. s. m.) [compounded "one" + "tenth"—actually a de
rivative of a c o m p o u n d hendeka "eleven", "one" + "ten"] which forms a regular
plural hendekatous ("eleventh" acc. pl. m.). "It is not necessary", he says, "for
everything that exists in the third person to be present also in the first and second",
that is, the analogy of heauton and heautous ("themselves" does not necessarily
entail *emauton, *emautous *myselves [i.e. "ourselves"], "since what h a p p e n s in the
first and second persons doesn't always h a p p e n in the third; for instance, there are
nominative dual forms in the first and second person, but not in the third." T o this
he might also have added that such third-person p r o n o u n s as ekeinos ("that (one)"),
autos("self', "same", "him/her/it/them"), houtos ("this/that (one)") a n d min ("him/
her") have no corresponding words in the first and second person [though he
"him/her" is matched by se "you" a n d erne "me"]. [[It is true that autos and houtos
may be used of first and second persons on occasion, but only in apposition with
overtly marked p r o n o u n s or verb forms.]] (b) All duals of the first and second
person are accented, while the third person duals are obligatory enclitics. (c)
T h e r e are possessives from first and second person duals [nōiteros "our(s), of us
two"', sphditeros "your(s), of you two"], but not third. A n d he also says that "usage 182
has by chance rejected some [possible] forms, and there's no reason to suppose
that this is not what has h a p p e n e d with the pluralized c o m p o u n d p r o n o u n s [of the
first and second persons]."
B O O K II: P R O N O U N S 147
152. (XXIX) O n the side of Aristarchus we may say that H a b r o n hasn't properly
understood Aristarchus' point. For he was not unaware of hendekatoi ("eleventh" n.
pl. masc), nor that a plural may be formed from c o m p o u n d s containing a singu
lar. But if heauton ("himself') consists of two singulars [i.e. "he" a n d "him"], a n d the
two singulars refer to the same single person [i.e. "he 1 . . . him 1 "], indicating the
influence of self u p o n self, how can there be indication of plurality for such a
word, in which the initial portion of the c o m p o u n d is in the singular n u m b e r and
the final part is plural ["i.e. "♦he 1 . . . them 1 "]?
153. Nor is hendekatos, hendekatoi relevant. For there is reference to two entities
[NP's, referents] in hendeka ("eleven"), "after ten, one more". So also hendekatoi
("eleventh" nom. pl. m a s c ) ; again the plurality next in o r d e r after tenth in
counting has a right to be called hendekatoi ("eleventh" nom. pl. m a s c ) . "But"
someone will say, "how is it possible, if "one" is present in the word, that it can
become significant of a second entity in the plural? T o this it must be a d d e d that
these words ("eleventh" etc.) are ordinals. Anyway, c o m p o u n d s referring to two
entities can signify a plural entity when containing a singular m e m b e r and also a
single entity though c o m p o u n d e d with a plural member. For instance, philan-
thropos ("lover of mankind") though c o m p o u n d e d with a plural element [-anthrōp-
here referring to all m e n ] , is singular in meaning, whereas, conversely, the plural
c o m p o u n d philoplatones ("lovers of Plato") contains the singular element Plato. So 183
it's not at all unusual that hendekatoi ("eleventh" nom. pl. masc.) signifies a plurality
though containing a singular; for they are different entities. But this is not the case
with heauton ("himself), where, as we said, both elements refer to the same person
i.e. ["he 1 . . . him 1 "]. [[This whole line of a r g u m e n t , that e.g. heauton is com
p o u n d e d like allelous, so that he- refers to the subject and -auton to the object,
breaks down on plural forms like hemas autous, in which both parts are in the same,
non-nominative, case. This does occur to him eventually, in Ch. 160.]]
155. A possessive p r o n o u n based on the third person dual ("of the two of them")
does not exist either, not (as some have supposed) because there is no genitive
form of the third person dual, from which the possessive must be derived—and
they p r e t e n d that all the a p p a r e n t genitives are really datives being used [by
enallage] for genitives; in fact the form exists and has the correct meaning [sphoin,
genitive dual third person "of them two"]. But no possessive could be derived
from it, because, as we have shown, no derivatives are m a d e from enclitics.
156. But anyway we have said elsewhere that it is necessarily the case that there
are more reflexive p r o n o u n s of the first and second person than of the third. And
you must grant that the smaller must be tested by the greater, not only in syntax, 184
148 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
but also when the unanimous usage of all writers does not offer a single instance of
the analogical forms [*emauton for "ourselves" etc.]; and you must also concede
that this agreement is practically perfect. So it cannot be by chance that a larger
n u m b e r of forms [all the first and second person dual and plural c o m p o u n d
forms, *emauto, *emautoin, *emautous, etc.—ten distinct forms] have been lost [and
only a smaller n u m b e r , three—heautbn, heautois, heautous, d i s r e g a r d i n g the
feminines heautais, heautas, as A.D. generally does—have been preserved]. As we
have said, the opposite is what is likely to have h a p p e n e d [that the smaller n u m b e r
would be lost].
157. But all possible cases of missing forms cannot be b r o u g h t u n d e r the head of
analogical change. For instance, we do not say *gunes ("of a woman"), etc. [from
nominative gune], and yet the other oblique first declension forms are not elimi
nated [i.e. ordinary nouns and adjectives like agathe, gen. agathes are unaffected by
the analogy of gune, gen.gunaikos]. Likewise we don't say *megalos ("large" n o m . sg.
masc), and yet the other adjectives e n d i n g in -los are unaffected, etc. [[i.e. the fact
that megalou gen. sg. has nominative megas, instead of megalos, has no effect on
homalos "even, level" gen. homalou; there is no nom. sg. *homas, etc. a m o n g the
many adjectives e n d i n g in -los]]. In place of the missing forms *gunes etc., *megalos
etc. other equivalent forms are used, gunaikos ("of a woman") for *gunes, megas
("large") for megalos, hudor ("water" nom./acc. sg.) for *hudas. [Cf. above Ch. 19.]
For how could *emautōn(for "ourselves") a n d the other analogous formations (in
all cases and both dual and plural) have been expelled, as if by ukase, both from
everyday usage and from the practice of the poets? In fact the poets are m o r e
likely to h u n t for forms which deviate from ordinary usage. A m o n g p r o n o u n s , for
instance, we may mention seo [[this seems an odd example, since it is paralleled by
emeo and heo, and all are related by rule to the prose forms sou, emou, hou. Possibly
A.D. wrote sphe "them", which is rather isolated]] and min ("him/her"; no related
forms in other cases or persons), etc.
observation about the conjoining of different persons does not apply, is easily
susceptible to the pluralization change; for a third person plural consists entirely
of third persons.
160. It is clear from the preceding a r g u m e n t s that no one could be in doubt
about the reason why there are no plural c o m p o u n d s from plural p r o n o u n s [i.e.
no *hēmdutous from hēōmas autous or *humautous from humas autous]. T h e internal
inconsistency (machomenon) involved for c o m p o u n d s can be seen from the simple
plural p r o n o u n s . For we said [just above, in 159] that these first a n d second plural
p r o n o u n s refer to groupings of different persons. It is foolish to c o m p o u n d plural 186
pronouns, because the singular p r o n o u n s , when c o m p o u n d e d , remain in the
singular n u m b e r [and so refer to the same person; but the plural p r o n o u n s
already refer to different persons]. But spheas autous ("themselves") and the other
cases of this p r o n o u n , though they are not of different persons, are not com
p o u n d e d , since the result of such c o m p o u n d i n g would have been r e d u n d a n t
because of the prior establishment of the plural forms [heautous etc.] which
blocked [such forms as *spheautous], and with which they are interchangeable; so
Iliad 13.152:
161. (XXXI) Since we have shown in the preceding why the plural reflexive
p r o n o u n s are not c o m p o u n d e d , we must complete our treatment of p r o n o u n s
with a discussion of hēmedapos ("native", "fellow citizen of ours") a n d humedapos
("fellow citizen of yours") since some people have cast considerable doubt on the
hypothesis that they are c o m p o u n d s of hemon ("our") and edaphos ("land"). T o
support this we may derive evidence both from the form and the meaning, since
agreement in these is a necessary feature of every c o m p o u n d . For instance, in
misogunes ("woman-hater") it is clear that both misein ("hate") and gune ("woman")
are present both in form and in meaning, and in philoponos ("work-loving") both
philein ("love") and ponos ("work"). A n d even if two or m o r e semantic parts are
present in one simple word, it is not [necessarily] a compound—e.g. toxotes ("bow
man") or shuteus ("shoemaker", "leather-worker"); although the notion of tem-
nein ("cutting" of leather) is present [in skuteus], it's not a c o m p o u n d unless "cut" is
present in form, as well, as in skutotomos ("leather-cutter", "shoemaker"). Hence
Aristarchos and similar words which are p r o p e r names are not counted as com- 187
p o u n d s by some, because they don't use the meaning of the parts to make a
meaningful whole. But on this point we tried to show [in a lost work] that the
original composition prevails, as in Bacchylides 12.58, aristarchos Zeus ("Zeus ruler
150 T H E SYNTAX O F APOLLONIUS DYSCOLUS
of the best" or "best of rulers") [in determining that it is a c o m p o u n d ] , and not the
later transfer by convention to use as a p r o p e r name. [[In fact, the great majority
of Greek personal names are c o m p o u n d s of this sort, and very few of them are
attested as ordinary nouns or adjectives, but most Greeks were certainly aware of
the meanings of their names, as appears in the beginning of Plato's Cratylus.]]
Therefore, since [hemedapos and humedapos'] satisfy the criteria of agreement in
form and meaning [with the proposed parts], nothing prevents us from saying
that they are c o m p o u n d e d . [[But we don't say so today, because (a) the agreement
in form is incomplete, since edaphos hasph, notp, and the third word of this form,
allodapos "foreign", lacks the e as well, (b) p r o n o u n stems e n d i n g in d- are known
from Sanskrit, and -apos has satisfactory cognates as a suffix in other languages.]]
162. "But if they really are compounds," they say, "they [hemedapos, humedapos]
o u g h t to be accented recessively, like áxiphos ("swordless"), dusrigos ("cold-
sensitive"), euteichos ("well-walled")." But that is not a crushing a r g u m e n t either,
since other c o m p o u n d s also show variation in accent. [Of s- stem adjectives, based
on neuters like telos "end, pay," mekos "length, height", etc.], some are accented on
the ultima, e.g. euteles ("cheap"), eueides ("well-shaped"), but others are not, e.g.
eumekes ("tall"), megakētēs("full of sea-monsters"), because of the e in the penult. So
why cannot we say "compounds derived from edaphos ("land") are accented on the
ultima, like those formed with ergon ("work") such as kakoergos ("evil-doer"),
hupourgos (from hupo-ergos "under-worker", "assistant")—unless they are to be
derived [as de-compounds or para-synthetic derivatives from compounds] from
hupourgō ("I serve, assist") and the like. "But, in that case, "they [hemedapos,
humedapos] ought to be two-termination adjectives [using the form in -os for both
masculine and feminine]." T o this we can reply that once you grant accent on the
ultima, you also grant distinct endings for feminine forms [as hemedape, humedape],
since all c o m p o u n d adjectives of two terminations also have recessive accent, e.g.
kakopastos ("badly salted"), dusrigos ("cold-sensitive"), kakopathos ("miserable").
163. T h e refutation [of the view that hemedapos and humedapos are compounds]
becomes more complete if we consider the fact that p r o n o u n s of the first and
second person never enter into composition with another part of speech, a n d that 188
plural pronouns are never c o m p o u n d e d at all. Pronouns make derivatives [with
suffixes] rather than c o m p o u n d with another part of speech, e.g. from emou ("of
me") comes emos ("my"), from noin ("of us two") comes nōiteros ("our" of two
people), from hemon("oi us") comes hemeteros ("our"). A n d if we suppose that
hemedapos and humedapos are compounds, they cannot be counted either as nouns
or pronouns. Since in compounds it is a rule that the end of the word determines
the part of speech—e.g. misanthropos ("man-hater") is a n o u n , t h o u g h it contains
the verb misein ("hate"), and eusebo' ("I revere") is a verb even though it contains the
adverb eu ("well"), and so also is cheirographo ("I hand-write") even though it
contains cheir ("hand"). For this reason it can be observed that every part of speech
remains the same, whether simple or c o m p o u n d [in final position]—every one,
that is, which can be compounded, except the preposition, although it has almost
unlimited potential for compounding, because it always comes first in a com
pound, and the first m e m b e r loses its identity—if this is true, how can hemedapos
help being a n o u n , if the n o u n edaphos is its final member? But how can it be a n o u n
if it is used as substitute for a noun? [In fact, it has the same morphology and
B O O K II: P R O N O U N S 151
167. T o the first question [why -dapos is only a d d e d to plural pronouns] we can
reply that it would be rather i m p r o p e r to make an ethnic derivative from singular
or dual forms,.since one cannot conceive of a nation in one person or two, but only
in uncounted n u m b e r s . W h e n one says houtos hemedapos esti ("This m a n is from our
hometown/ is our fellow-citizen.") the speaker is including not only himself, but
also his other fellow-citizens. Words like emos ("my") a n d hemeteros ("our") which
can be applied to anything whatever that can come u n d e r the possessive relation,
are naturally able to be derived from every n u m b e r indifferently. If you want to
say "our fellow-citizen" using hemeteros ("our"), you must apply polites ("citizen"), 191
but not so with hemedapos. In it, the derivative suffix is sufficient.
168. As for the question why this suffix cannot be a d d e d to the third person stem
[of the plural, i.e. sphe-], you might at first think it is because the p r o n o u n is not in
ordinary use. But that's not likely. For there is a derivative spheterizesthai ("to
appropriate") made from this unordinary p r o n o u n sphon ("them(selves)")—and
similarly there is a derivative mēlōtē("sheepskin") in use from the word melon
("sheep") which is not the ordinary word for probaton ("sheep"), and similarly from
the unordinary verb derkesthai ("to see") come fairly ordinary words like dorkas
("deer, gazelle" i.e. "bright-eyes" or "big-eyes") and the c o m p o u n d oxudorkein("to
be sharp-eyed").
169. It rather looks as if there is a m o r e natural explanation for the gap. J u s t as
there was a plausible explanation for the extension of c o m p o u n d i n g [for re
flexives] in the third person plural, which was not inherent in the first and second
persons [plural], so the fact that causes are inherent in first a n d second person
p r o n o u n s for the derivation [in -dapos] does not imply at all that the third person
could make this form. A n d obviously the inherent cause is this: the plural first and
second person pronouns, as we have seen (ch. 159), include within themselves a
n u m b e r of different third persons. Anyone who says hemas ("us") includes himself
and different persons who are with him, and if anyone says humas ("you" pl.) to
someone, he is including both the person present [the addressee] a n d others who
are absent. From these p r o n o u n s an ethnic derivative, given the domination of the
B O O K II: P R O N O U N S 153
1. (I) In the preceding book we have discussed the features peculiar to pro- 194
nouns; now we must try to cover the features which they share with the other parts
of speech, some of which cause related ungrammaticalities, while others are
indifferent in this regard, as it is possible to see in the first instance from the
p r o n o u n s themselves. [[This rather vague allusion seems to refer to what im
mediately follows, which deals with agreement and non-agreement of p r o n o u n s
in person.]]
2. (II) T h e p r o n o u n autos is called appositive (epitagmatike) because, as is said, it is
placed in apposition with p r o n o u n s of any person [as well as NP's in general]. This
does not h a p p e n exclusively with autos, of course, as is clear from examples like:
3.1 hod' ego cheiron ("I am he, worse. . . " possibly misquoted for
Odyssey 16.205 hod' ego toiosde "I am he, such as you see.")
or Odyssey 24.321:
3.2 keinos men de hod' autos ego, pater
("This T here, myself am he, father.")
or Pindar, Olymp. 4.26:
3.3 houtos ego tachutati
("I am he, in speed. . . ")
But if anyone wants to say that it h a p p e n s m o r e often with this p r o n o u n than with 195
others, then I will agree to the argument. [[That autos is more frequently used
appositively than any other p r o n o u n is correct; in this use it regularly means
" - s e l f , - s e l v e s . " ] ] No one would dare to call such constructions [as the three
quoted above] ungrammatical (alogos), since they are used by the most famous
writers and d o not violate grammatical rules, as I will show [in Ch. 23]. It is clear,
then, that the name "appositive" p r o n o u n [we would say "intensive"] is taken from
t h e fact t h a t this is a very f r e q u e n t c o n s t r u c t i o n [which h e h a s n ' t yet
exemplified—ego autos "I myself etc.]. This has h a p p e n e d often in other cases, as
well, where a predominant usage prevailed [in giving a name].
3. I'm a little puzzled by the fact that while first and second person p r o n o u n s can
never shift into the third person [i.e. be used as or in apposition with third person
pronouns], some third person p r o n o u n s do shift into the first and second, though
not all third person p r o n o u n s do this. T h e p r o n o u n hi, heo ("him, her, it") a n d its
other forms (suzugos) [e.g. hou, hoi, heoi, he, min, hethen, hee, heu, nin], which are
simple [nonreflexive] third person forms, never shift over to the first or second
person, but the p r o n o u n autos which replaces [hou etc. in ordinary Greek: auton
"him" etc.], does shift, as in Odyssey 4.649:
or 8.443:
3.5 autos nun ide pōma
("Now check the lid yourself.")
And, more surprising, everyone agrees that heauton a n d heauton ("himself') and its
other forms [heautoi; heautes, heauten, heautei, " h e r s e l f ; heauto "itself] are com
p o u n d e d of two third person pronouns, but that the plural forms heauton, heautous
("themselves") are used with the first and second person [i.e. to mean "ourselves"
a n d "yourselves"]. Isn't it illogical that a construction which is ungrammatical in
the singular should become acceptable when changed to the plural? For we do not
say *heauton hubrisa ("*I h u r t h i m s e l f ) or *heauton hubrisas ("*you h u r t h i m s e l f ) ,
but we do say heautous hubrisamen ("we h u r t ourselves"—lit. "themselves").
4. We might answer this briefly as follows. "Perhaps such constructions do
contain a solecism caused by the substitution of persons. If something is third 196
person, then using it as first or second makes a sentence ungrammatical. And,
conversely, if something is first or second person, using it as third person will cause
the same ungrammaticality, if one says, [as a conjoining of] grapho ego ("I write")
[and] grapheis su ("you write"), *graphousin hemeis ("we, they-write"). A n d if this is
correct, how can heautous, a third person form c o m p o u n d e d of third person
forms, make a grammatical sentence when used in the first and second persons?
T h e p r o p e r thing to say is heautous tuptousin ("they beat themselves") not heautous
tuptomen ("we beat (them)selves")."
6. (III) So we'd better stop a n d explain what the actual cause of the u n g r a m m a t i
cality is, not by the mere citation of examples as some [linguists] do, pointing out
the ungrammaticality without explaining the cause. But if you don't grasp the 197
cause, it's an exercise in futility to cite examples.
7. Besides, their arguments don't stand u p . Suppose we say that there is un
grammaticality of n u m b e r , as some have claimed, in Odyssey 8.48:
3.6 hourō de krinthente duo' kai pente'konta
("Young men, chosen, two and fifty" where "young m e n " and "cho
sen" are dual in form).
156 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
T h e y say that it's impossible to use the dual in place of the plural, but only the
plural in place of the dual [[i.e. that the dual means "exactly two" whereas the
plural means "two or more"]]. However, it is clear [that this text is all right] because
there is a reason, and we do not concede—as we will show later [not in our
text]—that krinthente ("chosen" dual) actually is to be construed with pentekonta
("fifty"). [[Presumably A.D.'s explanation involves a "gapping" rule by which
kouroi krinthentes ("young men, chosen" (plural)) has been deleted after kai. This is
a variant of what was called zeugma or prodiezeugmenon schema, similar to the
Alcmanic schema.]] O r take one [which they say is ungrammatical] in respect to
person, e.g. Odyssey 13.320 [with some mixture of 11.167]:
3.7 all' aiei phresin heisin echon alalemai o'izun
("but always I wander about with sorrow in my ("his") heart")
where they say [heisin "his"] is ungrammatical, since it should be emeis ("my"). In
reply to this, there is an obvious line of a r g u m e n t . If you are willing to accept
heautous hubrizomen (for "we h u r t ourselves") a n d other such examples, clearly you
can accept this phrase. O r is there some principle which will reject this case, but
allow heautous hubrizomen? A n d they also [cite examples which are deviant] in
respect to case, for instance when they require the [partitive] genitive [instead of
the nominative] in the sentence beginning in Odyssey 12.73:
3.8 hoi de duo skopeloi
("The two rocks. . . "[for the whole sentence see above Bk. 1.156
where the passage is discussed at length]).
I n this case, too, it is possible to consider a n d ask whether something has been
deleted or what the rule is which requires (apaitei) the genitive here. But let's stop
this accumulation of examples, or else we'll get too far away from o u r topic. [[One 198
man's figure (schema) may be another man's e r r o r (soloikismos).]]
8. (IV) I'm not forgetting the attempt by some to discredit the generally accepted
opinion that an e r r o r in a single word [i.e. wrong word, wrong pronunciation,
wrong inflection—errors of phonology a n d morphology] is a barbarism, whereas
errors in the grammatical combination of the words in a phrase [i.e. errors oi
agreement or government: wrong case, n u m b e r , gender, mood, tense, person]
constitute solecism—to discredit this view by alleging examples where a single
word in isolation involves a solecism, as for instance if someone.were to say houtos!
("this one" masculine, singular) with reference to a female or a g r o u p , and other
equally silly examples. But, in the first place, no nominative constitutes a complete
utterance (autoteleia) without at least a verb—and more, if the verb is not intransi
tive. T h e sentence houtos peripatei ("this m a n is walking") is complete, but nothoutos
blaptei ("this m a n is injuring"), because you must still ask "who [is he injuring]?".
But suppose we take houtos as answering a question tis se etupse ("Who beat you?") in
that case we must understand the verb [and object] as shared [from the question]
[i.e. houtos me etupse "this m a n beat me."] So also tis kaleitai Aias ("Who is n a m e d
Ajax?"), answered by houtos [i.e. houtos kaleitai Aias "this m a n is n a m e d Ajax"]. So it
is not true that a solecism can be contained in a single word.
alters the deictic interpretation [i.e. the reference of the word; hence it's like using
the wrong word, e.g. calling "boy" to a girl]. A n d obviously, this novel inventive
solecism involving the p r o n o u n houtos cannot take place in the dark, so a special
clause must be added to the definition [of houtos as solecism] "except in case the 199
utterance occurs in a dark-night situation; for clearly light is needed to determine
sex." But this is silly. Solecisms fall u p o n the ear, are specified by the presence of
irregularity in the way words are put together, a n d can be recognized by people
who have lost their sight. For they are not lacking in the sense which is specialized
for checking the form of words, namely hearing. This, too, will be granted, that if
something has changed merely in respect to reference [emending protaxeis "front-
ings" to apotaseis "references" as in 1.58], it will not fall u n d e r the theory of
solecism. [[Other editors e m e n d to prostaxeis "commands" or praxeis "actions".]]
This would be just as silly as the original example.
10. So when someone says houtos me etupsen ("this m a n beat me") of a woman, it is
not a grammatical error. T h e sentence obeys all the rules of agreement a n d
government (katallelon). But if, referring to a female, someone were to say *haute
me etupsan ("This woman, they beat me"), even if the g e n d e r reference is correct,
the sentence is nevertheless solecistic because of the e r r o r in agreement [singular
subject, plural verb]. Grammaticality or ungrammaticality does not d e p e n d on the
nature of the entities referred to, but on the construction of the words, which allow
the possibility of correction, while the external facts remain unchanged. [[He
should have said, "whether or not the external facts are changed", because
falsehood is not ungrammatical—and this is his real point.]] It's also possible to say
it in the neuter: touto me etupse ("This one [neuter] hit me"), without making the
sentence in any way deviant; for instance, it wouldn't be at all deficient if the
p r o n o u n referred to to gunaion ("the little woman", neuter) [in the mind of the
speaker], as for instance in Odyssey 12.74-5:
or even more so hendekate ("eleventh" f. gen. sg.). [[Because hen is the neuter n o m .
sg. form of "one", and dekatos, dekate, is "tenth", m. a n d f.]] In this case one can't
claim that the c o m p o u n d contains the [masc./neuter] genitive henos ("(of) one") [in
m o d e r n terms he means t h e masc./neut. stem, which here would be heno- before a
consonant]. C o m p o u n d numbers of this type are m a d e from nominatives, as in
treiskaidekatos ("thirteenth" composed, apparently, of treis "three" nom. pl. m./f.,
kai "and", and dekatos "tenth"), tessareskaidekatos ("fourteenth"). But hendekate
("eleventh" f.) is not made that way. [[The feminine of "one" is mid in the n o m . a n d
in the stem form; the masculine nominative is heis. ]] T h e simplest way to cope with
this form is to say that hendekatos ("eleventh") is a regular derivative from hendeka
("eleven"), w h e t h e r hendeka is to be r e g a r d e d as a c o m p o u n d or a p h r a s e
(parathesis) [hen + deka, "one + ten"], like hebdomatos ("seventh", Homeric form)
from hebdomos ("seventh"), from which the feminines are hebdomate, hendekate. [[It's
a little puzzling why A.D. drags in this rare Homeric form as a parallel; the ordinal
suffix -tos ("-th") appears plainly in tritos "third", tetartos "fourth", pemptos "fifth",
hektos "sixth", and enatos "ninth", as well as dekatos "tenth". It is true that only enatos
and dekatos have -atos, and the stem of enatos differs from ennea "nine."]] A n d we
have elsewhere [now lost] shown at length that a unified derivative word (parago-
gon) may be made from a two-word phrase [assuming that hendeka is such a
phrase], e.g. from tauto (to auto "the same") comes tautotes ("identity"), a n d from
parapotamon menei ho deina ("so-and-so lives by a river") comes parapotamios ("living
by a river").
e.g. the above-mentioned nouns and whatever else can make a distinction of
gender. [[Since this class is virtually identical with class one—case a n d n u m b e r —
excluding only the personal pronouns, and since Priscian 17.153 (p. 182.22 H.)
combines it with class one, substituting here the criterion of tense-aspect, it may be
that something is wrong with the text here. O r it may be that A.D. really means
inflected for gender, as metaschematizetai would mean if taken strictly, in which case
this class includes adjectives only, rather than all nouns, but includes also pronom
inal adjectives and demonstratives—and autos—, plus participles and articles.]]
Finally, (d) some are not inflected for any of these categories, namely those which
have only a single form, such as conjunctions, prepositions, and almost all the
adverbs. [[Varro's four-class scheme, e.g. 8.31.44, differs in using tense-aspect,
instead of person; his "b" class is verbs only, and "c" is participles.]]
14. T h e first [three] classes of words, changed according to their respective
inflectional patterns into the required concordance (akolouthia) of the aforesaid
categories of [case and] number, person or gender, in the process of sentence 202
construction are marked for association with whatever word they are construed
with, say plural to go with plural for association with the same referent [subject
NP],graphomen hemeis ['we are writing", where -men "we" agrees with hemeis "we" as
referring to the same persons], graphousin hoi anthropoi ["the men are writing"
where -usin "they" agrees with "the men"]. In case there is a non-reflexive com
plement (diabasis) [direct object, indirect object, genitive complement, etc.], the
verb and the complement do not have to agree in n u m b e r ; we can equally well say
tuptousi ton anthropon ("they are beating the man") a n d tuptousi tous anthropous
("they are beating the men").
agreement [[A.D. names no others here, but adds tense for things like "tomorrow
it will rain" vs. "yesterday it rained" in Ch. 19; one other possibility might be
so-called attraction of mood, where a verb is put into the optative if it depends on
an optative.]] For [if the word has no inflections] there's no test by which a failure
to agree (metaschematismos) could be detected.
18. We can construe the wovdkalos ("well") and others like it with any person and
any n u m b e r , e.g. holos grapho ("I write well") or kalos graphete ("You (pl.) write
well") and with inflection for tense, e.g. kalos egrapsa ("I wrote well") or kalos grapso
("I will write well"). All such constructions are obviously grammatical. For the
co-occurrence of an adverb, which is not inflected for the n u m b e r or person or
mood or tense of the verb, is quite free of hindrance in a construction, since there
is no test for agreement. T h e same cannot be said of the [adjective] kalos ("fine"
n o m . sg. masc.). It happens to be third person a n d singular in n u m b e r , a n d hence
can appropriately occur [as subject] only with a third person singular form of
graphein ("write"): kalos graphei ("A fine m a n is writing"), kalos peripatei ("A fine
m a n is walking"). But since, furthermore, kalos does not signify any particular time
relationship, it can be indifferently combined with verbs of different tenses.
19. However, the set of adverbs which are subdivided into different time-classes,
while they freely combine with different persons and numbers, do not combine
with all tenses; so [past tense adverbs do not freely combine] with future or present
tense forms, [e.g. echthes egraphon ("yesterday I was writing"), echthes egrapsa ("I
wrote yesterday"). But aurion "tomorrow" is incompatible with pastness, a n d will
not go with any forms except those which are inherently non-past, e.g. aurion
grapho ("tomorrow I write"), aurion grapso ("tomorrow I will write"), etc.]. [[The
insertion here is taken from "On Adverbs", and must correspond fairly well to the
lines lost from our text of the Syntax.]] However, those adverbs which are as
sociated with the whole extent of time [are not thus restricted], e.g. nun "now" and 204
similar adverbs [no doubt including tote "then", ede "already", eti "yet", etc.].
Similarly, those adverbs which have a semantic feature (semainomenon) of mood
(enclisis), such as eithe or age; for the imperative feature (enclisis) is incompatible
with the optative [feature], and so eithe [roughly "I hope, I wish" when combined
with an optative—for future wishes—or a past indicative—for unreal wishes]
rejects imperatives and age ["Come on (and). . ." with imperatives a n d hortatory
subjunctives] rejects optatives. T h e other adverbs, which contain no semantic
features of this kind [mood] have u n h a m p e r e d co-occurrence with all moods
{enclisis). We have discussed these restrictions at greater length in "On Adverbs"
[123,1—125,4], a n d will discuss them again as needed [perhaps in the lost section
of Bk. 4].
20. T h e same fact can be established in the case of conjunctions [as in the case of
adverbs], that any which are not subclassified with respect to any of the above [i.e.
persons, numbers, tenses, moods, cases] can combine elements indifferently as far
as concerns differences of gender, case, or distinctions of person. But if there is
such a subclassificatory distinction (merike diaphora) in these respects, then [the
conjunction in question] is no longer covered by this general rule (suntaxis),
because of the inherent meaning of the conjunction. T o avoid dragging the
subject of conjunctions in at this point, we will illustrate the rule with just one
example.
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 161
23. Obviously anything [i.e. any p r o n o u n or adjective] which takes no part in the
category of person cannot be wrongly used in respect to concord (akolouthon) for
person, so that if it agrees properly in gender, case and n u m b e r , it cannot be
considered to be incorrectly used in respect to the aforesaid constructions. [[This
sentence is very awkwardly worded, but it must mean essentially this.]] So erne
auton ("me myself) and hemas autous ("us ourselves") are all right. This point is
clear also in the case of the pronouns which, since they do not distinguish gender,
have no restriction against occurring in construction with [words of] all these
genders: so we say su autos or su aute ("you, y o u r s e l f masc. or fern.), and likewise
emoi autoi and emoi autei ("for me m y s e l f masc. a n d fern.), and there's no u n g r a m -
maticality, since there is no alternative form which could be substituted to correct
the e r r o r [i.e. no feminine *sue "you" to go with aute]. T h e same reasoning applies
tohoutos ("this, that") andhode ("this"); here, too, there's nothing to prevent saying
hod' ego ("I, here") or houtos ego ("this one is me" or "I am the one"), as we said above
[Ch. 2]. This also solves the problem of the extension oi heautous ("themselves") to
first person use (as "ourselves"), which would indeed be ungrammatical if there
existed a form *emautous ("*myselves") which could thus dislodge heautous as an
e r r o r in person. Since there is no such "correct" form, this fact makes the usage [of 207
heautous for "ourselves"] free from e r r o r as regards person. [[This ignores hemas
autous "ourselves".]]
24. T h e same logic applies to verbs. All the finite forms (enclisis), being character
ized for person and n u m b e r , admit grammatical e r r o r in regard to agreement for
person and n u m b e r . But the infinitives, since they have no such inflection [for
person and number], combine freely with all persons and all n u m b e r s : graphein
erne ("me to write") graphein hemas ("us to write"), graphein se ("you (sg.) to write"),
graphein humas ("you (pl.) to write"). But since they are not deficient in voice or
tense-aspect, they may exhibit grammatical error when those are incorrectl)
substituted.
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 163
25. Quite naturally, since this form [enclisis, the infinitive] has no illocutionary
force [mood, lit. "mental attitude"], it cannot be blocked from occurring in place of
all moods (enclisis), with the addition of words signifying the characteristic modal
feature (enclisis), and conversely, every mood (enclisis) can be converted into [a
construction with the] infinitive. For graphe("write!") can be equated to graphein soi
prostasso ("I bid you write") where we necessarily add the "bid" which is implicit in
the imperative, and the p r o n o u n ("you"), [also implicit]. For the infinitive has no
share in either of these [modality or person]. Peripatoies ("may you walk!") [is
equivalent] to euchomai se peripatein ("I hope/pray for you to walk"), and grapheis
("you are writing") to horizomai se graphein("I declare you to be writing"). T h e
conversion (metalepsis) from [finite forms to infinitives] is obvious also [in indirect
discourse]:graphoi Dionusios ("May Dionysius write") goes to euxato graphein Dionu-
sion ("He hoped/prayed for Dionysius to write"), a n d grapheto Dionusios ("Let
Dionysius write!" 3rd person imperative) to prosetaxen graphein Dionusion ("He
gave orders for Dionysius to write"). A n d the reason why the accompanying
nominatives [with finite forms] become oblique, we shall specify in the section on
verbs (Ch. 78-87 below), where we will also discuss in general the structures in 208
which infinitives take part. [[This seems to present a m o r e explicit performative
account of illocutionary force than 1.51, but similar to that in 3.61, below.]]
26. This can be confirmed also from the behavior of participles, which acquire
gender, case and n u m b e r when they are transformed (metalepsis) from verbs, but
lose the distinction of person and mood. A n d so we cannot be ungrammatical in
respect to the lost categories, e.g. in person grapsas anesten ("I wrote and got up"),
grapsas anestes ("You wrote and got up"), grapsas aneste ("He wrote and got up"),
a n d in respect to mood, grapsas anastaien ("May I write and get up!"), grapsas
anastethi ("Write and get up!"). T h e categories that remain, voice a n d tense-aspect,
fall u n d e r the theory of ungrammaticality, if anyone combines these [participles]
contrary to rule. [[The a p p a r e n t parallel here between infinitives and participles
does not lead A.D. to alter his classification system, in which the infinitive is one
form (enclisis) of a verb, whereas the participle is a distinct part of speech, each
instance of which is always derived by transformation (metalepsis) from a particular
tense-aspect and voice of a verb.]]
27. (VII) It is generally recognized that when forms [which are elsewhere distin
guished] coincide (sunempiptei), making homonymous either different genders, or
cases or persons or any other such categories capable of syncretism in form, the
possibility of ungrammaticality [in respect to the homonymous category] is re
moved. Sometimes [these forms] are used in an i m p r o p e r construction, but are
subject to reinterpretation as a different construction capable of occurring in the
sentence because of the h o m o p h o n y (sunkechumenon tes phones). Suppose we take
sophos ("wise") or any other three-termination adjective (trigeneia), and suppose we
combine it with theos ("divinity" "god or goddess") or any other n o u n capable of
interpretation in either gender. In this construction with theos, even if it is used in
the sense of "goddess", the phrase will be unambiguous in g e n d e r [i.e. theos sophos 209
means "wise god" and theos sophe means "wise goddess"]; but if the adjective is klutos
("famous") or agrios ("wild") [ which are used both as two-termination a n d as
three-termination adjectives by H o m e r ] , the same thing will not be the case. T h u s
we recognize that in Iliad 2.742 klutos Hippodameia ("famous Hippodamia") a n d
164 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
agrion aten ("wild blindness", acc. sg.) there has been hypallage of gender. [[I.e.
A.D. considers that klutos a n d agrios are really three-termination adjectives, so the
correct underlying forms would be klute Hippodameia a n d agrien aten, and H o m e r
has used the poetic license of substituting masculine forms for feminine.]]
28. Therefore, the c o m m o n Athenian oath ma to theō ["by the two goddesses" i.e.
Demeter and Persephone; it could also mean "by the two gods" since neither to nor
thed is marked as feminine] exhibits the figure [hypallage, wrong gender] not in the
word thed but in the accompanying article, which shows that the word theos is now
masculine, since the article is not ambiguous in gender. [[In fact, the feminine dual
to is quite normal in Attic, and most m o d e r n g r a m m a r s give that as the correct
form. But see Cooper, 1972.]] Obviously, if the article is removed, the figure
[hypallage of gender] is eliminated. This is not the same as in Hesiod's Works and
Days 198:
3.11 kalupsameno chroa kalon
("[they two, feminine] covering their fair skin")
or in Iliad 8.455:
3.12 ouk an eph' humeteron ocheōn plegente keraunoi
("[You two, feminine], struck by a thunderbolt, would not [come to
heaven] on your chariots." plegente "struck" is masc. dual, nor
mally.)
since h e r e the figure is in the head words [i.e. the two participles, kalupsameno a n d
plegente], but in the above case [of ma to theō] it d e p e n d s on the adjunction
iparathesis) of the article.
29. T h e same reasoning holds for all the tenses, at least if you have a reasonably
correct understanding of their distinctions. It is agreed that grapho ("I'm writing")
and egraphon ("I was writing") differ in tense, and that it is not possible to say echthes
grapho ("Yesterday I'm writing") [[actually this is a possible Greek sentence, since
Greek—much more than English—uses the present as a narrative tense. Most
likely, however, Apollonius would regard grapho, so used, as figurative language,
derived transformationally from an underlying egrapsa or egraphon]] but r a t h e r in
the imperfect tense echthes egraphon ("yesterday I was writing"). It is also agreed
that the participle corresponding to grapho is graphdn ("writing", "who is/was
writing") and so one shouldn't be able to say echthes graphdn ("writing yesterday"),
because you can't say echthes grapho. But in fact the participle allows the adverb 210
["yesterday"], inasmuch as it is, by coincidence, also the participle of the imperfect
tense. T h u s we say egraphon kai eniomen ("I was writing a n d having trouble"), which
construction, by transformation, becomes graphdn eniomen ("I was having trouble
writing"). [[Presumably this constitutes a general rule for the so-called cir
cumstantial participle of Greek; its underlying form is a coordination with kai
("and")joining two verbs in the same person a n d n u m b e r , a n d the transformation
converts the first verb into the corresponding participle a n d deletes kai. It's a
reasonable proposal.]] T h e same thing is true of infinitives: the infinitive graphein
("to be writing"), since it has the same form, whether it is transformed from a
present indicative or an imperfect indicative, can be grammatically combined in
the adverbial constructions here considered [i.e. with adverbs of present or past
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 165
time]: it is possible to say either sunebe echthes graphein Apollonion ("It h a p p e n e d that
Apollonius was writing yesterday") or sunebe semeron graphein ("It h a p p e n e d that
he was writing today"). [[Elsewhere A.D. realizes that "today" is not a good test,
since part of today is always past, but h e r e he forgets that.]] You won't find this sort
of behavior from infinitives that correspond to only a single tense, e.g. [in the
future] *grapsein echthes ("to be going to write yesterday"). [[In fact, however, this
can occur in Greek, as in English, in indirect discourse situations in a past con
text.]] But where there is no neutralization (coincidence of form, sunemptōsis), I
mean in the case of grapsai ("to write", aorist infinitive), [one can say] sunebe echthes
grapsai Apollonion("It h a p p e n e d that Apollonius wrote (or had written) yester
day"). [[There's something wrong with this last sentence; the future involves no
neutralization in A.D.'s sense so why "but"? Uhlig would e m e n d to "where there is
neutralization,. . . " but his a r g u m e n t that A.D. considers the aorist ambiguous
between a recent past a n d a remote past sense is irrelevant here; there's just one
indicative, egrapsa,to be transformed into grapsai in this construction. T h e r e are
other uses of the aorist infinitive, of course, which cannot be derived in this way,
and which can co-occur with "tomorrow". Most likely something has been lost
here. O r else one should simply delete "where there is no neutralization".]]
mean epoiethen ("I was made") or vice versa. According to this account, in examples
like Iliad 13.60:
3.13 amphotero kekopos
("having struck both men")
or 2.264:
3.14 peplegos agorethen
("having struck [him, I will drive him away] from the assembly")
or Odyssey 10.238:
3.15 rabdoi peopleguia
("having struck [them] with her wand")
or Iliad 1.56:
3.16 hoti ra thneiskontas horato
("that she saw them dying")
etc. [[In all these examples so-called "middle" forms, which frequently have a
passive sense, have a transitive, active use]], there is no hypallage of voice, given
this explanation of the "middle" but these forms have attained the sense of both
voices by the necessary logic of the structure. [[This is hardly a satisfying account
of the middle; but it does specify it in a distinct way.]]
3 1 . It is possible to show the same thing with regard to h o m o p h o n y (coincidence
of form) of person [in verbs]. T h e form niko("I a m winning") is of the first person
and cannot function as the second person, since the corresponding second person
form is nikdis. But since there is an imperative [passive] form of the second person
with the same phonological shape, it will be grammatical when used in the second
person, as an imperative (as we have remarked). This line of a r g u m e n t extends 212
even to the third person, which might be excluded from the h o m o p h o n y because
of the addition of the i in writing [i.e. nikdi]: however, as far as the actual sound is
concerned, it is perfectly h o m o p h o n o u s [since the -i is silent] with the first a n d
second person forms [which we have mentioned], a n d it will be grammatically
used if the appropriate mood (enclisis), namely optative, is in order, as is the case in
this passage of Alcman (fr. 91 Edmonds):
is I") a n d for ekeinos estin ("He is. . . " or "it is he"), then it should be the same case
for su ei ("You are. . . " or "it is you").
37. (e) If ei ("you are") is never construed with a [predicate] vocative, but only a
nominative—Aristarchos ei("you are Aristarchus"), grammatikos ei ("you are a
grammarian"), but not *Aristarche ei ("Aristarchus! you are"), how can it be denied
that su ei ("It is you") contains a nominative? (Incidentally, it is by no means the case
that nominative nouns are always taken as third person, as we have shown [2.47]
that verbs of being take nominative predicates in both first a n d second person.)
38. (f) Since the first person plural admits the inclusion of second and third
persons [along with "I", the speaker] in the same case, when we say analytically
emoi kai soi kai Dionusioi elalesen Truphon ("Tryphon spoke to me, you and
Dionysius"), and collectively (kata sullepsin) hēmin elalesen Truphon ("Tryphon
spoke to us"), a n d so for the other [oblique] cases, how can you deny that in ego kai
su kai Truphon [paraginometha] ("I and you a n d T r y p h o n [arrive]") all three words
are in the same case, since they, too, may be collected into a nominative plural:
hemeis paraginometha ("we arrive")?
39. (g) We can also offer evidence from the construction with conjunctions. We
conjoin nominals (ptotikon), when they share the same verb, invariably in the same
case; e emoi charizetai Truphon e soi e Dionusioi ("Tryphon favors either me or you or
Dionysius"), e emou akouei Truphon e sou e Theonos ("Tryphon is listening either to
me or to you or to Theon"), and it is obvious that if one were to insert a different
case [in such a conjoined series], the result would either be a grammatical e r r o r 217
(solecism) or require an inserted verb. W h e n the verb is shared, we have e emou
akouei Truphon e sou e Dionos ("Tryphon is listening either to me or to you or to
Dion"), and one can also say [with change of verb] eemou akouei Truphon e soi homilei
("Either T r y p h o n is listening to me [gen.] or talking to you [dat.]"). Admittedly,
then, when there is a single verb, and a grammatical sentence, [conjoined] cases
must be the same. So [one can say] e ego aperchomai e su eDionusios ("either I leave, or
you or Dionysius"), and if the sentence is ungrammatical, then you could have su in
a different case, but if it is grammatical, then you admit that [su] is nominative
[since ego and Dionusios are agreed to be nominative].
40. (h) But I don't even concede that su grapheis ("You are writing") normally has
a vocative structure. Such examples show not that su [is normally vocative and] also
may be nominative, but that [it is normally nominative] a n d also may be vocative.
W h e n it is united in a single [phrase], it is nominative, parallel to ego grapho ("I am
writing") and ekeinos graphei ("He is writing"), but if you hesitate after the su and
use it to call the attention of your addressee, then it will be admitted [that su is a
vocative] used unnecessarily and contrary to the n o r m for p r o n o u n s . For we have
shown that p r o n o u n s were invented in the first place because nouns cannot be
used [in the first and second persons]. But the vocative form moves a n o u n into the
second person, and so clearly the vocative p r o n o u n is superfluous. [[I.e., since you
can always say "George! come here!" or "Boy! come here!" you don't really need to
say "Hey you! come here!".]]
4 1 . And there isn't even the excuse available that there was for third person
p r o n o u n s (which were r e d u n d a n t since nouns could always be used instead). T h e
170 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
a r g u m e n t for t h e m was that they are b r o u g h t in for the sake of deixis (pointing) or
a n a p h o r a (reference), since bare nouns couldn't convey those values. But when su
("you") is used as a vocative, it can't convey any reference [back to an earlier 218
mention], for it's in the second person [and the second person is present and needs
no reference to identify himself], n o r can it give deixis, since addressees are
addressed, not pointed out, a n d the vocative can be applied to persons who are not
visible [and hence cannot be pointed out]. And, besides, addressees are likely to be
annoyed at being called by means of the p r o n o u n ["Hey, you!"], since naturally
they want to hear their own name, when called, although the p r o n o u n is expected
to be used in the other cases [i.e., "of you", "to you", etc., rather than "of George",
"to George", etc., when George is the addressee], since then the application of the
name is not proper, as we have shown [2.43]. How then can the purely optional
vocative use of the second-person p r o n o u n s overshadow the obligatory nomina
tive use and vindicate the claim [that su] is vocative only? I think it would be
superfluous to bring on more examples, since the case has been proved beyond a
shadow of doubt.
42. (IX) It doesn't seem inappropriate at this point to explain the vocatives of the
remaining p r o n o u n s [i.e. in our terminology, (a) personal p r o n o u n s , (b) demon
stratives, (c) autos, which is often now treated as two distinct items, an intensive
p r o n o u n "same" or "-self and, in the oblique cases, a third-person p r o n o u n ] . For
the same reasons as su, the second person dual ["you two!"] and plural ["you-all!"]
p r o n o u n s are used like it in the same way in the vocative, a n d these vocatives again
are h o m o p h o n o u s with nominatives [as is always the case with duals a n d plurals];
but first and third person p r o n o u n s are not so used, since (a) no one ever calls to
attract his own attention, and (b) no one speaks to persons who are absent and out
of earshot. T h e vocative is a case for a person who is in the speaker's presence, or at
least no farther away than his voice will reach. So naturally neither autos nor ekeinos 219
has a vocative, since autos refers to absent persons [i.e. as a third-person u n e m p h a -
tic anaphoric p r o n o u n "him", "her", "it", "them", it normally refers to entities
which have been mentioned but are not in sight] and ekeinos ("he", "that one") to
ones at a distance, which is inconsistent with the characteristic nature (idioma) of the
vocative. But houtos ("this one", "that one", referring to someone in sight, particu
larly if near an addressee), since it does not have the defects of the other two [autos
and ekeinos] can properly be used vocatively. T h e r e f o r e it is foolish to criticize the
classical (archaioi, Pre-Hellenistic) authors [who use houtos as a vocative] solely on
the g r o u n d that third person p r o n o u n s are incompatible with the vocative case. In
fact it is not the third person as such, but certain accidental properties of these
p r o n o u n s [namely absence or distance] that preclude vocative use, since all nouns
are third person and still are not excluded from the vocative case. [[A.D. again fails
to mention hode "this near me", which is never used as a vocative, no doubt because
of its strong first-person connotations.]]
43. Next we must discuss the vocatives of possessive p r o n o u n s . [[These, as
derivatives of personal pronouns, are the only other forms counted as p r o n o u n s
by A.D.; all other words which we might consider p r o n o u n s are classed with
nouns.]] Possessives of the first person have vocatives both by the relevant theory
and by actual usage. For it is possible to address one's possessions [i.e. relatives,
friends, etc.], just as it is to use first-person possessives with verbs of being in the
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 171
second person, e.g. hemeteros huparcheis philos("you are o u r friend") emos eignorimos
("you are my acquaintance"). [[I.e., since you can have first-person possessives
applied to an addressee, theory says you should be able to have them in the
vocative, the case of the addressee.]] A n d for examples of usage, see Odyssey 3.475
where Nestor addresses his own sons as paides emoi ("my sons!"), and again, in
Callimachus (fr. 97 Schneider):
3.23 o emai tou apiontos ankurai
("O my anchors, anchors of me the returner")
a n d collectively by Athena, including [Zeus'] other children in the p r o n o u n (e.g.
Odyssey 1.45):
3.24 o pater hemetere Kronide
("O our father Kronides!")
44. As far as the theory is concerned, the third person possessive should have a 220
vocative; it's quite possible to address someone else's possession, as can easily be
seen in the case of possessive [i.e. patronymic] names. For these nouns, which refer
to two distinct third persons, are used in the vocative, e.g.Aristarcheie ("O Aristar-
chean!" i.e. "follower of Aristarchus!"), Aianteie ("O descendant of Ajax"),
Telamonie ("O Telamonian!" i.e. "Ajax, son of Telamon"). A n d this might be
expected to h a p p e n with pronouns; for instance, one might expect sphetere ("O
their (son)!"). But in fact, I have found no examples of this usage.
4 5 . In the case of second person possessives it is not only that there are no
examples of a vocative in the tradition, but theory also shows such a form to be
impossible—not, however, for the reason given by Habron, that an internal
contradiction of n u m b e r would result. "For *humetere," ("O thou who belongest to
you-all") he says "is plural as regards the possessor, but singular for the possessed."
In that case, obviously, if the p r o n o u n is made doubly plural, if one says *o
humeteroi ("O ye who belong to you!"), the result should be an acceptable vocative.
[But this is obviously not the case.]
46. T h e reason for the non-existence of [the vocative of the second-person
plural] p r o n o u n is rather the conflict of case. T h e reasoning goes as follows. Every
plural person is a collection of singular persons in the same case, either a collection
(sullepsis) of forms of the same person or of forms differing in person; e.g., of
different persons:
3.25 erne kai se kai touton etheasato Truphon 221
("Tryphon saw me, you and this man")
which becomes, when collected, hemds etheasato Truphon ("Tryphon saw us") or emoi
kai soi kai tois peri Dionusion elalesen Truphon("Tryphon spoke to me, you and
Dionysius") which becomes hemin elalesen ("He spoke to us"). An example with
sameness of person:
3.26 se kai se etheasato Dion
("Dion saw you and you")
which equals humas etheasato Dion ("Dion saw you-all"), or touton kai touton mem-
phomai ("I find fault with this one a n d this one") which is equivalent to toutous
172 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
memphomai ("I find fault with these (men)"). Once this is granted, it must also be
given that all second person [plural or dual] forms include the addressee in their
reference, e.g. humon, humas ("you-all" gen., acc.). But a vocative *humeteroi ("Oh
you-all of you-all"!) is doubly second person, since the possessors are addressees
because of the p r o n o u n form [second person], a n d the possessed are in the
vocative. Therefore, it is not just a matter of usage that supports o u r restriction on
this p r o n o u n [that it has no vocative], but also the a r g u m e n t just given.
47. I n the case of emos ("my"), however, it is usage that has rejected the vocative,
which would be h o m o p h o n o u s with the accusative singular personal p r o n o u n erne
("me"). [[And in Latin, for example, the equivalent vocative form mi "my", is
common.]] A n d perhaps this is the only reason for the deficiency, since [a vocative
*eme "my"] if used, would give the impression of an e r r o r of g r a m m a r . So
inevitably, the Attic figure [nominative for vocative] was used by the poet [Homer]
when his text required him to use a vocative at Odyssey 19.406:
3.27 gambros emos thugater te 222
("O my son-in-law a n d daughter").
[[Even though gambre is possible, *gambre emos is ungrammatical.]] How could we
refuse him this privilege, in view of o u r present a r g u m e n t a n d the fact that in the
other genders of the singular he could freely use the vocative, [which, in those
genders, is identical with the nominative anyway], n e u t e r in Odyssey 11.155:
3.28 teknon emon
("My child!")
and feminine in 11.164:
3.29 meter erne
("My mother!")?
Given the existence of these forms, it was inevitable that the masculine should also
exist.
48. We explained above [Bk. 2, 117-124] about the h o m o p h o n y of emou [either
"of me" or "of my" masc. or neuter sg.], showing how to tell when it is the genitive
of the possessive p r o n o u n ["of my"] a n d when the personal p r o n o u n ["of me"],
and we discussed the Zenodotean readings. A n d it's also clear that the form emoi
[either "to me" or "my" n o m . pl. m a s c ] shows h o m o p h o n y of the [nominative]
plural [masculine] of the possessive ("articular") p r o n o u n with the dative singular
of the personal p r o n o u n ; a n d that this h a p p e n s in the second person [soi "to you"
or "your" nom. pl. m a s c ] as well as the first, but not in the third. For when not
enclitic, the singular personal p r o n o u n [hoi "to him"] has the circumflex accent,
whereas the possessive [hoi "his, her, its" n o m . pl. m a s c ] has the acute. And,
naturally, when the [dative singular] personal p r o n o u n s [moi, soi, hoi] are enclitic,
there is no h o m o p h o n y [since the possessive p r o n o u n s soi a n d hoi are never
enclitic, and emoi, in addition, has an e- on the front]. That's why in lliad 23.387:
3.30 (=2.150) hoi de hoi eblaphthesan
("But his [horses] were hampered")
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 173
from hos eiromenos me. [[The point of the r e o r d e r e d underlying forms in both cases
appears to be to make a verb form follow the relative p r o n o u n immediately. Note
that in the second example he is satisfied with the participle though that does not
count as a verb in the strict sense for A.D. Presumably "verb or participle next after
relative" is A.D.'s rule, and any other surface o r d e r is d u e to some sort of later
"scrambling" rule.]] It is impossible to p r e s e n t h e r e all the instances of
homophony in all the parts of speech. What we have offered is sufficient to
exemplify all other cases of this sort.
50. (X) It wouldn't be right, here, to ignore the question of n u m b e r agreement
with neuter subjects: why it is that with masculine or feminine plural subjects the
ungrammaticality is patent if the verb is put in the singular, e.g. *hoi andres legei
(*"The men says") or *hai gunaikes legei (*"The women says"), but not so for n e u t e r
forms (prophora) even if the meaning is exactly the same and only the surface
gender is altered, as in ta gunaia legei ("the women [diminutive form] say"—here
gunaia "women" is neuter plural, yet legei "says" in the singular is correct). A n d it's
no explanation at all to say this is a Boeotian way of speaking, as in Pindar [fr. 75, 18
Sandys]:
3.38 acheitai omphai meleon sun aulois 224
("There echoes the sounds of songs with flutes." [This is called "the
Boeotian figure"; in poetry the verb regularly precedes in such
cases. In prose it is a solecism.])
This is precisely the point that needs explanation, how it h a p p e n s that in the
neuter the ["Boeotian"] figure passes unnoticed. A n d that the choice of g e n d e r [in
itself] offers no explanation of the fact that neuters have this affinity for singulars
is clear from the fact that with indeclinable words (aptoton) [including verbs] the
choice of gender [regularly] makes no difference. For every indeclinable has one
and the same construction with all genders.
5 1 . One can also reason as follows. We have just shown [Ch. 13 etc.] that no part
of speech can be ungrammatical in respect to a category which it fails to distin
guish, e.g. adverbs cannot be wrong in r e g a r d to n u m b e r unless they are numeri
cal adverbs, nor in tense unless they are time-distinguishing adverbs like chthes
("yesterday"), aurion ("tomorrow") and the rest that we mentioned [in Ch. 19—
mainly in the gap in the MS that occurs there]. A n d conjunctions, which cannot
mark number, cannot be ungrammatical in n u m b e r , either—or gender, either,
for that matter. T h e application of the a r g u m e n t to all parts of speech is obvious.
(XI) Now verbs do not distinguish gender, but only n u m b e r , person, and the
other verbal categories. So they o u g h t not to treat neuters in one way a n d
masculines or feminines in a different way. So how can it be that we say legousin hoi
anthropoi ("the m e n say" with plural verb, masculine subject) and legei ta paidia
("the children say" with singular verb and neuter subject)? If verbs mark n u m b e r ,
obviously they should agree in plurality with associated plural [subjects], and
clearly, in Iliad 2.135, sparta leluntai ("the [ships'] rigging has gone slack" neuter
plural subject, plural verb) is more correct-in-theory {analogos) than doura sesepen 225
("the timbers are rotten" n e u t e r plural subject, singular verb) [see below, ex. 3.56].
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 175
52. It's possible also to test this figure with examples from the first a n d second
person. If it is right to say philoponoumen paidia onta ("we work h a r d when we're
children") but not *philopono paidia onta (*"I work h a r d when I'm children"), and
the same way in the second person, how can you help granting that philoponei
paidia onta ("He/they work hard, being children") is against reason? [[In fact, that
sentence would probably be avoided in fifth century Attic, as would any sentence
where the plurality and diversity of the subject, especially a h u m a n subject, is
relevant.]] And that is a proof of the natural correctness [of plural verbs for all
plural subjects].
53. T h e reason, I suppose, for the acceptance of the singular verb construction
[with neuter plural subject] is the h o m o p h o n y of nominative a n d accusative
[neuter], something which does not h a p p e n to masculines and feminines. [[A.D.
overlooks words like triereis "triremes", poleis "cities", pecheis "cubits" and often
basileis or basilēs "kings", as well as all masculine a n d feminine duals of any
declension.]] As a result the ungrammaticality (the non-agreement) is im
mediately obvious in the nominative, e.g. *hoi andres graphei ("The m e n writes"),
but this wouldn't be so in the accusative [since verbs do not agree with direct
objects]. T h e n it would seem as if the andras in andras graphei ("He is drawing
men") were the direct object ("person to whom transition is made"). W h e n cases
are fully differentiated in form, then it is easy to distinguish grammatical from
ungrammatical. But since this is not the case in the neuter form (schema), it
h a p p e n e d that the wrong n u m b e r seemed more or less natural—as if it were an
object e.g. graphei ta paidia ("The children are writing"; it could also mean "He is
painting the children") is accepted as if parallel to graphei tous paidas ("He is
painting the boys"). A n d since the two cases in the masculine differ in phonologi
cal shape—the nominative giving *graphei hoi paides, which is ungrammatical—but
not in the neuter, because of the h o m o p h o n y with the accusative the non-
agreement in the construction is h a r d to spot. Now I've shown both why the
construction is [theoretically] ungrammatical, a n d how it managed to become 226
accepted without the ungrammaticality being noticed. [[Buttmann has a nice
footnote here: "Zu solchen V e r i r r u n g e n gelangt die Grammatik, wenn sie syntak-
tische Erscheinungen lediglich aus d e r ausseren Form erklaren will." But, in fact,
A.D. is far more concerned with underlying form and abstract semantics than
grammarians of Buttman's day, and this whole a r g u m e n t is largely designed to
find a psychological key to the problem of linguistic change. Indo-Europeanists in
the 19th century as well as today assume that A.D. was dealing with a pseudo-
problem, that neuter plural subjects were originally collective singulars a n d never
had plural verbs until fairly late (e.g. sometimes in Homer), so there was no change
from agreement to non-agreement as A.D. believes. So far as I know, this is the
earliest discussion of the social psychology of language change.]]
54. (XII) Next we must speak of the general syntax of verbs, which I consider,
because of its many aspects, to need very serious study. (a) T h e inflections-for-
mood (enclisis) and their constructions need analysis; (b) the tenses into which they
are subdivided; (c) and the inflection for voice, active or passive, and the middle
voice which lies between these two; (d) the category of person, which appears
either t h r o u g h o u t a tense, or partially [as in the imperative, which has no first
person form], or not at all [as in the infinitive]; (e) a n d whether all verbs have both
176 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
56. Why not? Because they're adverbs? (a) "Different moods (enclisis) in the same
person [i.e. with same subject] are never construed together: we can't say *grapheis
lexais (*"you are writing may you say") or any such combination [[this is more or
less true in Greek; in Latin, however, it is not so unusual]], but we do say theleis
graphein ("You want to write") and anaginoskein boulei ("You want to read") [where
the words in -ein are infinitives]. (b) a characteristic feature of adverbs is to be
construed with verbs, both preceding a n d following them, and this h a p p e n s with
infinitives: graphein thelo Ithelo graphein ("I want to write") just like Hellenisti lego/ lego
Hellenisti ("I speak Greek"), (c) and just as an adverb like Hellenisti ("in Greek")
requires a verb, e.g. lego ("I speak") to make a complete sentence, so graphein ("to
write") requires (proslambanei) a verb like thelo ("I want"); (d) and just as adverbs are
indifferent to variations of n u m b e r , so are infinitives—graphein thelo ("I want to
write") or graphein thelomen ("we want to write"); (e) a n d as person does not affect
adverbs—Hellenisti lego ("I speak Greek") or Hellenisti legeis ("you speak
Greek")—so also with infinitives—graphein thelo ("I want to write"), graphein theleis
("you want to write"); (f) a n d if adverbs are sometimes derived from verbs, e.g.
Hellenisti ("in Greek") from hellenizein ("to speak Greek"), nothing is to prevent 228
graphein ("to write") coming from grapho ("I write"). A n d the variation in tense
proves nothing, since some adverbs also refer to time, and exactly the same
behavior is exhibited by participles [which count as a distinct part of speech, but
not as adverbs], since graphon ("writing") appears beside grapho ("I write"), and
grapsas ("writing" perfective) beside egrapsa ("I wrote"). In exact analogy graphein
("to write" imperfective) matches grapho ("I write") and grapsai ("to write" perfec
tive) matches egrapsa ("I wrote")."
57. T o all this we may reply as follows. As regards the construction of thelo
graphein ("I want to write") and the others like it, we may say that it is not a general
rule that excludes different moods (enclisis) with the same subject (proso'pon "argu
ment", "referent") from co-occurring in a construction: we can say ean anaginoskeis
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 177
proseche ("If you're reading, pay attention") ean dialegei epistrephe seauton ("If you're
talking, t u r n around") [[these are apparently schoolroom examples, like all the
sentences about beating a n d skinning alive earlier; a n d to us they also seem a little
off the mark, since there is a clear clause break in the middle, which is not the case
with object infinitives.]] A n d besides that, the construction [of verb plus infinitive]
is not completely general either; we do say proairoumai anaginoskein ("I prefer to
read") a n d philo graphein ("I like to write"), but not *gelo graphein (*"I laugh to
write") or *skallo legein (*I dig to speak") [[the English sentences are possible to
interpret with infinitives of purpose; this is not possible in Greek]] as we would say
if infinitives were just like adverbs in this respect, for there wouldn't be a thing to
stop us from using adverbs with [any of these] verbs. [[Of course no one adverb
will fit with every verb, except perhaps the negative ou ("not"), but there is no
special class of verbs which "take adverbs", as there is a class which take infini
tives.]]
58. T h e reason for this construction [verb plus infinitive] is as follows. Some
verbs contain the notion of events or acts, and are sufficiently s u p p o r t e d by the
indicative mood (enclisis) and the other forms, e.g. grapho ("I write"), eresso ("I
row"), tupto ("I beat"); but others specify merely a preference of the mind, without
mentioning the event or act [which is preferred,] such as thelo ("I am willing"),
boulomai ("I want"), prothumoumai ("I prefer") which are, so to speak, empty of
content until they are filled by the adjunction of the event or act, which is signaled
by nothing else but the infinitive, the most general (least specialized) of all verb
forms, as we will show in the next few chapters [and also cf. 1.51] thelo peripatein 229
("I'm willing to walk"), boulomai graphein ("I want to write"). T h o s e verbs, however,
which already contain the mention of an act or event inherent in their essence are
naturally in no need of infinitives to complete the structure; since the act tuptein
("beat") is contained in tupto ("I am beating"), it's futile to add on another act and
say *tupto graphein ("I beat to write") but not so in the case of thelo graphein ("I want
to write").
59. As we just said, the infinitive form (enclisis) is the most general [form],
necessarily omitting all reference to the categories we wondered about before—
person and its inseparable associate, n u m b e r , which is not as such a category of the
verb, but appears along with the subject persons ["persons who participate in the
action" one of A.D.'s clumsy substitutes for a technical term meaning "subject"].
T h e action (or event) itself is single—to graphein ("to write"), to peripatein ("to walk")
[here the infinitives are accompanied by a n e u t e r singular article to-"the writing",
"the walking"—which can never be plural]. A n d when this act occurs in connection
with persons it produces peripato', peripatoumen, peripatousin ("I walk, we walk, they
walk"). Nor is it the case that the verb is susceptible of person; this, too, arises by
association, in the following way. T h e persons [i.e. the referents of NP subjects]
who take part in the action are classified into [first, second, and third] persons—
peripato, peripateis, peripatei ("I walk, you walk, he walks")—but the verb itself,
unaffected by person a n d number, will freely go with all persons and all n u m b e r s .
F u r t h e r m o r e , the verb in itself has no mood ("mental attitude"); here, too, it is the
subject persons who express their own mental attitude by means of the verb. But
infinitives, since they have not yet acquired subject persons, naturally also have no 230
178 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
63. (XIV) T h e Homeric practice of using the infinitive form (enclisis) by hypal- 232
lage [i.e. as a figure] for the imperative construction, I think, is also explained by its
generality, the fact that all special [moods] can be transformed into infinitives. T h e
same is true of the use in the epistolary construction (the complimentary opening
of letters)—Truphon Theoni chairein ("Tryphon "to T h e o n , to rejoice" i.e. "greet
ings"), although it differs from the Homeric usage in that while the latter
[Homer's use] naturally resembles the imperative pattern and can always be
replaced by an imperative, but belongs to the class of figures that are conceded to
poetic practice, the former, the epistolary construction, cannot be so replaced.
[[I.e. Greeks never used an imperative chaire or the like at the beginning of a letter.
But A.D. h e r e argues that there is a grammatical reason for this, not just a
customary one.]] T h e reason is as follows.
64. As we said above [2.42], [in this formula] the nominative of the p r o p e r name
[of the sender] is used, a n d also the dative [of the addressee's name], which is
required by this construction, so that it cannot allow the imperative mood—or any
other [except the infinitive]. For Dionusioi chairein ("to Dionysius, to rejoice") is
equivalent to Dionusie chaire ("O Dionysius, rejoice"), which cannot be used h e r e
because the imperative mood and the vocative case are a p p r o p r i a t e only for
[speaking to] persons who are present, but the one to whom the letter is addressed
is absent, and the writer, though present [i.e. at the.moment of writing he is with
the letter] uses a construction appropriate for one absent [since he is writing from
the viewpoint of the addressee]. T h e nominative of names has been shown [2.43]
to be necessarily third person. (We a r g u e d [2.42] that the use of nominative [first] 233
personal p r o n o u n s is not p r o p e r [in letters] since they point at (refer deictically to)
present persons. And this [pointing] is impossible in this case, since the two parties,
the sender a n d the receiver, are out of each other's sight.) T h e r e f o r e , since both
names are in the third person, both the nominative one and the dative one, the
verb that belongs to them must also be in the third person, and be one which can
appropriately be construed with the nominative n a m e as subject.
65. So the verb in question might go with the nominative—Apollonios Dionusioi
chaireto ("let Apollonius rejoice in Dionysius") or chairei (". . . rejoices. . . ") or
chairoi ("May. . . rejoice. . . "). But while we may g r a n t that this makes the sentence
grammatical, it is perfectly clear that it is the subject (nominative) person who is
doing the "rejoicing", not the indirect object (dative) one. But in fact we want the
opposite to be the case, to have the addressee linked to rejoicing, just as when we
greet people in person we say chaire ("Rejoice!"), a n d d o not make ourselves the
subject—*chairoimi ("May I rejoice!") or *chairo ("I rejoice"). Since this solution is
impossible, we must clearly transform the construction into the general form, the
infinitive, u n d e r s t a n d i n g the usual ellipsis of legei ("says") or euchetai ("wishes"):
Truphon Theoni legei chairein("Tryphon says to T h e o n to rejoice"), j u s t as we have
other c o m m o n ellipses, like [that of chersin "hands"] in Odysseyl8.2S kopton am-
photereisin ("punching with both [hands]"), or [the ellipsis of Zeus or the god] in huei,
brontdi ("It's raining", "it's thundering"). A n especially good parallel is the implicit
[performative] verbs which are inherent in the moods (enclisis), as we have shown
[Ch. 61] that chairois ("may you rejoice") contains the verb euchesthai ("pray, wish")
and chaire ("rejoice!") contains a form of prostaxai ("order, command") which
180 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
preceded by the adverb eithe, a n d the imperative by age, so also the infinitive has as
its special adverbs chre and dei.
69. (b) And no one can suppose that they are conjunctions, since they do not
require a second clause (logos), as conjunctions do by definition, a n d they signify
intensity, like mala ("very") and agan ("too"), and, besides that, they form com
p o u n d s [with prepositions], which never h a p p e n s to conjunctions, but only to
verbs and adverbs: we say, e.g., apodis ("twice") apechthes ("yesterday") [[neither of
these words is well attested outside A.D., so they may be contemporary items with
somewhat different meanings; see 4.3]], apelthen ("he went away"), apestin ("he's
absent"); just so we have apodei ("it lacks") and apochre ("it's enough").
70. (c) Another supporting a r g u m e n t is this, that adverbs can occur with verbs of
all persons and numbers—kalds graphd ("I write well") a n d kalds graphomen ("we 236
write well") a n d so on, [kalds "well" being always the same]. But verbs change in
person and n u m b e r to agree with their subjects (prosdpon)—hemeis graphomen ("we
write") but humeis graphete ("you [pl.] write"). But this is not so of chre and dei—chre
hemas graphein ("We must write"), chre erne graphein ("I must write"), a n d dei se
anagindskein ("You [sg.] should read"), dei humas dialegesthai ("You [pl] should
talk"). And for this reason even more, dei a n d chre are adverbs.
7 1 . (d) Neither is it possible to claim that they are infinitives, on the g r o u n d that
they make no distinction of n u m b e r or person, a point we discussed above [Ch.
55]. Unfortunately neither the morphological characteristics n o r the syntactical
ones fit. What infinitive is there with an ending -ei or -ē? W h e n do two infinitives
make a grammatical sentence (logos) with one nominal? But we can say dei peripa-
tein Dionusion ("Dionysius must walk"). T h e r e f o r e these forms cannot be p u t into
the infinitive category. A n d they won't fit as verbs in general [i.e. present indic
atives], because, once again, they are not inflected for person, or n u m b e r , either,
which is a defining feature shared by all verbs.
72. In reply to this one may say (a) if they are adverbs, specialized to occur with
infinitives, as eithe is specialized for optatives a n d age for imperatives, why can't
they go with nominative personal pronouns? You can say eithe ego graphoimi ("I
h o p e I'm writing"), eithe hemeis graphoimen ("I h o p e we're writing"), age graphdmen 237
hemeis ("Come on, let's write"), age graphete humeis ("Come on, you guys write!"),
but you can't say *dei egō graphein ("*Me must write") or *chre humeis graphein (for
humas, to mean "you[pl.] ought to write"), although the rules for adverbs in no way
prohibit these constructions, [since all cases may co-occur with them] e.g. eithe
humeis akououte ("I hope you listen", where "you" is nom.), eithe humas theasaito ("I
hope he sees you" where "you" is acc.), a n d eithe humōn akouoi ("I h o p e he hears
you" where "you" is gen.), and so on. [[Of course this is a lousy a r g u m e n t since the
case of the p r o n o u n s is determined by their relation to the verb, a n d nominatives
may occur with finite verbs, as subjects, but not with infinitives. T h e adverbs have
nothing to do with it, nor would dei a n d chre if they were adverbs.]]
73. (b) O n e of the distinguishing features of verbs, but not of adverbs, is to take
the syllabic a u g m e n t in past tenses. We say semeron grapho ("Today I'm writing")
a n d semeron egraphon ("Today I was writing" [[with the prefix e-, called "syllabic
augment", a d d e d to past tenses of verbs with initial consonants; vowel-initial verbs
182 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
lengthen the initial vowel, which is called "temporal augment"; both are probably
meant by A.D.'s expression "chronos exothen" "a time externally" i.e. "one m o r a
initially"]]), with an initial addition to the verb, but n o n e on the adverb. But we
likewise say dei graphein ("[He] has to write") and edei graphein ("[He] had to write"),
and the form edei exhibits the same augment, and closely resembles eplei ("He was
sailing" from pleo) and epnei ("He was breathing" from pned), but there is no
a u g m e n t on graphein ("to write"). H e r e we have a clear e n o u g h proof that the form
that narrowly escaped being called an adverb is actually a verb, while graphein,
about which doubts were expressed as to whether or not it was a verb, we proved
above [Ch. 55-61] beyond any doubt to be, not merely a verb, but the most general
verb (i.e. the most purely verbal) of all verb forms. Besides our arguments there,
we should mention that it also takes reduplication, which is an exclusive feature
(idion) of verbs [[not really, since perfect participles have reduplication, and A.D.
does not count them as verbs]], as in gegraphenai ("to have written" from grapho "I
write"; [[the syllable ge- h e r e with the vowel e preceded a n d followed by the same
initial consonant, is called the syllabic reduplication]]) etc. We do not mean to say
that all past tense forms [[he really means all non-indicative forms, which were
regarded by A.D. as corresponding both with the basic past tenses a n d the basic
present tenses of the indicative]] which lack augments are not verbs, since n o n e of
the other moods (enclisis) a d d the augment, only the indicative; but it is true that all 238
forms which have the a u g m e n t belong to the verbal category,, a n d we have shown
that both echrin and edei have the augment. Also evidence that they are verbs is the
fact that they form infinitives—chrinai and deisthai [[actually deisthai doesn't really
belong to dei, whose infinitive is dein]] just like stenai ("to stand") a n d pneisthai ("to
be blown or breathed on", said e.g. of wind instruments).
74. (c) But one can also get evidence from the accentuation. Dei has a circumflex
just like plei ("He is sailing"). "Well," you may object, "aren't there some adverbs in
-ex with circumflex, e.g.pei ("where?"), autei ("right here"), toutei ("here")?" I n the
first place, those adverbs belong to the Doric dialect, which is not the case with dei.
In the second place, the forms pei, hex ("where") a n d autex coexist as [Doric ]
equivalents with pou ("where") hou ("where?") and autou ("right here"), which is
not the case with dei [i.e. there is no variant *dou]. In fact dei is contracted from deei,
just as plei is frompleei ("sails"), rhei fromrheei ("flows"), or chei from cheei ("pours"),
and its first person form is deō ("I lack, I miss"), like pneō ("I breathe"). T h e r e has
also been a phonological change in the case of chrē, one which can be paralleled in
another verb. T h e meaning of dei is endeomai ("I lack, I need"), which has as
synonym chrd (contracted from chreo) [[actually only the middle of this verb, which
we treat as contracted from chrao, has the meaning "want", "lack", "need"]], a n d
the two verbs have the parallel derivative n o u n s deos ("fear"—not in fact related to
ded or dei) a n d chreos ("need"). T h e verb chrd gives rise to the derivative chrimi
[[A.D., like o t h e r a n c i e n t g r a m m a r i a n s , d e r i v e s all -mi v e r b s f r o m a n
underlying—usually "abstract", i.e. unattested—regular -6 verb. I n this case, how
ever, there is no *chrimi attested but only a kichrimi.]] like phimi ("I say") of which
the third person is chrisi likephisi, from which chri is derived by apocope of the last
syllable, parallel to phē as used by Anacreon [fr. 16 E d m o n d s ] :
3.42 se gar phi Targilios emmeleds diskein
("For Targelios says that you throw the discus elegantly.")
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 183
[[Actually we now believe that chre was in origin a n o u n "necessity, need", and that
echren a n d chrenai were later developments which had the effect of shifting it into
the verbal category.]]
75. (XVI) (d) So much for phonological-morphological arguments: now we
must supply syntactic ones. Since [dei and chre] are verbs in the indicative, it seems
that they are deficient [in remaining unchanged] with different persons and 239
n u m b e r s [reading endeos "deficient" with Uhlig rather than henikōs "singular"];
this would not h a p p e n if these different persons and numbers occurred [as
subjects]. But in fact in dei hemas graphein ("we ought to write") it is graphein ("to
write") which is the [subject] of dei ("is necessary"), not hemas ("us"). This is why in
this construction the verb [dei or chre] never changes for person and n u m b e r , since
its subject (hoi sumpheretai) is always an infinitive [which counts always as neuter
singular], and therefore, since [the subject] never changes [in this respect], the
verb naturally remains unchanged along with it.
76. (e) T h e semantics (deloumenon) of the construction is as follows. W h e n there is
some failure or lack of some acts or events, the presence of these verbs u n d e r
discussion [dei and chre] urges the performance of the deficiency. W h e n you say dei
philologein ("[We] ought to study"), you're essentially saying epei leipei to philologein,
philologomen ("since there's a shortage of studying, let's study"). If we just use this
verb leipei ("is lacking") by itself, the construction will be quite as deficient [as the
one with dei]: leipei to philologein hemas ("Our studying falls short"), leipei to philolo
gein erne ("My studying falls short"), and yet no one is likely to claim that leipei is an
adverb, even if this pattern were repeated without limit, because of the fact that
the subject (he sunousa suntaxis) is an infinitive.
77. (f) This one reason is also enough to block the insertion (paralambanetai) [of
dei or chre] with the infinitives in epistolary [openings]. We have already said [Ch.
66] that a verb such as legei ("says") or euchetai ("wishes") must be understood with
such infinitives—Dionusioi Apollonios legei chairein(or euchetai) ("To Dionysius Apol- 240
lonius says [or wishes] joy"). A n d it's impossible for a single infinitive to have two
indicatives governing it, rather than just one. T h e r e f o r e we say thelei graphein ("He
wants to write"), or dei graphein ("[One] must write"). But since legei or euchetai is
implicit (enkeitai) inDionusioi chairein ("for Dionysius to rejoice"), it's impossible to
p u t in a chre or dei on top of it. But if we leave the epistolary construction, then we
can, of course, say dei chairein ("[One] should rejoice"). But in a letter the addressee
is supposed to have joy added, not subtracted, which is what dei chairein implies [in
the sense "there is a lack of joy"].
78. Next we must discuss the construction of cases with infinitives. A n d first we
must consider whether it is true that infinitives always have-subjects-in (pheretai
epi) the accusative, as was the case with chre and dei-chre anaginoskein Dionu-
sion("Dionysius should read")—and other similar verbs. T h e correct account is
rather not that infintives as such always require accusative [subjects], but that they
take the case that is governed by the indicative or other mood (enclisis) [of the verb
on which they d e p e n d ] . [[I.e. the accusative here is really an object of dei or chre.
This analysis becomes very awkward with articular infinitives like the example in
Ch. 76—leipie to philologein erne, which can also be o r d e r e d as leipei to erne philolo
gein, but not leipei erne to philologein, which last might mean "studying deserts me",
184 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
but not "my studying is deficient" as in the other orders. A n d many similar verbs
cannot take an accusative in any sense.]] T h e theory of this is as follows.
79. It is chre and dei themselves which govern the accusative construction
(parathesis), requiring it in their capacity as verbs just as many other verbs have a
similar requirement, calling for a genitive or a dative or an accusative. I n exactly
the same way that leipei ("is lacking", "abandons") governs ipheretai epi) an 241
accusative—leipei Diona ("he abandons Dion" or "Dion r u n s out of it") leipei erne ("it
leaves me", "I r u n short of it"), just so dei takes an accusative dei erne, dei se. [[These
actually do not occur; he p e r h a p s has constructed them as underlying forms for
deomai "I need (it)", deei "you need (it)", derived by passivization from dei erne, dei
se.]] This is the basic grammatical [structure] of the sentence dei erne akouein ("I
must listen" [i.e. akouein "to hear" is subject of dei, and erne is direct object "hearing
is-needful-to me"]). It is not so, as we said, that the use of infinitives in itself
requires [an accusative subject]. An instance to the contrary is the epistolary
construction we discussed earlier, where there is no accusative [but only a dative,
as required by lego ("I say") or euchomai ("I pray")]. Similarly in sentences like the
following—toi peripatein hedomai ("I enjoy walking"), peripatein thelo eper graphein
("I'd rather walk than write"), ethelei koimasthai e peripatein ("He'd r a t h e r sleep than
walk"). [[In these sentences, where the u n d e r s t o o d subject of the infinitive is the
same as that of the main verb, an accusative in fact is never used—*"I enjoy for
myself to walk" etc.—though some contemporary linguists would assume that it
h a d been present underlyingly and then deleted by "Equi-NP".]] If you substitute
dei for the main verb in any of these, then certainly there is an understood
(hupakouetai) accusative dei peripatein ("[I] must walk"), chre dialegesthai ("[We] must
talk"). [[In this case, of course, there is no "Equi-NP" a n d the identity of the subject
can only be inferred from the situation.]]
80. Don't infinitives presuppose accusatives? No, not in general, only where the
accusative is also associated with the other moods (enclisis):philei Theona ("He likes
T h e o n " ) , philei Truphona ("Love Tryphon!"), ean philei Truphona ("If h e likes
T r y p h o n " ) , and so also phesi philein Truphona ("He says he likes T r y p h o n " ) .
Because of this you may sometimes have two accusatives with an infinitive.
81. Infinitives enter into construction with genitives or datives also when the
other moods (enclisis) take the genitive—akouei Truphonos ("He hears T r y p h o n
[gen.]")—or dative—didosi Truphoni ("He's giving [it] to T r y p h o n [dat.]"), and so
on for the other moods, of which the transformed (metalambanetai) infinitives will
take the same cases—phesin akouein Truphonos ("He says he hears Tryphon"),phesin
didonai Truphoni ("He says he's giving [it] to T r y p h o n " ) . Now if we replace the 242
phesin in these sentences by dei, as we have shown, an accusative will be added—dei
akouein Truphonos Apollonion ("Apollonius must listen to T r y p h o n " ) , dei eran
Theonos erne ("I must love T h e o n " ) , dei soi Truphona charizesthai ("Tryphfon must do
you a favor"). [[Many infinitives, like these indirect discourse ones, A.D. derives by
transformation from other moods, usually indicative; but at a d e e p e r level most
i n d e p e n d e n t indicatives, optatives a n d imperatives are equivalent to infinitives
d e p e n d i n g on performative verbs, as we have seen.]]
82. If one were to delete the accusative from such sentences, it would be clear
that the accusative was missing, e.g. if we were to say dei soi charizesthai ("[one]
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 185
should favor you") dei sou akouein ("[One] must listen to you"). For an accusative is
invariably u n d e r s t o o d here, which must be supplied to make the sentence com
plete. If it were not j u d g e d missing [in such sentences], its addition would have to
produce the effect of pleonasm—dei soi charizesthai erne ("I must favor you"), dei
akouein sou Diōna ("Dion should listen to you"). [But this is not the case.] [[Where
two variants exist, one with one m o r e word or phrase t h a n the other, the g r a m m a r
ian must always decide which variant is the underlying one; if the longer variant is
correct, t h e n the shorter one is derived by ellipsis, but if the shorter version is, then
the longer one arises by pleonasm. Such underlying forms are not really " d e e p " in
A.D.'s scheme.]]
Theona hubrisai Diona ("They say that T h e o n h u r t Dion"). H e r e it is obvious that the
first accusative is the subject, the one which is transformed from nominative—
Truphon hubrise("Tryphon hurt. . . ") goes to legousi Truphona hubrisai ("They say
that T r y p h o n hurt. . . ") and after these is a d d e d the person affected, namely
Diona ("Dion") or Theona ("Theon"). T h u s the first accusative with an infinitive
belongs to the actor's (energetike) relation [i.e. is the subject], whether one should
say periechei ho ouranos ten gen ("The heaven s u r r o u n d s the earth") from which is
derived legousi ton ouranon periechein ten gen ("They say that the heaven s u r r o u n d s 244
the earth"), or, conversely, periechei he ge ton ouranon ("The earth s u r r o u n d s the
heaven") from which comes legousi ten gen periechein ton ouranon ("They say that the
earth surrounds the heaven!") [[A.D. here must be arguing for an underlying
"shallow structure" obligatory word o r d e r Subject-Infinitive-Object, since other
orders do, in fact, occur—by hyperbaton, as he says in the next chapter.]]
87. This, too, is generally conceded, that the active aspect of the event precedes
the passive one. T h e recipient suffers only after he has received the effect of the
active behavior, as is clear from the respective denials (apophatika). For a m a n who
says ouk edeira("I didn't skin [you]"—i.e. "beat [you]") denies his prior relation to
the person who alleges a subsequent passive relation to the event of darenai ("being
skinned"). [[This is an awkward sentence, but seems to mean that a denial of the
action "I didn't beat [sc. anyone]" automatically entails a denial of the result.]] A n d
if this is correct, it is quite p r o p e r that the verb [infinitive] with the first accusative
should express the actor's (energetike) relation [i.e. agent, subject], a n d the second
accompanying accusative correspond to the passive one [i.e. object], since the
effects of the act are second [to the act]: sunebe erne philein Apollonion ("It h a p p e n e d
that I liked Apollonius" cf. above, ex. 3.44) which is an exact equivalent of ego philo
Apollonion ("I like Apollonius"). [[The tense seems to be wrong, but otherwise this
is very Generative Semantic, extracting the tense and the " h a p p e n " from a finite
verb Cf. Lakoff 1972.]] A n d it's obvious that we have hyperbaton in Iliad 5.118:
89. This is why the so-called asseverative (diabebaiotikos) and causal (aitiologikos)
conjunctions occur with this [mood]. We are 'determining' when we say gegrapha
("I have written") and 'asseverating' when we say hoti gegrapha ("that I have
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 187
93. (XXI) T h e indicative mood, (enclisis), which we are now discussing, some
times lays aside its affirmativeness and its right to the n a m e indicative, when it
enters into yes-no questions (eperotesis), as when we say gegraphas? ("have you
written?") or lelalekas? ("have you spoken?"), a n d if the corresponding affirmation 247
would not be true, we answer ou ("no", "not"), but if it is true that we have written
[or spoken] we say nai ("yes"). A n d so the interrogative, when completed by an
affirmative response, turns again into an indicative. T h e best proof of this is that
sometimes we do not use the word nai ("yes"), but repeat the same indicative verb
[with appropriate change of person and, presumably, intonation] as if the affir
mation were already implicit. I n reply to the question grapheis? ("are you writ
ing?"), we reply grapho ("I am writing"), or else, with reinforcement of the indic-
ativeness, doubling the affirmative force, nai grapho ("yes, I am writing"). [[This is
more natural in Greek than in English, a n d is even more so in languages like Latin
which have no p r o p e r word for "yes".]] This is the pattern in Iliad 1.286:
95. (XXIII) We should note that there is a difference between the optative
meaning (enclisis) conveyed by the verb form and the adverbial one, in that the
verbs signify the event-or-action along with the attitude (illocutionary force,
diathesis) of wishing—so graphoimi ("I h o p e to write") is a wish for the act of writing,
and philologoimi ("May I study!") for the act of studying—but eithe is no m o r e than a
word for "wish", with no indication in it of what is being wished for.
96. T h e same p h e n o m e n o n may be illustrated with many more examples, (a)
T h e word leukoteros ("whiter") contains an intensification (epitasis) of the quality
"white", and glukuteros ("sweeter") one of the quality "sweet". But words for
intensified quality in general are such as beltion ("better") a n d ameinon ("better").
[[The mss. all have the superlative aristos ("best") added here, which seems confus
ing a m o n g a g r o u p of comparatives.]] (b) A n d if these examples do have specific
meanings from which they take their form, a n d have merely been figuratively
(tropikos) e x t e n d e d to t h e g e n e r a l sense of intensification (just as halieus
["fisherman" derived from hal- "salt, sea"] is extended to one [who fishes] in a
river, and puxis ["box" from puxos "boxwood"] is extended to cover [boxes] made 249
of other materials), it is certainly beyond debate that heis ("one") differs from Aias
("Ajax", singular; there were two heroes of the name inlliad), in that heis is only the
name of the n u m b e r , whereas Aias conveys the notion of singularity in addition to
its own individual quality [this is a Stoic notion; the "quality" meant is, from our
point of view, the quality of being n a m e d Ajax. Cf. 2.45]. (c) Similarly, in compar
ing egō ("I") with grapho ("I'm writing"), we see thatgrapho contains the notion of
" I " along with the meaning of the action ("write") and the mood and tense,
whereas ego is a p u r e p r o n o u n , merely the n a m e for [the first] person, (d) Iliothen
("from Troy") differs from allothen ("from elsewhere") in the same way, since
Iliothen conveys the idea of [motion from] a place along with the indication of the
place-name, whereas allothen merely presents the place relationship, (e) A word
like tachistos ("fastest, quickest") signifies intensification along with a certain qual
ity, but agan ("very, too") is only a word for intensification, (f) Similarly with
grapson ("write!") and age (imperative particle). H e r e , too, age is only a word
signifying c o m m a n d , but the verb grapson specifies the act [of writing] a n d the
person and n u m b e r [of the person c o m m a n d e d ] alongside the inherent notion of
imperativeness. Similar examples can be a d d e d indefinitely.
97. T h e combination of aithe with opheles ("you ought") inlliad 1.415-16:
3.49 aittiopheles para neusin adakrutos kai apemon
hesthai
("I wish you had sat by the ships, free from tears a n d troubles.")
or of hos a n d opheles in Iliad 18.86-87:
3.50 hos opheles su men authi met' athanateis halieisi
naiein, Peleus de thneten agagesthai akoitin 250
("I wish you had stayed back with the immortal sea-goddesses, and
Peleus had taken a mortal wife."—here, too, most editors read
aith'opheles)
however, should not be considered a p u r e case of pleonasm. [[The imperfect and
aorist forms, with and without augment—ophellon, ophellon, ophelon, ophelon, -es, -e,
190 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
etc.—of the verb opheilo "owe" are freely used with d e p e n d e n t infinitives to
express unreal or contrary-to-fact wishes—"I wish it were so", "I wish it had been
so"—and the common addition of aithe=eithe or hos does not add or clarify
anything, but may perhaps intensify.]] For in the verb ophelon you have also the
specification (diakrisis) of person, which is not present in eithe. Obviously, this is
why eithe is a d d e d to optatives, which can then specify person, as in Odyssey 14.440:
3.51 aith' houtos, Eumaie, philos Dii patri genoio
("I pray, Eumaeus, that you may be as dear to father Zeus [as you
are to me]")
or in Iliad 4.178:
3.52 (=3.48) aith' houtos epi pdsi cholon telesei' Agamemnon
("I pray A g a m e m n o n may fulfill his anger on everyone as
much as. . . ")
And we also showed that eithe appears with [past tense] indicatives, as ine i t h 'egrapse
Truphon ("I wish T r y p h o n had written"), [which likewise are inflected for person].
But it doesn't go with infinitives, since neither the adverb nor the infinitive is
capable of making clear what person is the subject of the wish/prayer. A n d to fill
that gap, ophelon is inserted, for the expression of the person for whom the prayer
is made [i.e. the subject of the e m b e d d e d verb]. For in example 3.49, the subject is
understood as Achilles, to whom the speech is addressed [by his mother, Thetis],
for no other reason than the second person ending of opheles. So much for eithe a n d
related words. [[What A.D. does not make clear here is the variety of alternatives
for unreal wishes: "Would that T r y p h o n h a d written", to take his example, may be
either eith' egrapse Truphon or ophele grapsai Truphon or eith' ophele grapsai Truphon.
Clearly eithe is added to the second formula, not ophele to the first, as A.D. seems to
suggest, nor to an ungrammatical *eithe grapsai Truphōna.]]
98. (XXIV) Some people are puzzled by the use of [past] tense forms in this 251
function (enclisis, i.e. wish), arguing that the occurrence of past tense forms in this
modal-function {enclisis) is out of place, since this is an impossibility; just as in the
case of other words the inherent meaning of the base blocks the occurrence of
certain forms of the word, e.g. in verbs like ploutō ("I am rich") and huparcho ("I
exist") and similar intransitives there is no passive voice, or in verbs like machomai
("I fight") there is no active [since it is d e p o n e n t , and the passive form itself has the
active meaning], or, in regard to gender, a word like ektrousa [a poorly attested
word, apparently a second aorist participle of a verb meaning "to miscarry" or "to
have an abortion"] has no masculine, nor the word arsen ("male") any feminine,
[[this is not a good example; actually arsen can modify a feminine n o u n , so it's
really just a two-termination adjective, with feminine the same as masculine]], so in
the present case [past-tense wishes] the combination [of past tense with eithe]
doesn't make sense. A n d these objectors say "If wishes (or "prayers") are for the
coming to pass of something which does not yet exist, how can what is past have
any need for wish/prayer?"
99. T o this objection we can reply that, in fact, a wish/prayer in past time is a
necessity. Imagine that the appointed time for a particular Olympic contest has
gone by, and a father is praying for victory for a son who participated in that
B O O K III: ON SOLECISM; VERBS; PARTICIPLES 191
contest. Obviously he will not pray in the future tense or the present progressive,
since they are inappropriate for a past event; so consequently his prayer would be:
3.53 eithe nenikekoi mou ho pais
("May my son have won!")
3.54 eithe dedoxasmenos eie
("I pray he has been honored.")
[[Of course these are not the contrary-to-fact wishes which he set out to explain;
they are perfect optatives, not past indicatives.]]
100. We can also say that it is true that wishes/prayers are for things we don't
have. It's when we don't have philologein ("study") that we would say philologoimi ("I 252
hope to study"), and when we lack ploutein ("to be rich") that we say ploutoimi ("I
pray to be rich!"). We must note that the thing requested by the optative is either
something present which is to be extended so that it will keep on, as when one says
zooimi o theoi ("O gods, may I live!") or else it is something not in existence, which is
to be accomplished, as when A g a m e m n o n prays [not a quotation]:
3.55 eithe o theoi porthesaimi ten Ilion
("Grant me, O gods, to sack Troy!")
here the prayer looks to a time when the event [of sacking Troy] is past and
finished. [[This is one of the passages where A.D. sketches his theory of aspects—
which he does from tense—with imperfective aspect (as in philologoimi, ploutoimi
and zooimi) represented as continuing and incomplete, a n d perfective (aorist, as in
porthesaimi) as finished and complete.]] For he [Agamemnon] would consider
prolongation [of the sack] as something not to be wished for. As H o m e r says in
Iliad 2.134-5, while he has been sacking [really only besieging] Troy:
3.56 ennea de bebaasi Dios megalou eniautoi
kai de doura sesepe neon kai sparta leluntai
("Nine years of great Zeus have come and gone, and the ships'
timbers are rotten and the rigging slack.")
And just the opposite is the case with zōoimi, since no on would put into his prayer a
wish for the completion of life, e.g. zesaimi ("May I finish life!"). For the fulfillment
of such a prayer would put an end to the continuance of life. [[Note that A.D. gets
side-tracked in a discussion of aorist, perfect and imperfective optatives, and
never actually discusses unreal wishes, in which real past tenses of the indicative
are used. This is partly because he speaks of all perfect a n d aorist forms, whether
indicative or not, as "past tense", since he has no distinct name for aspect.]]
101. T h e same problem arises also with imperatives. H e r e , too, as in the case of
wishes, one makes commands for things that have not h a p p e n e d , and clearly what
is past has h a p p e n e d . So, by the same arguments one should never use the
imperative of a past tense [i.e. an aorist or perfect imperative]. H e r e , too, one can
offer the same defense, that there is a distinction between kleiestho he thura ("The 253
door be closed!" i.e. "Let it continue to be getting closed!" imperfective aspect) and
kekleistho ("Let it be closed" i.e. "Let it be in a closed state" perfect or stative aspect),
in that the form from the present [i.e. imperfective] signifies the near-future
192 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
command, since this is the special function of the present progressive, but kekleistho
indicates [a command] for a condition which ought to have been b r o u g h t about
long ago.
102. As we said, one type of c o m m a n d [i.e. imperfective imperatives] is con
cerned with duration (paratasis). Anyone who says graphe ("keep writing"), sarou
("keep sweeping"), skapte ("keep digging"), is giving a c o m m a n d for continuance
of a disposition (diathesis), as is the case in Iliad 8.282:
3.57 ball'houtos, ai ken ti phoōs Danaoisi geneai
("Keep shooting like this, a n d perhaps you may be a light to the
Greeks.")
where he [Agamemnon] is telling [Teucer] to continue occupying himself with
shooting in the battle. But [for the other type], anyone who says, using a past form
[i.e. the aorist or perfective imperative] like grapson ("Write!"), skapson ("Dig!") is
not only c o m m a n d i n g an act which is not already going on, but he is excluding the
act from duration, since to those who take too much time writing we say grapson
("Write!"), as if to say "Don't prolong things, finish u p your writing!".
103. (XXV) Second person imperative forms are already identifiable, even when
they are h o m o p h o n o u s with indicative forms (enclisis) [as in 2nd person dual or
plural, imperfective]. A vocative accompanying the form removes any ambiguity,
as does the imperative adverb age, as we will show in what follows (Ch. I l l , Ch.
177]. T h i r d person imperatives, since they always have their own special endings
(phone), are never ambiguous at all. [[Why A.D. believed that vocatives are a sure
sign of the imperative is a puzzle, since Greek, like English, can easily say "You're a
fool, Demosthenes" or "Are you coming, Demosthenes?" H o m e r , A.D.'s favorite
source, is full of such examples, beginning at Iliad 1.74 and continuing to Odyssey
24.511. See also the quotation from T r y p h o n in Chapter 35 above.]]
104. T h e possibility of first person imperative forms is denied by some people,
for the following reasons. T h e y say that the one giving the c o m m a n d must be 254
distinct from the one receiving it, "and this is not possible in the first person; the
first person is the way the speaker refers to himself, and the second the way he
refers to the one to whom he is speaking, a n d that's the one referred to by the
imperative mood—pheuge ("Scram!"), lege ("Speak!"), graphe ("Write!")."
105. It's also plain to see that vocatives involve two persons, the one calling and
the one called, and therefore there are no vocatives of the first person, but there
are second person vocatives. So, obviously, the facts of syntactic structure show
that there are no first person imperatives. For imperatives are construed with
vocatives, and it has been shown that first person vocatives are impossible. [[A
doubly fallacious a r g u m e n t , like many of A.D.'s.]] A n d j u s t as it is impossible to
conceive of anyone calling to himself, because of the inseparability of a person, so
we can't have anyone o r d e r i n g himself, for the same reason. For every imperative
is directed from a d o m i n a n t person to a subordinate one. [[A.D. h e r e forgets, as
other linguists sometimes do, that imperatives are regularly directed by h u m a n s to
divinities and by slaves to their masters ("Forgive me!" "Don't beat m e " etc.).]]
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 193
107. "It's also quite clear from the construction with imperative particles—age
legomen ("Come on, let's speak"), age arithmesomen ("Come on, let's count. . . ")
whose singulars are phere lego ("Come on, let me speak!"), phere arithmeso ("Come
194 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
on, let me count!"). Note that the second person indicative form, which is not
h o m o p h o n o u s with an imperative, does not allow this construction [with those
particles]—who would ever say *phere legeis ("Come on, you are saying") or *phere
arithmeseis("Come on, you will count")? Examples are from Odyssey 13.215:
111. It is evidently the fact that the imperative lacks first person forms a n d the
suggestive lacks second and third person forms which leads to the uniting of the 258
196 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
two moods, so that each one supplies what the other lacks; and the two moods are
very close semantically. If some people try to claim that there are second person
suggestives, a n d that they just h a p p e n to coincide in form with imperatives, we
could reply that the construction of a form with the particle age will establish the
imperative mood, since this hortatory intensifies (epiteinei) the imperative notion.
[[It is not obvious what A.D. is proposing. If he is saying that the 'suggestive' a n d
the imperative are really distinct moods, a n d age infallibly marks the imperative,
he is forgetting the fact that age is quite c o m m o n with first person singular a n d
plural forms, as appears in some of his H o m e r quotations above—Chs. 107, ex.
3.60, 3.61 (singular), 106, ex. 3.59 a n d 109, ex. 3.64 (plural). If, as B u t t m a n
suggests, he is p r o p o s i n g that t h e r e is really only o n e m o o d , a combined
suggestive-imperative, h e seems to be g o i n g c o n t r a r y to his implications
elsewhere.]]
112. (XXVII) Next we must discuss the third person imperatives, which are also
simultaneously second person, they say, but differ from the p u r e second person
forms in that the latter imply a direct acceptance [by the second person] of a
c o m m a n d to do something himself, whereas in the former, the second person is
c o m m a n d e d to pass on a c o m m a n d to other persons. [[This analysis, in short,
treats third person imperatives as causatives,—as the English translations also
suggest—with the second person as the causer a n d the third person as direct actor.
Sometimes the identity of the second person may be in doubt, as when the Creator
says "Let there be light!"]] "For", they say, "legeto ("let him speak") a n d the like
clearly involve two commands, one from the first to the second person, [e.g. "let. . ."],
and the other going from him to the third person, as can be u n d e r s t o o d by
comparing one-person [i.e. ordinary personal] p r o n o u n s and two-person ones
[i.e. possessives]. For ego ("I") a n d its class [namely "thou" and "he, she, it"] are
singly singular and indicate a single person, whereas a form such as emos ("my"
nom. sg. masc.) is recognized as doubly singular a n d refers to two persons ["I" a n d
the one possessed]. Likewise, in a sense, the accented [emphatic] personal pro
n o u n s [e.g. erne "me" etc.], t h o u g h simple, presuppose a second person because of 259
the contrastive force, (diastole), a n d similarly comparatives "A is m o r e so than B "
a n d relative nouns (ta pros ti) [e.g. father presupposes a child, half presupposes a
whole, etc.]. Just so legeto presupposes a second and a third person."
113. In reply to this it is possible to argue that while it is quite true that [forms like
legeto] involve a double person, and that the second person is involved, it is not true
that this is something exceptional about imperatives only, since the same thing
h a p p e n s just as much in the indicative mood (enclisis). For [when we make an
assertion] we are talking to somebody about something—orthos dialegetai Truphon
"hemera estin" ("Tryphon is right in r e m a r k i n g 'it is day'")—but we don't say, just
because dialegetai ("he remarks") is addressed to someone, that it is second person.
T h e people who defined the second person as "the one to whom the sentence [is
addressed]" [[this is, perhaps, the famous Ars attributed to Dionysius T h r a x , to
which A.D. often seems to allude in this manner]] are at fault for not a d d i n g "and
[the sentence is] about the addressee himself." [[I.e. the word "you" occurs only
when you're mentioning the person to w h o m you're talking—but he is always
presupposed.]] Certainly forms like legeto are orders addressed to a second per-
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 197
son, but they are not themselves second person forms, since the explicit c o m m a n d
is not to be carried out by the addressee, but by some third person or persons.
114. This fact is also obvious from the n u m b e r agreement. W h e n an o r d e r is
given to a second person or persons it is either singular— lege ("speak" addressed
to one)—or plural—legete ("speak" addressed to several). But this doesn't h a p p e n
in the case of legetō. A c o m m a n d to be carried out by a third person makes its
agreement of n u m b e r in the third person, legetō ("let him speak") legetdsan ("let
them speak"), whereas the implicit (enkeimenos) second person does not accept any 260
such agreement. But if the form really were classifiable as second person, it would
accept n u m b e r [agreement] as singular or plural, just as the [possessive] p r o n o u n s
do: hemeteros ("our(s)" n o m . sg. m a s c , [where the stem shows plural possessor, and
the ending shows singular possessed]), emos ("my/mine" [where the stem shows
singular possessor, and the ending singular possessed]). But, in fact, legetd, legetd
san and the rest [i.e. all third person imperatives] do not indicate any distinction of
n u m b e r for the second person, whether one is talking to a single person about a
single [third] person or to several about one, or [to one or to several] about several.
So it is possible to say to a g r o u p of people egregoreitd ho strategos humdn ("Let your
general wake up!") and also to a single person egregoreitd sou ho despotes ("Let your
master wake up!"), a n d there is no change except for that indicated for the third
person, which is the actual verb form. [i.e. the verb would be plural if the o r d e r
were "Let the officers wake up!"].
115. Forms like legetd, then, convey a c o m m a n d to be carried out by persons who
are not present, and necessarily a second person must take his t u r n as a relayer of
the command, since, as we have said, a c o m m a n d without a second person is
impossible. That's why [a third person imperative] may be accompanied by voca
tives, which are always semantically (nooumenos) second person. [[Again, any
sentence, in any mood a n d any person may be accompanied by a vocative; A.D.
usually forgets this.]]
116. We mentioned above (Ch. 103) that many [second person dual and plural
imperative] forms are h o m o p h o n o u s with indicative forms, e.g. legete ("Speak!" or
"You are speaking"), legesthe ("Choose!" or "You are choosing.") and dianoeisthe
("Consider!" or "You are considering"), etc. [[This includes all imperfective 2nd
dual or plural imperatives, and also perfect middle 2nd du./pl. imperatives.]] A n d
since vocative plurals are h o m o p h o n o u s with nominatives, and indicatives go with
nominatives while vocatives go with imperatives, then dianoeisthe anthrōpoi is two-
ways ambiguous [[actually it's four-ways ambiguous, because in spite of A.D.,
vocatives may freely occur with indicatives a n d nominatives occur with im
peratives]], a n d the ambiguity cannot be resolved until something is a d d e d to
disambiguate one or the other of the words. For we can say with the nominative 261
anthrdpoi ontes dianoeisthe ("You, being men, are thinking"—but, in fact, it may also
mean "Since you are men, think!"), not, as some maintain [and quite correctly, in
spite of A.D.] in the imperative mood, but only indicative, as is easy to recognize if
you shift to first or third person—anthrdpoi ontes dianooumetha ("Being men, we
think"), anthrdpoi ontes dianoountai ("Being men, they think"). [[However, one can
also say anthrdpoi ontes dianoeisthdn "Since they are men, let t h e m think!"]] It's also
clear as day from the singular anthrdpos on dianoei ("Being a man, you are think-
198 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
117. If we add the definite article, saying hoi anthrdpoi dianoeisthe ("You m e n
think") then dianoeisthe cannot be a c o m m a n d , but is an assertion (horismos) of the
existence of thought in each one, as if we were to say huparchei en humin to logistikon
("Reason exists in you-all"), because the article hoi signals the nominative. A n d if
we add 6 ("O!"), then the dianoeisthe becomes imperative, d anthrdpoi dianoeisthe
("Think, men!"—or else, "You are thinking, m e n " in spite of A.D.'s rule). But if we
add age [then it is indubitably imperative], age dianoeisthe anthrdpoi ("Come on and 262
think, men!"). O n the other hand, if we prefix hoti to this construction, hoti
dianoeisthe ("that you think" or "because you think") the construction will be
converted to an unquestionable indicative m o o d (enclisis). [[See above, Ch. 89, for
hoti as the special modal particle of the indicative.]] So much for the h o m o p h o n o u s
[forms of the imperative and vocative].
118. As far as the unambiguous forms of nouns and verbs are concerned, no
theoretical account is necessary for this construction: anthropos egraphen ("A m a n
was writing") anthrope graphe ("Write, man!"). A n d if one of the two [i.e. either the
n o u n alone or the verb alone] is ambiguous, the presence of the other eliminates
the ambiguity. Suppose we illustrate with Helikdn a n d similar names [which do not
distinguish nominative sg. from vocative sg.]; here the addition of an indicative
form reveals the n o u n as nominative: Helikdn graphei ("Helicon is painting"); of an
imperative shows the vocative: Helikdn graphe ("Paint, Helicon!").
119. Conversely in the case of echei ("it sounded" or "sound (it)") [[i.e. verbs—
mainly contract verbs—beginning with long vowels, so that the a u g m e n t makes no
change, in the active 2nd sg. imperative a n d 3rd sg. imperfect.]], namely in the
third person of the imperfect paradigm (prophora), e.g. echei pote to en Ddddnei
chalkeion ("Once the bronze cauldron at Dodona sounded"), and in the imperative,
as when I say echei su ("Sound it, you!"). W h e n it is accompanied by a [subject]
nominative form, echei is disambiguated as imperfect—echei ho anthropos ("The
man sounded [it]")—and, conversely, an associated vocative determines it as
imperative—echei anthrope ("Sound, man!"). W h e n an ambiguous nominative
form is put with such a verb form, the structure is completely ambiguous as in echei
Helikdn ("Helikon sounded (it)" or "Sound (it), Helikon!"). In this case we would
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 199
123. (XXVIII) Next, [having discussed the infinitive, the indicative, the optative
and the imperative], we must speak about the subjunctive (hupotaktike) mood,
which some people call the 'hypothetical' (or 'dubitative' distaktike) on the basis of
its meaning, just as the other moods are n a m e d [whereas 'subjunctive,' meaning
that it is a mood of subordinate clauses, is a syntactical name]. For clearly ean
grapho ("if 1 write...") and the like express a hypothesis (or "doubt") about an event
conceived as potentially future.
124. But perhaps someone may respond that it is not the mood itself which
signifies hypothesis (or "doubt"), but the accompanying conjunction (ean " i f ) is
responsible for the hypothetical ("dubitative") sense. A n d if it's O.K. to name
verb-forms after the force of conjunctions, then nothing prevents us from trans
ferring the same name to the other moods, when they accept this force from
conjunctions. For ei egrapsa ("If I h a d written. . ." contrary-to-fact), a n d ei
ephilologesa ("If I had studied. . .") [[by themselves egrapsa and ephilologesa mean "I
wrote" a n d "I studied", aorist indicative]] are no longer "indicative" (or "deter
minative" or "declarative"), nor is philologeso ("I will study") when it is in the
sentence etoi philologeso e poreusomai eisperipaton ("I'll either study or go for a walk")
[since it indicates a future uncertainty], nor does grapsaimi an ("I would write" or "I
might write") indicate a wish (or prayer), but an assertion of a potential future act.
T h e a m o u n t of "doubt" or "hypotheticalness" present in ean peripateis kinethesei ("If
you walk, you will move" ["whereperipateis is subjunctive]) is pretty well matched by
that in ei peripateis kinei ("If you're walking, you're moving", [where peripateis is 265
present indicative]), and yet we don't call ei peripateis "dubitative" [or "hypotheti
cal"].
125. T o this objection it is possible to reply that the other moods, being capable
of occurring without conjunction, have i n h e r e n t meanings (deloumenon) of their
own from which they have received their names. Obviously grapsaimi ("I hope I
write") by itself indicates wish (or prayer), a n d peripato ("I'm walking") by itself
conveys determinativity (declarativeness, indicativeness,homraas). So if one places
them in conjunctional constructions, they won't t h e n receive their names from
such constructions, but only from their natural inherent mood (enclisis), just as
these conjunctions themselves are given names based on their own inherent
semantics, some being called "disjunctive" [e.g. "or"] or "potential" [the particle
an, which is not now generally considered a conjunction], or "synaptic" [the
conjunction ei " i f used with the indicative or optative], and so on [[an implicit
reference to the Stoics and the tradition represented by Dionysius T h r a x , who
classified conjunctions in this way]]. So if it were the case that the so-called
hypothetical ("dubitative"—i.e. the subjunctive) had a special significance when
used without any conjunction, it would certainly have taken its n a m e from that.
[[A.D. has already excluded the main use of the i n d e p e n d e n t subjunctive, the
first-person hortatory uses which he called "hupothetike" or "adhortative" ("sug
gestive") back in Ch. 109, a n d he has never mentioned the prohibitive uses of the
aorist subjunctive With me,—though he cites an instance of it at the e n d of Chapter
90—which he would probably treat as transformed imperatives.]] But as it is, the
subjunctive never occurs alone, without a conjunction, a n d so its i n h e r e n t mean
ing is not clear, and therefore, since it has no meaning of its own, it has taken its
n a m e from the force of the accompanying conjunction.
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 201
or Odyssey 23.13:
3.69 kai te chaliphroneonta
("And light-minded. . ." here both kai and te mean "and")
And we could add thousands more. It isn't even peculiar to conjunctions, this
pleonastic insertion; nearly every part-of-speech (lexis) can be treated like this.
129. If the n a m e ["expletive"] comes from their pleonastic use, what is the whim
of fate which assigned this name to these conjunctions, the expletives? Perhaps it is
the fact that the other conjunctions, along with their differences of form, show
subclasses with single meanings from which they can receive their names, whereas
this is not the case for the expletives. For practically every single one of t h e m has a
special sense: restriction (meiosis) in the case of ge as in:
3.66 touto ge moi charisai
("Grant me this, at least")
logical transition for de, adversativeness (enantiotes) for per along with emphatic 268
strengthening (auxesis). So it was impossible for them to take a n a m e from their
shared meaning, for the aforesaid reason. But the one thing which they did share
was their pleonastic use in passages where nothing of the sort was needed, a n d
from this shared characteristic they took the name, which they do not in fact belie.
[[I.e. they all are, if only sometimes, used expletively, i.e. unnecessarily.]]
130. A similar observation applies to the classes of nominal (paronumon) a n d
verbal derivatives. [[I.e. they can only be referred to by a name based on their
formation rather than their meaning.]] For they, too, took their names from their
underlying substance. They couldn't be n a m e d from their meaning because of the
vast variety of senses that they have, although, again, the other classes of second
ary nouns [e.g. comparatives, patronymics, possessives, etc.] have single meanings
from which they could take their names.
131. T h e mood we're discussing [i.e. the subjunctive], when used with ean and its
equivalents [[he ought to be thinking of words like hotan and epeidan "when", but
probably he just means eanper "if indeed" a n d similar expressions]] is accom
panied by a [main verb in the] future or present tense: ean philologo paragenesetai
Dion ("If I study, Dion will come"), ean anaginosko paraginetai Truphon ("If I [ever]
read, T r y p h o n comes"). A past tense is ungrammatical (akatallelos). Similarly with
final hina—hina philologeso paragenesetai Truphon ("Tryphon will come so that I may
study"), or, with paraginetai ("Tryphon comes so that I will study"). If it occurs in a
past construction, it may be understood as causal [see above, Ch. 126], hina
philologeso paragenethe Truphon ("Tryphon came for my studying"). [[A.D. never
mentions the occurrence of optatives in secondary sequence, except for a passing
reference in "On Conjunctions" 235 = 502 for an Attic use in hoti clauses.]] This is
equivalent to dioti ephilologesa paregenethe Truphon ("Tryphon came because I 269
studied"). Now I don't mean that it may not also be taken as a purpose (apotelestikos)
clause; it can have the sense of eis to philologesai me paregenethe Truphon ("Tryphon
came in o r d e r for me to study"). But the construction cannot have the causal
interpretation in a future context; causes are attributed only in the past. That's
why the causal interpretation is more grammatical in a past context: if I start with
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 203
hina hubriso Theona in the causal sense, ("because I injured Theon") I won't finish
aganaktesei Dion ("Dion will be angry") but iganaktisen ("was angry"). But in the
final sense, it's all right to say hina hubriso Theona parestai Truphon ("Tryphon will
come so that I may insult Theon"). [[The "because" use of hina plus subjunctive is
u n k n o w n outside of A.D.; see above Ch. 126.]]
132. (XXIX) We must next explain about the derivation of subjunctive con
structions, the question of the underlying mood from which the introductory
(paratithemenos) conjunctions transform the verbs. [[Apparently A.D. assumes an
underlying non-subjunctive for every surface subjunctive, a n d a "triggering"
function for the conjunctions. ]] We already said just above that subjunctive verbs
never occur without conjunctions. It is clear that [otherwise] every word (mews
logon) which occurs in construction with a conjunction can also invariably occur as
an i n d e p e n d e n t word, by itself. Expressions like ean dramiis ("if you run") or ean
grapseis ("if you write") are not single words (mews logon i.e. compounds); and it
must be granted that not every c o m p o u n d element can occur by itself, since we can
find thousands of forms (schema) occurring in composition which have been
transformed (metapoieitai) by the c o m p o u n d i n g , so that, when taken apart, the
parts cannot be said separately. So from sebomai ("I honor") comes eusebis ("well-
honored") and yet one cannot say *sebis alone. Similarly with protopagis ("newly-
joined") homepage ("it was joined"), and *page~s cannot occur alone; similarly with
agenes ("low-born, ignoble") from geino (actually geinomai "I was born"). [[Most
such adjectives in -is, however, correspond to neuter nouns in -os; here sebos
"honor" and genos "race, family" exist, a n d probably there once was a n e u t e r n o u n
*pagos "joinery". T h e stem-forms of these n o u n s e n d in -es-, just like the adjectives,
a n d their genitive and dative case forms agree.]] But ean labiis, as we said, is not 270
one word in that respect, so there's no absolute necessity for labēis to be capable of
occurring alone. [[The logic is a bit weak here.]]
133. But if this needs further proof, we will add a little more to show that
conjunctions are not c o m p o u n d e d with verbs into single words (mews logon). [If
they were,] how could other words intervene? E.g. ean semeron kai anrion akouseis,
antilepsei ton legomenon ("If you listen today and tomorrow, you will u n d e r s t a n d
what is being said.") But, [you may say,] this also h a p p e n s to c o m p o u n d s , as in Iliad
17.542 leon kata tauwn ededos ("a lion which has devoured a bull" [kata is in classical
Greek a prefix to ededos, katededos; this separation is called tmesis by the g r a m m a r
ians; see below, quotation 4.11] or in Odyssey 1.8-9:
3.70 nepioi hoi kata bous Huperionos Eelioio / esthion
("Fools, who ate u p the cattle of Hyperion the Sun")
for katisthion); but in the analogous cases with conjunction, like hina dramiis ("in
o r d e r that you may run") or ophra pepoitheis ("that you may believe", [e.g. in Iliad
1.524]), both words have i n d e p e n d e n t accents, showing that they form phrases
(parathesis), not compounds. A n d anyway, conjunctions are never c o m p o u n d e d
with any word, except for the synaptic conjunctions [e.g. eiper "if indeed", or the
"parasynaptic" ones like epeidē "since"].
134. T h e basic form (ekphora) underlying the subjunctive is the indicative, and
the conjunctions which govern it change it into its characteristic [subjunctive]
204 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
phonological form (phone). This is the origin of the name, since, if the form elabes
("you took") had remained unchanged in the phrase ean labels ("if you take"), it
would have kept the same modal name ["indicative"], even though it no longer
signifies an assertion, just as we have shown above that egrapsa an ("I would have
written"), even though it does not convey an assertion of the act, is still called
"indicative" because of the inherent association of the form with the mood. So
along with the change of form mean labels ("if you take"), ean dramels ("if you run"), 271
there is also a change of the name of the mood.
135. Subjunctives are derived from indicatives by lengthening the short vowels
[before the endings], leaving the rest of the syllables u n c h a n g e d . (a) T h e [first
person middle] indicative ends (legel) in -mai, legomal, but so does the [subjunctive]
legomai—and what kinship is there with the optative legolmen? or the impossible
imperative [[remember that the alleged first-person imperative forms that he
rejected back in chapters 104-111 are, in fact, subjunctives]]? or the infinitive,
legesthal? (b) In the second [person singular middle, subjunctive and indicative] are
identical, hoti legel ("that you are called . . .") and ean legel ("if you are called. . ."),
whereas legolo (optative) a n d legou (imperative) have no phonetic similarity
(olkelotes) to the subjunctive. (c) In the third person plural [active, the optative and
imperative] are legoien [optative] and legetosan[imperative], whereas the indicative
is legousln, with which legosln agrees, as in ean legosln ("if they say"), (d) In the
second conjugation of contract verbs [the so-called alpha-contract verbs like timao]
the present indicative and subjunctive are identical t h r o u g h o u t [both active and
middle], as they are for all verbs in the first person [singular] active, ean lego ("if I
say" subjunctive), hotl lego ("that I say" indicative), a n d in all second person
[singular] passive forms, hotl legel su ("that you are called . . .") ean legel su ("if you
are called. . .") hotl noei su ("that you are thought. . .") ean noel su ("if you are
thought. . ."). I think it would be a waste of time to extend the a r g u m e n t further;
the case is obvious. [[Of course, he doesn't explain how to derive first aorist
subjunctives, which have no corresponding "present aorist indicative", or perfect
subjunctives, where the endings of the indicative are quite different. But he's got
the right starting point.]]
136. Anyhow, we have already proved [Ch. 62] that the indicative is the primary
or basic mood from which the others are derived [i.e. performs the same function
for verbs that the nominative does for nouns], on the g r o u n d that it is the most
vivid, and has the most morphologically distinct tense forms. T h e r e f o r e , if the
optative and the imperative take their standard from the indicative, it's r e d u n d a n t
to inquire if the subjunctive also is transformed (metalambanetai) from it. [[This 272
chapter at least hints at the matrix form of verb paradigm discussed by Varro in
Bk. 10 of De lingua latlna. In one dimension the first person singular is the base for
all the other seven persons (2s., 3s., 2d., 3d., lp. 2p., 3p.,—there is no first dual,
except for three middle forms in Homer), in the other the indicative is the base for
the other moods, with subjunctive next.]]
137. (XXX) We must examine the construction of these conjunctions which take
the subjunctive, to determine why it is that they reject verb-forms with past-tense
endings (phone). For it is not acceptable to say *ean elegon [for "if I said", which
would have to be el elegon] or *ean pepoltha [for "if I trust", ei pepoltha] and the like,
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 205
138. Evidently the reason for this ungrammaticality (akatallelia) is the semantic
conflict between the past tenses and the force (dunamis) of these conjunctions [ean
and hina, mainly]. For they present uncertainty (distagmos) about future events,
including those which are going to be completed (whence the n a m e apotelestikoi
[meaning something like "completive" or "conclusive"; for us "final conjunctions"
or "conjunctions of purpose"]). For how can a past fact harmonize with what is
going to be? So combinations like *ean elabon (for "if (in the future) I took") or
*hina anegndn("in o r d e r that I had read") are ungrammatical (asustaton), a n d
similarly with other conjunctions of these classes, but it is grammatical (sustaton) to
say hina anagno ("so that I may read") or ean anagno ("if I read"). For this verb form
[anagnd] has an ending which can never signify past time in the first person. [[The
ending -6 is a primary ending, as m o d e r n grammarians say, found only in present
and future indicative and all subjunctives.]] T h o s e verbs whose second a n d third
person [subjunctive] forms must be the same in length [as the corresponding
indicatives] will either take the same vowel, di, or the equally long ēi [ both with the
so-called iota subscript, silent in A.D.'s time], so that even if the ō of the first person
remains in the second a n d third persons, it takes this a d d e d i; we discuss this in 273
more detail in our book "On Spelling." [[Here A.D. is talking, in a r a t h e r confusing
way, about athematic aorist contracted subjunctives, which either go -ō, -eis, -ex like
ordinary subjunctives, or, in the case of the aorists of gignosko "know" a n d its
compounds and of diddmi "give" a n d its, go -o, -dis, -di.]]
139. Evidently the causal conjunction hina [see above, Ch. 126, Ch. 131] takes the
same [subjunctive] construction as the purpose hina because of its identity of form,
perhaps influenced also by the h o m o p h o n y of the [relative] adverb hina ("where").
W h e n indicatives are construed with hina, it regularly has the locative sense, as in
Odyssey 10.417:
3.71 hina t'etraphen ed egenonto
("where they were b o r n and b r o u g h t up").
For it is well known that, otherwise, causal conjunctions are construed with past
tenses, e.g. hoti egrapsa ("because I wrote") or hoti enoesa ("because I thought").
140. But we must note that the subjunctives which occur with hina and ean are
derived both from presents a n d from pasts. [[Since A.D. considers the aorist a n d
perfect indicatives as past tenses, the aorist and perfect subjunctives must be
derived from past tenses, for him. Nowadays "tense" is attributed only to indic
atives, and these forms are said to differ in "aspect", not "tense".]] W h e n they
come from pasts [i.e., here, are of aorist aspect], they are used as in edn mathō ("if I
206 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
learn") which means ei anusaimi to mathein ("If I finish the act of learning"), or ean
dramo ("If I r u n . . ."), i.e. ei anusaimi to dramein ("If I finish the act of running");
whereas ean trecho ("If I r u n " or "If I am running") means ean en paratasei genomai
tou trechein ("If I get into the continued process of r u n n i n g . . ."). [[This account of
the meaning of the two aspects may be c o m p a r e d with those above in the segments
on optatives [Ch. 100] a n d imperatives [101-102]. This is essentially the m o d e r n
view. But A.D. is h a m p e r e d by the lack of any term for aspect, a n d by his feeling
that all subjunctives (and optatives) are derived from underlying indicatives.]]
This is why the construction with a future subjunctive doesn't exist. For the
conjunctions themselves [ean and hina] signify the potential future, whether
progressive (paratasis) or perfective (anusis).
141. T h e people who specify that the Dorians don't form contracted future
subjunctives, then, and investigate the question why they don't are plain silly.
[[Most Attic-Ionic futures in-so" -seis, etc. are matched in Doric by e-contract forms
(from -seo, seeis) -so, seis.So if the future did have a subjunctive, you would expect
the Doric forms to be -so, -seis (from -seo, -seeis). But, as A.D. proceeds to show,
there is no future subjunctive.]] T h e i r a r g u m e n t is based on an untenable as
sumption. What lured t h e m into this stupidity was the h o m o p h o n y [of such a
future subjunctive, if there were one] with [the subjunctive of] the a-final [i.e.
"first"] aorist, which is like this. T h e m o r a (chronos) that is prefixed [i.e. the syllabic 274
or temporal augment] in the [aorist] indicative, is d r o p p e d when the mood is
changed from the indicative. T h e optative based on elexa ("I said") is lexaimi ("May
I say!") the infinitive is lexai ("To say") a n d the imperative lexon ("Say!"). So, in the
subjunctive mood, the same thing h a p p e n s when the e n d i n g changes to -ō as
happens in the case of the other past 'tenses' [i.e. aorists a n d perfects]—ephagon ("I
ate" aorist), ean phago ("If I eat..."), edramon ("I ran" aorist), ean dramo ("If I r u n . . . " ) ,
and similarly elexa ("I said") ean lexo ("If I say..."), which is h o m o p h o n o u s with the
future indicative lexo ("I will say"). But that [the subjunctive lexo, lexeis] is not the
subjunctive of the h o m o p h o n o u s future [but of the aorist] should be obvious from
the preceding part of the Syntax. A n d if those who listen to the complete a r g u m e n t
refuse to share this opinion, they're behaving extremely stubbornly.
142. T o make the demonstration clear to everyone, I will present the mor
phological arguments for the position taken above [that lexo in ean lexo is aorist
subjunctive, not future subjunctive], first offering the point from which the
investigation began, that the Dorians preserve -contract [["circumflexed"]] con
jugation t h r o u g h o u t all forms of the future system, but do not show it in
subjunctives—because, obviously, there is no future subjunctive.
143. T h e second bit of evidence comes from the Attic conjugation [of verbs in
-izo]. In the future [of such verbs], Attic speakers say [e.g.] lurid ("I will play the
lyre") and komio ("I will bring") [both these forms are also r-contract forms, from
lurieo and komieo], but in the aorist elurisa ("I played the lyre") and ekomisa ("I
brought"). So how does it happen that they do not say *ean lurid (for "If I play lyre.. .")and
*ean komio (for "If I bring. . ."), but rather use -5- and recessive accent, agreeing
with the aorist [i.e. ean luriso and komiso]?
144. In the third place, in the liquid-and-nasal "("fifth") conjugation, the penul
timate syllables of first aorists are opposite in quantity to the penultimate syllables
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 207
of futures: the aorist always has a long penult, so that if it's derived from a short 275
present stem, it is lengthened—dero ("I skin") edeira ("I skinned"), nemo ("I man
age") eneima ("I managed"), but the future has a short vowel, so that if it is derived
from a long present stem, it is shortened—keiro ("I am cutting") kerb ("I will cut"),
phtheiro ("I destroy"),phthero ("I will destroy"). Given those facts, how can we deny
that ean neim ("If I manage") and ean deiro ("If I skin") come from the aorist alone,
since they contain the one feature impossible for futures, a long stem-vowel, and
have captured that one special mark of the first aorist, the lenthened penult?
145. In the fourth place, those who ask why the Dorians do not circumflex their
future subjunctives [i.e. conjugate them like e-contract present subjunctives] are
behaving a bit ridiculously; evidently they are forgetting some ordinary [Attic or
koine] forms when they ask why Doric future [subjunctives] are not circumflexed,
since some common forms behave in exactly the same way. T h e y could just as
easily have asked in this connection why pluno ("I will wash") is circumflexed, but
not ean pluno("If I wash. . ."). Obviously it is because the circumflex is a [necessary
and sufficient] mark of the future here, which proves our theory correct. If in
these verbs the future had the same accentuation as the aorist [subjunctive], the
structure would be ambiguous, as is the case with steso ("I will set u p " future), and,
from estesa ("I set up") ean steso ("If I set up").
146. Fifth is the fact that doso ("I will give") a n d theso ("I will put. . .") if combined
with the 'adjunctive' (epizeuktikos i.e. subjunctive-governing, like ean and hind)
conjunctions produce an ungrammatical (akatallelos) construction. But how does it
h a p p e n that the adjunctives will not go with a form in -5, when they have the power 276
to transform other endings into-o? [[i.e., xihina or ean is placed before a past tense
or an optative, the ungrammaticality can be corrected by changing the verb to the
corresponding subjunctive.]] No doubt it is because *edosa and *ethesa with -s- do
not exist, from which by changing the -a to -5, one could derive a grammatical
form. T h e actual [first] aorists [edoka "I gave" and etheka "I put"] are defective
(oligoreitai), and are lacking outside the indicative mood [indeed they are almost
limited to the three persons of the singular], and so, since there are no p r o p e r
[first] aorists, there are no derived modal forms. T h e r e are no "second futures" of
mi-verbs [such as didomi "give" and tithemi "put"], but there are second aorists like
*edon, *apedon, *ethen, [[In fact, these three forms are non-existent, but must be
postulated by A.D. as underlying or base forms to derive the plurals edomen "we
gave", apedomen "we paid", ethemen "we put" and the modal forms.]] from which
comes ethemen ("I put for m y s e l f ) . A n d since these second aorists exist, there are
also subjunctives derived from them—eantho ("If I put...") ean do ("If I give..."). It
has been shown that where a given stem exists for futures, but not aorists, there a
subjunctive formation is ungrammatical, but where there are aorists but not
futures, there the subjunctive is grammatical.
147. (XXXI) [Having finished with the category of mood], we must now discuss
the category of voice which is present in every mood, not even excluding the
infinitives, because of the logical necessity for all tenses to be marked as either
active or passive or middle. First we must investigate the question whether all verbs
participate in the two-voice system (along with the aforesaid middle), just as in the
case of nouns we consider the category of gender, whether all nouns take part in
208 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
the inflections for feminine and neuter; [[of course those that do are the adjec
tives, considered by A.D. as a subclass of nouns. Note that masculine is assumed as
the base form.]] or is it that some verbs are inflected only for the moods: indicative, 277
optative, etc.—but do not vary for active or passive, [[he seems to be thinking of
deponents]] or that some verbs, along with indicative and the other moods indi
cate activity, but are not inflected for passivity? [[Here he seems to mean active
intransitives.]] T h e theory goes as follows.
148. J u s t because a verb occurs in the indicative (or any other mood) it does not
necessarily follow that it will have a true active voice. For you must consider that
activity is something that passes over (diabibazetai) to some object (hupokeimenon,
"entity"), as in [transitive] verbs like [derei "he skins"], temnei ("he cuts"), tuptei ("he
beats"), and similar verbs, and from these basic actives is derived the passive voice
deretai ("he is skinned"), [temnetai "he gets cut"], tuptetai ("he is being beaten"). But
there are quite different verbs [which are not transitive, have no passive, and really
no "activity"] such as huparchb ("I exist"), zb ("I live"), eimi ("I am"), pneb ("I'm
breathing"),phrono, ("I'm sensible") etc. [[The fact that some such verbs in Latin,
e.g. vivo ("I live"), do have passives, e.g. vivitur "there is life", "people live", led to
some fascinating debates in the Renaissance a n d the Age of Reason, when Sanctius
maintained that all a p p a r e n t intransitives have a deleted object—vivo (vitam) "I live
(life)"—being followed in this, as in nearly everything, by Lancelot a n d Arnauld.]]
150. Some verbs signify a mental or physical disposition, which cannot take the
passive inflection because the passivity is already present even with the active
endings (katalexis). While the verb anib ("I annoy") involves action passing over to
another person, so that it will allow a passive anibmai ("I am annoyed"), this is not
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 209
the case with verbs like kopio ("I'm getting tired") or opthalmio ("I've got eye
trouble"); for such verbs involve self-suffering (autopatheia) [i.e. nothing "passes
over" to a n o t h e r person]. A n d since this passive experience may belong either to
the class of desirables (euktaion) or those which are not so, it's generally agreed that
there are no passives for pascho ("I experience (something)", "I have (something)
done to me, h a p p e n to me"; [this may have an agent construction with hupo, and an
accusative of the particular thing which happens]), chairo ("I rejoice"; [a few
passive forms occur but the meaning is the same]), eruthrio ("I blush"), thneisko ("I
die"; [this, like pascho, may even take hupo + gen. "by so-and-so" for the agent of
death, like a true passive]), gerō ("I grow old"; [this has a causative first aorist,
egerasa "I caused to grow old"]), thallo ("I flourish") ouretio ("I have to urinate"),
gaurio ("I exult"). T r y i n g to passivize these would be like a d d i n g a masculine suffix
to masculine nouns or a feminine suffix to feminine nouns. You cannot make
something passive if it already is passive.
151. Verbs with "middle" present form, formally like the passive, but signifying
a [transitive] activity, are incapable of taking the -5 endings of the active because
the possibility of using it is destroyed by the aforesaid present "middle", as for 279
instance, biazomai se ("I force you"), machomaisoi ("I'm fighting you"), chromai soi ("I
use you") and many, many more. [[But biazomai is used in the aorist and perfect
passive—occasionally even the h o m o p h o n o u s present passive—e.g. ebiasthen ("I
was forced.")]] It is clear that every passive form in -max etc. has a corresponding
active, provided that the syntax-and-semantics (ta tes suntaxeos) agree [in passivity]
with the form: histamai hupo sou ("I am stood u p by you") histemi se ("I stand you
up"), deromai hupo sou ("I am skinned by you"), dero se ("I skin you"), helkomai hupo
sou ("I am drawn by you"), helkō se ("I draw you"), but not *petamai hupo sou ("*I fly
by you"), hence not *petemi se (? "I cause you to fly" or whatever); the same holds
good for agamai ("I admire..."),dunamai ("I can...", "I am able..."),eramai ("I'm in
love with. . .").
152. Some other verbs signify an activity, yet have no corresponding [complete]
passive paradigm (ekphora) [i.e., not just third person forms] because the inani
mate objects (apsucha) affected by these verbs cannot be considered to experience
or feel anything, unless someone makes u p a speech as if spoken by them; so
peripato ("I walk"). You cannot make a *peripatoumai ("I am being walked") or a
*peripatei ("You are walked") since we do not address speech to inanimates [for the
second person] and inanimates cannot make assertions [to use the first person
form], but we do talk about them, and can say peripateitai he hodos ("The road is
being walked"), oikeitai he ge ("The land is inhabited"). T h e same account works for
pled ("I sail"), trecho ("I run") and all of that sort [[mainly verbs of motion, taking
objects like "trip", "journey", "race", "sea", etc.]]. Verbs like dorizo, aiolizo, etc. ("I
speak Doric", "I speak Aeolic", etc. with -izo for "I speak X language or dialect")
are also semantically active, but cannot make first or second person passives,
*dorizomai ("I am spoken in Doric"), *aiolizomai ("I am spoken in Aeolic"), but [only
third persons]: aiolizetai ta Alkaiou poiemata, dorizetai ta Alkmanos ("Alcaeus' poems
are in Aeolic, Alcman's in Doric"). [[This is almost the only mention of verbs which
take inanimate objects; A.D. never classifies verbs like take, paint, give, sell, bring,
know, eat, raise, etc.]]
210 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
153. T h e same observation applies to aristo ("I eat breakfast") and deipnō ("I eat
dinner"). T h e case is so obvious that it is not worth the time to give examples.
Obviously deipnizo ("I entertain.. .at dinner") and aristizo ("I have.. .to breakfast"), 280
since they take (epidechetai) an oblique case—accusative—as complement, referr
ing to an animate being (empsuchon), have complete passive paradigms aristizomai
("I am entertained at breakfast") and deipnizomai ("I am dined"). For [aristizo and
deipnizo] are equivalent to ariston parecho tini ("I give someone breakfast") and
deipnon [parecho tini] ("I give someone dinner"), so that aristo differs from aristizo in
that aristo means aristou metalambano ("I partake of breakfast") whereas aristizo
means aristou metadidomi tini ("I share breakfast with someone"). Therefore, as we
said, the possibility of this construction allows the formation of the [complete]
passive. T h e r e is a similar difference between gamō ("I marry [a woman]") and
gamizo ("I give [someone] in marriage"). For the former means gamou metalambano
("I partake of marriage") a n d gamizo means gamou tini metadidomi("I give a share of
marriage to someone"). Butgamo can be inflected t h r o u g h o u t the passive, since its
complement [in the active] refers to an animate person. [[In short, -izo derives
causatives from both transitive and intransitive verbs.]]
154. Now that this is clear, we must consider the verb eripo, to determine whether
it is equivalent to pipto ("fall") [and hence an intransitive, unpassivizable verb]
whose participle, accented according to the dialect, is eripon. [[This accentuation in
Attic is the mark of a second aorist participle.]] And if peson ("having fallen",
second aorist participle) has no passive, a n d it is possible to say pesonti ("having
fallen", dative sg. masc.) then obviously the phrase in Pindar's Olymp. II, 48:
3.72 eriponti Poluneikei
("to Polyneices, having fallen")
is more correct with o [i.e. eriponti rather than eripenti, the other reading of the mss].
But if it were the case that eripo is a synonym of pipto, then there would not be any
passive eripetai, just as there is no piptetai. But perhaps it is synonymous with hallo
("I throw") instead, and like hallo se ("I throw at you") there is eripo se, and like
hlethenti ("to one who was thrown (at)") is eripenti. For there certainly is a form
ereripto in Iliad 14.15:
3.73 ereripto de teichos Achaion 281
("And the wall of the Achaeans had been thrown down")
which is not, as some suppose, derived from the verb ripto ("I throw"). For it is more
natural to say that a wall has been "overthrown" than that it has been "thrown". So
then [the form in Homer] is from eripo, in the passive voice third person of the
pluperfect tense, with Attic shortening of the second vowel from e to e, as in
xeron/xeron ("dry" Ionic a n d Attic) pothesailpothesai ("to long for"). From eripo comes
*eripto, then from that [by Attic reduplication] ereripto, a n d then [by shortening]
ereripto. [[Modern scholars give ereipo as the present. *eripto must be considered an
intermediate stage of derivation which never surfaces; it would be the hypotheti
cal "regular" pluperfect.]]
155. [To r e t u r n to the subject matter as of Ch. 153 or so], the above [intransitive]
verbs in construction with a nominative [subject] express a complete thought—
B O O K III: ON SOLECISM; VERBS; PARTICIPLES 211
157. We can make a m o r e general statement, that verbs which are construed
(sunartatai) only with a nominative [subject] a n d do not require an oblique case
[complement, object] will not have a passive form, either. But verbs which require
a complement can always be transformed from active to passive voice, a d d i n g a
genitive p h r a s e governed by the preposition hupo ("by"), e.g. deromai hupo
Truphbnos ("I am skinned by T r y p h o n " ) , timbmai hupo Theōnos ("I am h o n o r e d by
T h e o n " ) . This is the only construction restricted to passives; with actives you may
have the genitive without the preposition hupo, kurieuō sou ("I am your master"), or
the dative,palaibsoi ("I wrestle with you"), or the accusative, timō se ("I h o n o r you").
[[Of course, some verbs d o take prepositional complements, but not hupo +
212 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
genitive, unless they are semantically passive. Note that A.D. nowhere restricts the
notion of "transitive" or of passivization to verbs which take the accusative. In fact,
some verbs which govern other cases d o passivize in Greek. But see below, C h a p t e r
178.]]
158. (XXXII) T h e foregoing account is adequate for anyone who wishes merely
to u n d e r s t a n d the facts of usage a n d tradition (paradosis). But for anyone who
wants to investigate in detail the underlying theory of g r a m m a r (ta tes suntaxeos tou
logon), it will be necessary to inquire which verbs take (apaitei) the genitive a n d why,
which the dative, a n d again for what reason, a n d finally also the accusative.
Obviously it's going to be a big and difficult task to cover the large n u m b e r s of such
verbs [that exist in Greek] along with their syntactical characteristics (idiomata), but
I think [in what follows] that I will achieve the requisite degree of precision.
159. Almost all actions (drasis) which proceed from a nominative [subject] affect
an accusative [object], so that the sentence presents both an active agent a n d an
entity which receives the effect (pathos), as in dero se ("I skin you"), tupto se("I beat
you"); or else the recipients are transformed, in the passive construction, [from
accusative] to nominative, while the agents are changed to the genitive with hupo 284
("by") e.g. ego deromai hupo sou ("I am skinned by you"). Why the preposition is hupo
will be explained later [presumably in a lost part of Book 4]. This transitive
construction [whether active or passive], whether it signifies physical action or
mental, as we have said, is grammatically the same. A n d since there's a great
variety of possible [subject-object] relations (diatithesthai), there is also a great
variety of subtypes of verbs according to the characteristic kinds (idiotes) of actions.
160. T h e following are [transitive] verbs of physical relation {somatike diathesis):
tribo se ("I thrash you"), nipto se ("I wash you"), resso se ("I break you"), helkō se ("I
d r a g you"), biazomai ("I force. . . "), chalo ("I loosen. . . "), gumnazo ("I train. . . "),
nusso ("I prick. . . "), knetho ("I scratch. . . "), xuo ("I scrape. . . "), smō ("I
cleanse. . ."), brecho ("I soak. . . "), tupto ("I beat. . . "),paio ("I strike. . . "), loud ("I
wash. . . "), desmeuo ("I bind. . . "), luo ("I u n d o . . . "),plēsso ("I strike. . . "),phoneuo
("I m u r d e r . . . "), kteino ("I kill. . . "),phtheiro ("I destroy. . . "), kaio ("I burn. .'. "),
phlego ("I burn. . . "), kaihizo ("I seat. . . "), therizo ("I mow. . . "), zemio ("I damage. . ."),
blapto ("I harm. . . "). A n d both physical a n d mental is hubrizo ("I injure. . ."), for
that covers both use of hands and mental attitude (psuchiche diathesis), which is
involved in verbs like loidoro ("I curse. . . "), kakologo ("I Curse. . . "), anio ("I
distress. . . "), lupo ("I grieve [someone]"). A n o t h e r subclass [of mental verbs] is
those of praise: humno se ("I sing your praises"), megaluno se ("I magnify you"), dido
("I sing of. . "),melpo ("I sing of. . ."),doxazō ("I glorify. . ."),kleio ("I celebrate. . .")
—from which comes kleos ("fame") [[this note is a d d e d to indicate that he does
not m e a n the verb kleio "close. . ."]]—, aino ("I praise. . ."). A n d a n o t h e r subclass
is verbs of deception, such as paralogizomai se ("I cheat you"), klepto ("I steal, I
cheat"), apato ("I deceive"),perigelo ("I ridicule"),paizo ("I make fun of"),apapho ("I
cheat"), exapato ("I deceive"), piano ("I mislead"). A n d verbs related to separation
from things e.g. zeto se ("I look for you"), heurisko ("I find"). A n d verbs of control:
echo ("I have"), krato ("I own"), phulassō ("I guard"), tero ("I watch"),eirgo ("I shut. . .
up"). Some verbs of action, like erizo soi ("I quarrel with you") or palaio soi ("I
wrestle with you"), take the dative, for reasons we will explain later [Ch. 185]. But
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 213
niko ("I defeat") being a verb of control takes the accusative. Similarly the struc- 285
tures of a more honorific type; there are many verbs asserting things which may be
covered by the verb sebomai ("I honor")—entrepomai ("I respect"), aischunomai ("I
revere"), proskuno ("I make obeisance to"), thopeuo ("I flatter"), kolakeuo ("I fawn
on"), hazomai ("I revere"). And some verbs involve a mental relationship of
thought with some goal: oiomai se ("I think that you. . . ") [[this depends on A.D.'s
theory—above, Ch. 78-9—that the so-called subject of the infinitive is really the
object of the governing verb]], hupolambano ("I u n d e r s t a n d that. . . "), huponoo ("I
suspect"), huphoromai ("I suspect"). [[Note that this whole list consists almost
exclusively of verbs which regularly or optionally take objects which are h u m a n .
Cf. above, Ch. 153.]]
161. Verbs of will-and-desire (proairetika), too, involve an attitude directed at an
[entity in the] accusative case, any person other than the subject, e.g. from first
person to second or third, [second to first or third], or third person to second or
first [or third], or to put it more briefly, when there is transition from one person
(entity,prosopon) to a different one. For in boulomai philologein ("I want to study") or
proairoumai anaginoskein("I prefer to read"), there is no need of a contrastive
[accusative] p r o n o u n , since the construction is one involving self-suffering [i.e.
identity of subject, Equi-NP]. T h e underlying structure (logos), again, is boulomai
emauton ploutein ("I want myself to be rich"), boulomai emauton peripatein (*T want
myself to walk") i.e. "I am disposed toward walking" or "being rich". [[These forms
with reflexives are just as queer in Greek as in English, and seem clearly to be
conceived as "deep structures."]]
162. But if there is a change of person [i.e. if the infinitive has a different
subject], then the p r o n o u n must be used: boulomai se graphein ("I want you to
write"), boulomai se Dionusion philein("I want you to like Dionysius"); if no p r o n o u n
[or NP] is inserted, then the action of the infinitive must go back to the subject of
boulomai, and so be "self-suffering". [The subordinate infinitive whose subject is 286
the same as that of the governing verb of wishing] requires the addition of the
object p r o n o u n emauton ("myself) if the infinitive is from a transitive verb, but not
otherwise. T h e first case [i.e. where the infinitive is transitive] is as in gumnazo se ("I
exercise you"—i.e. gumnazo is transitive), boulomai gumnazein emauton ("I wish to
exercise myself), dero se ("I skin you"), boulomai derein emauton ("I want to skin
myself). An example of the second case is, e.g., plouto("I am rich"), which is
intransitive, so that it does not absolutely require boulomai ploutein emauton ("I want
myself to be rich"), but may also be used without the p r o n o u n [i.e. boulomai ploutein
"I want to be rich".] [[In fact, it is almost always without the "myself. O n the rare
occasions when a p r o n o u n is used, it is more likely to be the simple personal
p r o n o u n , as in English "I want you to do it, not me", or "I want you, not me to do it",
and, as in those sentences, is contrastive. I have not yet found an emauton used in
such sentences, as A.D. uses it.]] T h e same applies in the case of boulomai peripatein
("I want to walk") a n d its class, and so the upshot is that we invariably need the
reflexive p r o n o u n emauton in the first case [transitive infinitive, with no other
"person" involved], e.g. with boulomai blaptein ("I want to hurt. . . ") or boulomai
philein ("I want to like. . . "), but not with peripatein ("to walk"), trechein ("to run"),
and their class [i.e. intransitives]. [[A.D.'s lack of terms for "subject" a n d "object",
214 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
and his resolute efforts to get along with "nominative", "accusative", "person",
"same", "different" and "transitive", really cause him trouble in this chapter.]]
163. Perhaps someone may think that boulomai ("I want"), proairoumai ("I pre
fer"), thelo ("I wish"), etc. d o not themselves govern (pheretai epi) the accusative, a n d
that the accusative belongs rather to the d e p e n d e n t infinitive. Suppose we con
sider blapto se ("I h u r t you") a n d grapho se ("I'm painting you"), from which we
derive thelo se blaptein ("I want to h u r t you"), a n d thelō se graphein ("I want to paint
you"), so that the accusatives belong to the infinitives, not to the governing
indicatives. A n d besides we can a d d this clincher: consider the verb akouo ("I
hear") which governs the genitive. T h e n when we say thelo akouein Dionusiou ("I
want to hear Dionysius") with the genitive, the reason is that akouō takes the
genitive.
164. But that's just nonsense. As we have just said, verbs of will-and-desire 287
always take an accusative, a n d [the construction] acquires a n o t h e r [oblique] case if
the d e p e n d e n t infinitive is derived from a verb which governs (pheretai epi) an
oblique case. Otherwise the only accusative will be the one d e p e n d i n g on the verb
of will or desire, (a) A n example of the first case isthelo se akouein Dionusiou ("I want
you to h e a r Dionysius"), where the accusative se ("you") is governed by thelo ("I
want"), whereas the genitive Dionusiou d e p e n d s on akouein ("to hear"). A n o t h e r
example is the verb soi charizomai ("I favor you") [which takes the dative], from
which we may form thelo se charizesthai Dionusioi ("I want you to favor Dionysius"),
with the accusative here, too, being governed as we said [by thelo] a n d the depen
dent dative [Dionusioi] by charizesthai. A n d so also two accusatives a p p e a r if the verb
[in the infinitive] is one which regularly governs an accusative, as is the case with
blapto se ("I h a r m you"), whence we get thelo se blaptein Dionusion ("I want you to
h a r m Dionysius") and thelō Dionusion blaptein Apollonion ("I want Dionysius to h a r m
Apollonius"), about whose possible ambiguity we spoke [in Chapter 84-85]. (b)
Examples of the second case, intransitives, are plouto ("I am rich"), peripato ("I
walk"), zo ("I live"), a n d m a n y more. In a sentence like boulomai se zen ("I want you
to live"), there is no case [i.e. NP] deleted, because the verb is intransitive, a n d the
accusative [se "you"] which is present d e p e n d s on the verb of will. So we have now
explained how verbs of will and desire govern the accusative. [[This is the same
rejection of subject accusatives that we had above in Ch. 78.]]
172. T h e inconsistency of the construction [i.e. the genitive for other senses,
accusative for sight] has forced us into this digression. A n d it is clear that philein
("like, love"), and eran ("love, be in love with") differ in the same way, in that the
disposition present in philein shows that it is a word of action. Those who philein are
influencing [others], since the disposition which this verb shares with other [tran
sitive] verbs discussed above entails an accusative object. This verb is like didaskein
("teach") and peithein ("persuade"). But the verb eran entails counter-influence
from the beloved. That's why Sappho was correct in using the stronger word [in fr.
13 Bergk]:
3.85 ego de ken' ottotis eratai
[[This passage has not yet been successfully explained; possibly
"whatever one loves, that I. . . "]]
So it is clear that philein is something for good and wise m e n to do, as fathers love 292
their children, hut eran is for the foolish or perhaps rather the insane. So it would
no longer be a puzzle to anyone why philo takes the accusative and ero the genitive.
173. Likewise the verbs kedesthai ("care for"), pronoeisthai ("provide for") and
phrontizein ("be concerned for"), which also take the genitive, include along with
218 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
the activity of caring also the passivity of being affected (paschein) by other people,
so it is quite according to theory that they take the genitive.
174. T h e genitive is similarly governed by verbs indicating domination or rule in
relation to those who are surpassed or mastered, quite naturally according to o u r
theory. Obviously there can be no notion of possession (ktema) without a genitive.
That's why all possessive adjectives (ktetika) are reducible to and derived from
genitives, both in nouns a n d pronouns, in relation to all things which are capable
of being possessed. [See above Bk. 2, Ch. 103-115, 136-7]. This is why Nea Polls
("Naples" = "New City") and similar names composed of two words with separate
accents are declined in both parts [i.e. where one part is an adjective agreeing with
the other part], whereas names like Korakos Petra ("Crow's Rock"), huos kuamos
("pig's bean", the plant called "henbane"), Onou Gnathos ("Ass's Jaw"), Achaion
Limen ("Achaeans' Harbor") are not [since the first part is a genitive modifying the
second]. If the genitive were changed to another case, the whole idea of possession
or domination would disappear. That's what this class of verbs depends on: a
king's subjects are his possession, in a sense. That's why we say basileuo touton ("I
reign over these men"), and so with hegemoneuo ("I dominate"), stratego ("I com
m a n d as general"), turanno ("I am dictator over"). That's why the dative must be 293
considered ungrammatical [i.e. a poetic figure] in Iliad 1.180:
a n d both n o u n a n d verb take the same construction [i.e., like turanno humon "I rule
you", isturannos humon "ruler over you"]. T h e same thing is true of basileus ("king")
a n d basileuo ("I reign over"), strategos ("general") a n d stratego ("I command"),
despotes ("master") and despozo ("I am master of"),kurios ("lord") and kurieuo ("I am
lord of"). [[Recent grammarians have generally-regarded the genitive with verbs
of ruling and c o m m a n d i n g as a partitive genitive—the King of the Persians is one
of the Persians, the general of the Greeks is one of them—but a fair case can be
m a d e that A.D. is right in calling it a possessive. We d o say, in English at least, " T h e
Persians have a king", " T h e army has a general", " T h e city has a dictator", a n d we
d o not often say " T h e king has Persians" or " T h e dictator has a city." Still, Greek
echo is not identical with English "have", t h o u g h it, too, takes a direct object in the
accusative, in spite of A.D.'s notion of an essential genitive.]]
177. Now we must move on to the verbs that take the dative. All verbs which
indicate someone's profit or acquisition (peripoiesis), whether verbal or physical,
take the dative; for instance lego soi ("I tell you") (soi is dative) is equivalent to "I
share speech wiih you", and, contrariwise, it is clear that lego se klepten ("I call you a
t h i e f where se is accusative) signifies (semainei) "I assert by means of the speech
which I a m uttering that you have d o n e the act of stealing." [[This is amazingly like
the d e e p structures of Generative Semantics, e.g. McCawley, CLS IV (1968)
71-80.]] A n d necessarily the act of speaking relates to a dative entity [i.e. the
person addressed], just as in a physical act like temno soi ("I am cutting for you" [soi
dative]), which means "I a m benefiting you with some p a r t of something [i.e. a
piece of meat]." But temno se ("I'm cutting you" [se acc.]) directs the act against the
object (hupokeimenon) a n d so takes the accusative construction. [[This is the same
claim m a d e later, e.g., by Sanctius, that all true datives are datives of advantage or
disadvantage, "acquisitio" which is a good equivalent for A.D.'s peripoiesis.]]
178. I n reference to this construction [of verb with dative] it must be noted that
the subject of the passive form temnomai ("I am cut") can never be the person who is
referred to in the active by the dative, since the action is not directed at that person,
but at the implied (enkeimenos) accusative which alone can become subject of the
passive, e.g. temnomai ("I am cut"). [[In English, of course, we say "he was given a
piece of meat," but not in Greek. T h e r e are a few Greek verbs taking the dative of
which this is not true; they are listed in Smyth (1956) pars. 1745 a n d 1748, but do
not include verbs m e a n i n g give, show, or tell. ]] T h e same a r g u m e n t fits all verbs of 295
similar meaning: dido soi ("I sing to you") clearly does not have a passive didomai ("I
am sung"), which is the passive mate to aidō se ("I sing you"). Similarly with komoido
soi ("I j o k e with you" dat.) a n d komoido se ("I ridicule you" a c e ) , humno soi ("I sing to
you" dat.) vs. humno se ("I sing of you" acc.). So also kitharizo soi ("I play the lyre to
you"), tragoido soi ("I declaim to you"), anaginosko soi ("I read to you"), phaino soi ("I
disclose to you"), kirno soi ("I mix. . . for you"), stornuo soi ("I spread. . . for you"),
doroumai soi ("I present to you"), charizomai soi ("I do a favor for you"), aulo soi ("I
play the flute for you"), which is analyzed as signifying a benefiting (peripoiesis) [of
you] with flute-playing. [[It's curious that he says nothing h e r e about verbs which
require both dative and accusative. But see below, Ch. 183.]]
179. T h e dative has a different sense in aulo tois aulois ("I play the flute") [than it
has in aulo soi ("I play the flute for you")], in which aulein ("play-flute") conveys
220 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
both the meanings present separately in the former construction, both the act of
playing a n d the use of the flute, as also in aulo tois theatais ("I play the flute for the
audience"). But in the other construction [aulo tois aulois "I play the flute"], aulō
signifies only the musical skill and tois aulois "on the flute" indicates the particular
instrument; i.e., tois aulois aulein ("to play on the flute") is equivalent to "to exhibit
one's musical skill by means of the flute", as becomes clear if we change the
instrument [or the verb], givingsurizei tois aulois ("He pipes on the flute") or aulei tei
suringi ("He plays on the pipes"). [[In short, besides the dative of advantage used
for personal NP's, there is a dative of means or instrument, used with NP's
referring to things. A n d some verbs are derived from instrument nouns, so
contain an inherent instrument, which may be canceled or repeated by an explicit
instrumental phrase.]]
180. T h e same thing can be observed in the case of other verbs, since akouei ("He
hears") sometimes signifies perception by h e a r i n g (akoe), as of noises, buzzes, roars
and other non-linguistic sounds, with reference to which we call a m a n keen of
hearing. But sometimes the verb refers to the u n d e r s t a n d i n g (sunienai) of what is
heard. In Iliad 14.1:
3.87 Nestora d'ouk elathen iache
("Nestor could not help noticing the din.")
T h e m e a n i n g is not simply that he h e a r d the sound, but rather that on hearing and
u n d e r s t a n d i n g the sound he also understood what was going on in the fighting. 296
We also use the expression ouk akouei ho deina ton legomenon ("So-and-so does not
hear what he's told") m e a n i n g that he does perceive the words, b u t does not
u n d e r s t a n d their significance. T h e r e is also a third sense for akouei mou ("He hears
me"), when it signifies "he consents to what he's told by me", akouei mou ho deina
("So-and-so listens to me"), i.e. he is obedient to me, a n d conversely ouk akouei ("he
does not listen") in the sense "So-and-so is disobedient."
pupil says "I have not read yet". But sometimes it means both that a n d the mental
apprehension of the poems [which are being read]. T h u s we say ouk anegno ho deina 297
Alkaion ("So-and-so has not r e a d Alcaeus"—i.e. has not p r e p a r e d his homework,
presumably) or ouk anegno Homeron ("He did not r e a d Homer"). T h e same sort of
thing can be demonstrated for other verbs, too, but this is not the object of o u r
present discussion, which is to complete the account of the construction of oblique
cases [complements] with verbs, not the variations of semantics in the verbs.
A n d if [you say that] aulō tei noumeniai ("I play at the new moon") does not
agree [with the dative uses mentioned in 177-9], it is, nevertheless, clear that the
"benefit" or "acquisition" belongs to the audience [not to the new m o n t h ] . It is at
that particular time (katastema) that the action takes place, but the benefit goes to
the listeners. [[This passage scarcely fits h e r e , t h o u g h it might be squeezed in
somewhere in Ch. 179. All editors bracket it. It mentions still another use of the
dative, the indication of the time when an event occurs, provided that the head
n o u n is a n o u n of time. Like the instrumental use, it cannot be used of persons,
a n d the three uses are automatically distinguishable by their semantics.]]
183. Verbs which govern physical-objects (somata) [as direct objects], also behave
in this way, as we said above [177]: phero soi ("I bring you [something]"), ago soi ("I
bring you [something living]"), which contrasts, as we saw, with phero se ("I bring
you [to some place]") a n d ago se ("I take you [somewhere]"). So it is generally
understood that such datives include the accusative within themselves [i.e. entail
or imply direct objects, perhaps], which is then a d d e d at the end of the sentence:
ago soi ton paida ("I bring you the boy"),phero soi ton oinon ("I bring you the wine"),
temno soi to kreas ("I am cutting you the meat"), anaginosko soi Alkaion ("I r e a d you
Alcaeus"),kōmōidō soi tous Epitrepontas ("I play you the comedy ' T h e Epitrepontes'"
[of Menander]). A n d such sentences have passive equivalents, as we said [Ch. 178].
So I think it quite natural that H o m e r in these constructions with both cases
usually puts the dative before the accusative, since the dative includes it. [[The idea
seems to be that you can take something somewhere without implying that it's for
someone, but if there's a taking to someone, there must be something taken. This
is A.D.'s only treatment of verbs which regularly take direct a n d indirect objects.]]
184. Similarly the dative is governed by huperetō soi ("I serve you"soi dat.), douleuo
soi ("I work for you"), hepomai soi ("I follow you"), akoloutho soi ("I follow you"), heko
soi ("I have come to help you"??). [[This is a rare construction, t h o u g h Uhlig says
it's O.K. All the other examples are verbs which regularly and frequently occur
with the dative. Conceivably prosekb soi ("I belong to you") or eikb soi ("I yield to
you") would be better.]] T h e datives indicate beneficiaries of the implicitly be-
nefactive (peripoietikos) acts in the verbs. Douleuein ("to work, to slave") is a generic 298
term including all kinds of services, but if these detailed jobs are specified in the
active voice, they require a [direct object] accusative, e.g. tribose ("I massage you [se
ace.]"), louose ("I bathe you"), niptose ("I wash you") a n d similarly keiro ("I shear"),
kosmo ("I adorn"), smo ("I scrub"), anado ("I crown"), empleko ("I braid"), lampruno
("I dress. . . up"),phaidruno ("I clean"). Andeikein ("to yield"), which covers every
kind of submissive behavior, and hupochorein ("to give up") take the dative for the
same reasons. [[There are at least two textual cruces in this sentence, b u t the
general idea is clear enough.]]
222 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
185. T h e verbs of equal (reciprocal) influence, that is, those referring to the
action of two persons directed at each other, govern a dative, e.g. machomai soi ("I
fight you" soi dat.),palaio soi ("I wrestle you"), gumnazomai soi ("I train with you"),
hippazomai ("I ride"),anthistanomai ("I resist"), monamacho ("I duel"),pankratiazo ("I
pancration-fight"—[refers to a blend of boxing a n d wrestling]). It is obvious
e n o u g h that such verbs must take a complement (plagia), since there is a clear
transitive activity directed toward an object (hupokeimenon) person, as we said
originally [Ch. 159), a person who can perceive the activity as it reaches him. T h e
notions of wrestiing or pancration necessarily presuppose two living beings
(empsucha). But the aim of a verb of physical action which requires the accusative is
simply to have the object influenced by the subject without any counter-influence
[of object on subject], e.g. dero se, tupto se, paid se [all m o r e or less m e a n i n g "I give
you a beating", e.g. a schoolmaster to his pupil]. For any such assertion does not
entail a counter-beating [of teacher by pupil]. But the verbs we are now discussing
do not have such a meaning, a n d clearly that is why they do not take the accusative.
For they all entail the counter-influence from the person n a m e d by the dative, the 299
fighting back of the wrestler.
186. But why do these [reciprocal action verbs] take the dative [rather than the
genitive]? Because the genitive is specialized for the passive relation (diathesis).
This [reciprocal] construction, therefore, cannot accept either of the other two
cases [genitive or accusative], and therefore has nothing left except to take the
dative, from which it is easy to get the mutual benefit (antiperipoioumenon) of the
relation. In gumnazo se ("I train you") we have a one-sided relation, but in gum
nazomai soi ("I train with you"; [this construction is not otherwise attested, so we
cannot be sure that this is what it meant]) we have a mutually benefiting relation on
equal terms. Hippazo se [[presumably means "I ride you" a n d is addressed to the
horse, but the construction is poorly attested]] [is also a one-sided relation] a n d its
passive is hippazomai hupo sou ("I am r i d d e n by you" spoken by the horse), and,
again, the reciprocal relation is hippazomai soi [[the only attested m e a n i n g is again
"I ride you", said to the horse, but A.D. seems to mean something like "I race you
on horseback"]]. We could add many more examples of the same sort, but refrain
from doing so. But I think we should consider in this light the phrase peithomai
hupo sou ("I am convinced by you", passive) and peithomai soi ("I believe you" or "I
trust you" middle), which means precisely this mutual trust between two parties.
[[Grammarians classify the datives with these verbs (exceptpeithomai) as comitative
or dative of association, a use which belonged in I n d o - E u r o p e a n to the instrumen
tal case.]]
187. T h e r e are also some verbs that take the dative which occur only in the third
person [singular], e.g. melei ("care") and metamelei ("repent"), as in melei Truphoni
("Tryphon cares", "It is a concern to T r y p h o n " ) or metamelei [Truphōni]("Tryphon
repents", "It is a matter of regret to T r y p h o n " ) . T h e s e verbs vary from the syntax
of all others, in that the others may always have a nominative [subject], as in Theon
peripatei ("Theon is walking"), and there may also be an oblique [complement],
whose construction we have been discussing. But this is not the case with melei or
metamelei. W h e t h e r you say melei Sokratei or metamelei Sokratei ("Socrates cares" or
"Socrates repents"), there is no subject nominative present. This is why the Stoics
called these verbs parasumbama ("pseudo-predicate"), whereas the other verbs, 300
B O O K III: ON SOLECISM; V E R B S ; PARTICIPLES 223
189. This is the end of o u r account of the syntax of verbs with oblique cases, it
being well understood that the different tenses, different persons, and all the
other moods [as well as the indicative] partake of the same syntax. So temnoimi ("I
h o p e I may cut!"), though optative, takes the accusative [just like the indicative],
and so does an imperative or an infinitive or a subjunctive, a n d past tenses or
futures — so long as the voice is active. A n d many examples could be cited to
illustrate this.
224 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
190. Participles (metochē), too, use the same construction [as the verbs they are
derived from], even though they give u p the features of person a n d mood and so
differ from verbs. [[He's not really confident of this, and might have felt happier if
participles were considered another mood, like infinitives.]] As we have said
elsewhere, participles govern (pheretai epi) the same [complement] cases as verbs,
although other nominal (ptotika) derivatives of verbs do not. For you say kopto
touton ("I p o u n d this m a n [acc.]") but not *kopeus touton [with an accusative], but
rather kopeus toutou ("Pounder of this man"). O n e says aulei touton ("He flutes (to) 302
this man" [acc.]), but auletes toutou ("fluter of (to) this man" [gen.]) a n d gumnazei
touton ("He trains this man" [acc.]) but gumnastes toutou ("Trainer of this m a n "
[gen.]). So it is clear that all complements of verbal nominals (nouns and adjec
tives) are in the genitive, but not complements of participles. T h e y take the same
cases as the verbs they come from, and thus share (metechein) in the characteristic
features of verbs —gumnazon touton ("who exercises this m a n " [acc.]) and aulon
touton ("fluting (to) this man"). We have explained this point in more detail in our
discussion of the participle [i.e., no doubt, in the lost treatise "On the Participle"].
Book IV
1. (I) Now that we have finished the constructions of the verb, discussed in the 303
preceding book, which is Book III of this treatise, we can proceed to the construc
tions of prepositions. This calls for very careful investigation, since it appears that
prepositions form c o m p o u n d s with some parts of speech (meros tou logou), but
phrases (parathesis) with others, and characteristic features of accent are not always
reliable, though with most words the unitary nature of a [compound] word, i.e.
that it is just one word (meros logou), is shown by its single accent, whereas the fact
that two separate accents remain on the individual words (morion) of a phrase
shows that there are two words. [[Criteria for w o r d h o o d did not really trouble
linguists again until the 20th century, or at earliest the 19th, when many new
languages were being studied by Europeans. T h e separability test, mentioned in
3.132, is still used, as is word accent, in languages of this type.]]
2. T h e expression Dios koros ("Son of Zeus" or "one of the twins Castor and
Pollux") when accented with an acute-on-the-penult (paroxunetai) i.e. Dios koros
conveys the genitive relationship [i.e. that Dios is a genitive modifying koros],
parallel to Dios huios ("Son of Zeus"), but if accented only with an acute-on-the-
antepenult (proparoxunetai) [i.e. Dioskouros], it is a c o m p o u n d , similar to Diognetos or 304
Diodotos [proper names, etymologically "Zeus-born" and "Zeus-given"]. A n d when
eu nod ("I consider well" or, as a c o m p o u n d , "I am friendly") has two circumflexes
i.e. eu nod, it is clearly a phrase consisting of adverb plus verb [whereas eunoo, with
one accent, is a c o m p o u n d ] , Helles pontos ("Helle's sea") with an acute on the first
syllable, [i.e. Helles pontos], is a phrase, whereas Hellespontos is a c o m p o u n d , "the
Hellespont", and emou autou with a double circumflex is a phrase [meaning "of me,
m y s e l f ] , whereas emautou is a c o m p o u n d , [the reflexive "of myself, "my own"].
[[It's surprising A.D. did not use the accusative erne auton, which may elide to
em'auton, whereas emou autou cannot be shortened.]] These examples and the
thousands of others like them, when they are c o m p o u n d e d , take the [single]
accent [tonos] of composition, according to the rule (logos), but if they are not
c o m p o u n d e d [but are phrases], their tones remain separate, and as we said, by the
retention of the accent on each separate part, they show their status as phrases [not
compounds].
3. But katagrapho ("I write down", "I list").[for instance], does not indicate by its
accent whether it is two words or one [since kata grdpho is phonetically identical
with katagrapho, the grave accent having the same value as no accent]. A n d other
words of this sort, e.g. apoikou (either apoikou "of a colonist" or ap'oikou "from
home"), katapherontos (either kata pherontos "concerning a bearer" or katapherontos
"of a down-bringer") and the rest suffer from the same ambiguity (amphibolia).
[[Word division in books was rare until long after A.D.'s time—and so was accen
tuation, for that matter.]] So do apopse(now M.G. apopse "tonight" "this evening"
t h o u g h Uhlig evidently thinks it's a n A l e x a n d r i a n poetical e x p r e s s i o n ) ,
226 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
apodis[[since this word, a n d its parallel apotris, are known only from A.D., we have
no idea what they mean. Dis and tris mean "twice" and "thrice" respectively; maybe
apodis a n d apotris are m e r e synonyms, since apo is sometimes prefixed to adverbs in
late Greek without much effect. Since apo with n u m b e r s in Modern Greek means
"apiece", they may mean "twice each" and "three times each". T h e other passages
are 3.68,4.64 and 69.]] and, a m o n g conjunctions kathoti ("inasmuch as" vs.kath'hoti
"according to what"), dioti ("because" vs. di'hoti "because of which")—in these two
we will show [Ch. 26] that the prepositions [kata "according to" a n d dia "because
o f ] are present—and many more. We have u n d e r t a k e n to account in detail for
these. In the case of some parts of speech, there is no possibility of this uncertainty
about prepositions arising, since they take prepositions neither in construction
iparathesis) nor in composition—e.g. the nominative cases of all p r o n o u n s . T h e
reasons for this impossibility we will discuss in the a p p r o p r i a t e place.
6. Some words take a unique position (thesis) in the phrase a n d never change
it—hence never get involved with that sort of ambiguity— words such as e'toi
("either" in "either X or Y") a m o n g conjunctions, which is always prepositive
(protaktikos) and never postpositive (hupotaktikos), or, a m o n g verb forms, the sub
junctive verb, which never occurs without a subordinating subjunctive (epizeuktikos
"adjunctive") conjunction before it [e.g. edn "if, hina "in o r d e r that"]. T h e enclitic
personal pronouns, too, are incapable of occurring without the words (morion)
[preceding them] on which they d e p e n d accentually—timeson me ("honor me!")
and do smoi ("give me . . . !"). [See the long discussion above in Book 2, Ch. 57-89.]
T h e a r g u m e n t offered by T r y p h o n in his treatise on articles is not at all acceptable,
where he claims that the postpositive article [relative p r o n o u n ] becomes preposi
tive in sentences like hos an elthei meináto me ("Let whoever comes, wait for me!"
["headless relative"]) W h e r e is the n o u n it [hos] precedes? [[A.D. forgets that it is, in
fact, quite possible to have such a noun—hos an elthei stratiotes meinato me "Let
whatever soldier comes wait for me!"]] How can such a word be a postpositive
article when it does not refer back (anapolei) to any antecedent (proiipokeimenon)
n o u n , but, quite the contrary, occurs with indefinite (aoristos) reference at the
beginning of a sentence? Not only is hos in hos an elthei n o article [accepting Uhlig's
supplement for a gap in the MS], but neither is it when combined with another
nominal tis ("anyone") in hostis ("whoever"). It has to be combined with other
nominals, not indefinite ones, in which combination it shares in indefiniteness;
a n d [hostis] is even more indefinite than tis, a n d this is the exact opposite of the
essence of articles. Nouns which by themselves are semantically indefinite are
m a d e definite by the adjoining (parathesis) of the article to specify identity (horis-
mos) of referent [as being known], e.g.ho anthropos paregeneto ("The m a n [known to
both parties] arrived" whereas without ho the meaning would be "a m a n arrived"),
tonpoieten ethaumasa ("I a d m i r e d the poet"). [[The ancient identification of relative
p r o n o u n and article crippled A.D.'s treatment of relatives. Greek has a large a n d
varied set: besides hos, which A.D. considers t h e only t r u e "post-positive
article"—but only when it has an explicit antecedent before it—, there are hostis, an
indefinite relative, hosos, a relative of quantity, hoios of quality, helikos of size or age,
a n d hoposos, hopoios and hopelikos, indefinite relatives of n u m b e r , kind and size, as
well as relative adverbs hou, hoi andhothen of place ("where, whither, whence"), hote
a n d henika of time ("when") and hex of direction or way, hos ("as") of m a n n e r , plus
the indefinites hopou, hopoi, hopothen, hopote, hopenika, hopei, a n d hopo's. All of these
can be used with explicit antecedents, one of A.D.'s tests.]]
("What kind")], with indefinites and final-accent (oxeia) [the indefinites, which are
otherwise homonymous with interrogatives, are, when not enclitic, accented on
the ultima, e.g.poios "of some kind"]. So it is not totally strange for prepositions to
have two accentuations, a feature that they share with personal p r o n o u n s , which
have accented (orthotonoumenos) a n d enclitic forms, with the adverb hds initial (en
archei) as opposed to medial (en mesei lexei) [accented by A.D. with circumflex in the
meaning "so" or "thus", but with an acute or grave otherwise, where m o d e r n
editors make it a proclitic], which is circumflexed [in the "nevertheless" combina
tions kai hds "even so" a n d oud'hōs "not even so"], with the expletive e ("indeed" or
some change of stress or tone) which is circumflexed initially in sentences, but not
when postpositive, as Iliad 15.244:
4.3 ti e de su nosphin ap'alldn
("Why in the world [are] you [sitting here] away from everyone
else?")
or in this (comic?) [fragment]:
4.4 hoti e kalliston e kata Xanthippon
("because a very fine thing was I according to Xanthippus"??).
A n d one could add many more.
9. But conceivably the double accentuation of prepositions is a necessary
development. All the other parts of speech have a simple construction pattern that
they take (pheretai epi), e.g. adverbs are construed with verbs [A.D. never considers
adverbs that modify adjectives or other parts of speech, t h o u g h he might have
d o n e so at I.102 or C h a p t e r 20 below], even though there may be other words
(meros logon) in between [the verb and its adverb], articles go with nominals, 308
[subject] nouns with their verbs, a n d the verbs, in t u r n , are in relation to
[complement] nouns or pronouns, which may substitute anywhere for nouns.
[[One might deduce a basic sentence formula from this.]] But prepositions have
two distinct constructions, one with nouns [to make prepositional phrases "in the
house", etc.] and the other with verbs [ to make c o m p o u n d verbs, "overcome",
etc.]; a n d so it is natural that they have shifting accent, since the accent, by shifting
back [to t h e p e n u l t ] , signifies t h a t the p r e p o s i t i o n has b e e n t r a n s p o s e d
(anabibazetai), and by remaining [on the last syllable], indicates that the preposition
is in its true position (suntaxis).
10. In Odyssey 1.247:
4.5 Ithaken káta koiraneousi
("They rule throughout Ithaca."),
because of the initial (arktikos) accent, i.e. the anastrophe, we learn at once that the
preposition belongs with the preceding word [i.e. the basic o r d e r is kata Ithaken],
whereas if it were not so accented, it would be construed with the verb [i.e. make
the c o m p o u n d katakoiraneousi ("They rule completely")]. T h e r e f o r e , the dual
accentuation of prepositions is very desirable. T h e same a r g u m e n t applies to Iliad
19.4:
B O O K IV: PREPOSITIONS; A D V E R B S 229
Short vowels are easily prone to change of accent; note that in Aeolic, when all
accents are shifted (metatithesthai), they leave behind, for the most part, a short
syllable. [[Uhlig's text; the MSS say "they go, for the most part, onto short
syllables." Neither statement is true, so it's hard to choose. Uhlig says A.D. is
thinking of personal p r o n o u n forms such as humin to which the Aeolic equivalent
is ummin.]] W h e n prepositions get prolonged second syllables, they cease to be
liable to anastrophe, e.g. katai ("down along", for A.D. prolonged from kata),hupai 310
("under", normally hupo) etc. "But not all prepositions u n d e r g o anastrophe" you
say. Well, neither does every personal p r o n o u n have enclitic and accented variants
[discussed above 2.57-89], but only most of them. Neither does every interrogative
have recessive accent [the one exception is podapos? "where from?"], a n d similar
non-uniformity appears in many other parts of g r a m m a r . [[This is almost the only
hint of an allusion to the anomaly-analogy controversy in the Syntax.]]
13. We can also j u d g e by inserting the definite article. W h e n the article is a d d e d , 311
it yields first position to the preposition if the construction is a phrase, and follows
that on which it depends. But if the expression is a c o m p o u n d , this does not
h a p p e n , since the preposition here is a p a r t of the [compound] n o u n a n d so will
B O O K IV: PREPOSITIONS; A D V E R B S 231
have before it the so-called prepositive article. Examples of the phrase: peri
Aristarchou ("about Aristarchus")—peri tou Aristarchou ("about Aristarchus"—
addition of the article makes no translatable difference), kata Ktesiphontos-kata tou
Ktesiphontos ("Against Ctesiphon"),pari stephanou ("on crown")—peri tou stephanou
("On the Crown"—title of Demosthenes' speech opposing Aeschines' Against
Ctesiphon). Examples of the c o m p o u n d case: ho periklutos ("the renowned man";
never *peri ho klutos), ho sundoulos ("the fellow-slave"), ho metoikos ("the resident
alien"). Because of this, even potentially ambiguous sequences may have their
ambiguity resolved by the insertion (proslambanei) of the article. So there will be a
distinction between para tou nomou ("from beside the law"—presumably "the law"
is inscribed on a stone stele) or para tou pherontos ("from the one bringing. . . ") etc.
and tou parapherontos ("of the one serving") a n d tou paranomou ("of the illegality"),
which I have discussed in detail elsewhere [possibly in "On Prepositions."]
14. T h e r e f o r e we have a single word (meros logou) in proprokulindomenos ("rolling
on and on", used by H o m e r , e.g., in Iliad 22.221) and hupopepteotes (2.312 "cower
ing down"), and kata is part of a c o m p o u n d in 17.542:
4.11 leon kata tauron ededos
("A lion which has devoured a bull")
rather than a separate word in the sentence. For words (meros logou), properly
speaking, are members of syntactic phrases, not separated by tmesis from com
p o u n d s [i.e. kata and ededos form a c o m p o u n d katededos underlyingly], as will be
shown elsewhere [in "On C o m p o u n d s " no doubt]. Neither can peri ("around") in
19.362:
4.12 gelasse de pasa peri chthon
("The whole earth smiled all around")
be construed as a preposition with the nominative chthon ("earth"), but must be an
adverb meaning perix ("round about"), or else be shifted by hyperbaton from its
position as a prefix on the verb, making the underlying form periegelase [which
then (1) has a u g m e n t -e- deleted, (2) has -s- doubled, (3) has peri postposed],
meaning periss6s elamprunthe ("it gleamed excessively"). T h e accent shows us that it
can't be c o m p o u n d e d with chthon [into *perichthon], because it doesn't have penul
timate accent like [the g e n u i n e c o m p o u n d s ] autochthon ("earth-born") a n d 312
enosichthon ("earth-shaking"), and also because of the sense (he tou semainomena
ennoia). For it remains a feminine n o u n [as shown by pdsa "all", n. sg. fern.], [and
the c o m p o u n d would have to be endocentric with chthon as head], whereas com
p o u n d s generally transform the sense. For in autochthon and enosichthon the head is
not chthon "earth", but the respective meanings are (a) a man who comes from the
earth, a n d (b) a man [or god] who rules the earth. [[This last should be "who shakes
the earth"; probably A.D. is just absent-minded, but there may be something lost
from the text.]]
15. Such illustrations derive from simple observation, but more t h o r o u g h obser
vation shows the natural explanation. [[Here I am r e n d e r i n g Uhlig's emenda
tion.]] Nominatives and vocatives always refer to (deloi) the same persons as [i.e.
the subjects of] the verbs which they go with—Truphon anaginoskei ("Tryphon is
reading"), Truphon anaginoske ("Read, Tryphon"), Truphon paralambanei
232 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
16. Back at the beginning of this treatise [Bk. I, Ch. 9] we said that the elements
(stoicheion) of the sentence behaved similarly to the elements of the word (lexis).
Just as there are certain letters (stoicheia: phones, phonemes) which are prepositive
to certain others, not all to all, so also some words are prepositive but cannot go
before all words, but only before those which are capable of entering into con
struction with them. Suppose you take a syllable like tra or kla a n d insert an s. It
won't go before the r or the l, because s can never precede them [i.e. Greek has no
sr- or si- clusters], but only before the k or the t [since sk- and st- are common], so
yielding skla or stra. O r suppose you take a word Tirus and add (prostithetai) an n
after the second syllable. You can't put it after the s, although n is a common final
[in Greek], because s can never form a final cluster *-sn#, but only the reverse
[-ns,yielding the name Tiruns of the Argive town], and this [phonological] struc
ture is especially frequent in the Argive dialect. [[In fact, there are many Doric and
other dialects which show final -ns- where Attic-Ionic a n d Aeolic have lost it, e.g. in
the accusative plural endings -ons and -ans for which Attic shows -ous a n d -as with
loss of -n and compensatory lengthening.]]
taken as the abstract underlying present of the future periopsesthai "to look a r o u n d
upon"; "to ignore", whose actual present infinitive is perioran), nor epikouros ("ally"
from epi "in addition" a n d kouros "young man") from epikourein(" to be an ally",
actually from epikouros; there is no verb *kourein). A n d it's no objection to this if
epikouros (e.g.) is used as a name ("Epicurus"): for [in such cases] the primary
derivation (sunodos) yields an adjective form (prophora) [i.e. "assisting"], though it
may then be applied as a person's name. [[In fact, most Greek personal names are,
formally speaking, c o m p o u n d adjectives.]] (But in cases like Antitheos ("An-
titheus", "god-like") and Antipatros ("Antipater", "foster-father"), the c o m p o u n d
preserves the force (dunamis) of a prepositional phrase with a n o u n in the genitive,
i.e. the sense is anti theou ("in place of a god") or antipatros ("in place of a father"),
and, obviously, when the compounds are formed, they keep this value (idioma).
[[This type of c o m p o u n d is discussed in Smyth, par. 899.]] So since the combina
tion of prepositions with verbs, according to the above a r g u m e n t , is always com
position, not phrase construction, as will be proved [below, Ch. 32], necessarily
when there is a combination of a preposition with a n o u n in the nominative which
has an appropriately verbal sense (ennoia) the combination behaves like that with
verbs [i.e. is a compound].
18. Nouns in oblique cases, i.e. those which are not subjects [to auto prosōpon
delousa, "referring to the same person-i.e. as the verb"], do not take the verbal form
of composition [with prepositions], [[at least that's not the most usual way of
c o m b i n i n g with a preposition]]. Q u i t e t h e c o n t r a r y ; they p r e f e r p h r a s e -
construction to compounding. Wherever the n o u n and the verb are united [by
person-number agreement, as subject and predicate], there, as we said, the two
elements [preposition and verb or noun] are united in composition, but where the
n o u n is oblique and avoids verbal agreement (i.e. is not a subject), it also avoids 315
c o m p o u n d i n g . A n d in p h r a s e s t h e r e is m o r e case r e s t r i c t i o n t h a n in
compounds—peri Apolloniou ("about Apollonius" gen.), en Apollonioi ("in Apol-
lonius" dat.), etc. [[i.e. some prepositions take only one case, others two, others
three—see above, Ch. 12—but a c o m p o u n d noun-may occur in any case]].
19. Now maybe someone will say that there is composition even in oblique forms,
e.g. bolou ("of a cast" gen.) peribolou("of an enclosure") gen.; c o m p o u n d e d of peri
" a r o u n d " and bolos "casting"), dromou ("of a course", "of a race" gen.) katadromou
("of a playing field" gen.; c o m p o u n d e d of kata "down" and dromos "running"), and
so on. But this disregards the fact that these forms are not c o m p o u n d e d like this
(in the oblique case), but are merely inflected (klisis) into [that case] after they have
been c o m p o u n d e d in the nominative, as we will explain below [not in extant work].
20. Another plausible conjecture is this: " T h e prepositions are not really (under-
lyingly) construed with the oblique cases, but'have been transposed by hyperbaton
(huperbibasmos) from c o m p o u n d i n g construction with verbs [[this is a c o m m o n
assumption in interpreting Homer]], so that kata Apolloniou elalesa ("I spoke
against Apollonius") would be derived from katelalesa Apolloniou ("I talked Apol
lonius down"), sun Apollonioi ēmēn ("I was with Apollonius"—note the late middle
formemen "I was" for classical ēn) from sunemen Apollonioi ("I met Apollonius"),peri
ton kamptera edramon("I ran a r o u n d the turning-post"), from periedramon ton kampt-
era ("I r a n all a r o u n d the turning-post"), pros Truphona eipon ("I spoke to
234 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
to say para poian aitian hemarten Truphon? ("In respect to what cause did T r y p h o n
err") T h e same applies to meta mikron ("after a little") and met'oligon eleusetai 317
Truphon ("Tryphon will come after a little"); it is u n d e r s t o o d as meta mikron diastema
tou chronou("after a small interval of time"). T h e same a r g u m e n t applies to all such
examples, so that they must all be interpreted as prepositional phrases with the
prepositions in construction (parathesis) with an accusative case. So the question
para ti leipei ho logos'? ("In what is the sentence elliptical?") is equivalent to para tina
lexin leipei ho logos? or para tinaphrasin leipei ho logos? ("In respect to to what word or
expression is the sentence elliptical?"). Since it is not possible to use para with the
nominative, *para tis? ("to who?"), but only with the accusative para tina? ("to
whom?"), one must interpret ti as incapable of being construed as nominative in
prepositional phrases but only as an accusative, when we say para ti? ("because of
what?"). [[This problem depends u p o n the fact that there is never any difference
between nominative and accusative in the n e u t e r gender. This makes one of A.D.'s
expansions a little puzzling: meta mikron diastema tou chronou "after a short interval
of time" is not unambiguously accusative, since diastema is also neuter. All his other
expansions or parallel examples use feminine nouns (aitian, phrasin, lexin) or the
masculine-feminine p r o n o u n tis, acc. tina. T h e r e is no ordinary word for "inter
val" of time except diastema, but he could have said meta mikron chronon "after a
short time", since chronos is masculine. Perhaps he t h o u g h t he had.]]
24. But there may be a c o u n t e r a r g u m e n t . We said above [Ch. 13] that with a
prepositional phrase, the article stands after the preposition, but with a c o m p o u n d
it's placed before. Now we d o say to para ti ("the why"), which certainly wouldn't
h a p p e n unless the preposition were already united [in a c o m p o u n d ] .
25. But this, too, has an answer. First, the word ti ("what") cannot take an article
at all [except in quotation or hypostasis, as h e r e to "ti" = "the word ti"]. Interroga-
tives (peusis) are incompatible with articles, as we showed [in 1.86]. Second, the
example cited, to para ti, does not take its article because it contains an interrogative
(pusma), but the article applies to the whole phrase (suntaxis) and to the understood
action {pragma) [i.e. the whole sentence with its verb, b u t without the person, mood
or voice features], as we may u n d e r s t a n d in the construction with infinitives, to 318
graphein ("writing"), to peripatein ("walking"). For the article does not apply to
voices or tenses, b u t to the bare action (pragma) indicated by the verb. T h e
a r g u m e n t can be confirmed by applying the pattern (suntaxis) to a sentence.
Suppose we consider the sentence me para touto poiesometha ("Let's not act contrary
to this"), where again some action is to be u n d e r s t o o d , a n d the article is called for:
to me para touto poiesometha ("Let's do the thing (which is) not contrary to this"). So in
to para ti we do not have a case of c o m p o u n d i n g shown by the addition of the
article. [[This a r g u m e n t is not crystal clear. It is true that in Greek almost any
phrase or sentence can be nominalized with the n e u t e r article to, but A.D. seems to
d r a g in extraneous considerations, and also to use the prepositionp a r ain two quite
different senses.]]
and, if the latter, whether it is the neuter of hotis (an epic variant of hostis "who,
whoever"), a single word (meros logou) formed like hopoion ("of whatever kind, of
which kind, which") and hoposon("of whatever quantity, which") [by prefixing ko
to an interrogative tis "who"], or is itself two words, apposed in a kind of parallel
structure [of two near synonyms], whose masculine is hos tis ("whoever, who"—we
usually write it hostis in the masculine, though ho ti in the neuter), and whose
feminine is also by apposition (parakeitai) of two words, he tis ("whoever, who"
fern.), thus making the neuter also a two-word phrase ho ti ("whatever, which"
neut.). A n d I am well aware that the conjunction hoti is itself h o m o n y m o u s , having,
in addition to the causal sense ["because"] also the asseverative ("affirmative",
diahebaiotikos) sense, as when we say hoti nikō se ("that I conquer you", [[which A.D. 319
apparently considers an intermediate state leading from a deep "I assert that I
conquer you" to a surface "I conquer you"; hoti then is a kind of declarative
particle, just as eithe is an optative one a n d age an imperative one. See above
3.117-119]]), hoti pleiona sou anaginosko ("that I am reading more than you"),
patterns which differ syntactically and semantically from the causal use, as in hoti
pleiona sou anaginosko, sunetoteros sou kathesteka, ("Because I read m o r e than you, I
have become cleverer than you"). We have e x p o u n d e d this difference [between
the two hoti's] in detail in o u r "On Conjunctions" [p. 235], but since o u r present
problem concerns only the prepositional construction, we will discuss only that
part of the problem which involves prepositions.
accents, one for h ("rough"), one for no h ("smooth"). Hence A.D.'s treatise dealing
with the presence and absence of h is called "On Breathings".]]
29. But [this analysis] is confirmed by the properties [of dihoti a n d kathoti]. I n the
first place, dia never has the causal sense in c o m p o u n d s , e.g. diadromos ("passage"
or "errant") or diatrecho ("run across", "run through"), but only in phrases with the
accusative case, dia Truphona ("because of T r y p h o n " ) , dia ten hemeran ("because of
the day"). A n d so it is also clear that the hoti [in dihoti ] cannot be the conjunction
hoti ("that" or "because"), but must be a nominal in the accusative case, since we
have already shown [Ch. 12] that the nominative is never governed by a preposi
tion and that when there is a c o m p o u n d of dia in the nominative, the dia can't have
the conjunctional use [synonymous with heneka "because o f ' ] .
30. It remains now to determine whether hoti is the accusative case of a unitary
word [i.e. hotis, see above Ch. 26] or is a two-word phrase iparathesis) consisting of
ho plus ti, which is the neuter form derived from the masculine hos tis ("who",
"whoever"). T h e latter seems to suit the theory (logos) better. W h e n you consider
the other genders, masculine and feminine, which never have h o m o p h o n y of
nominative and accusative, the two-word analysis becomes obvious. A n d it is the
accusative that we find in di'hon tina logon("for whatever reason") or di'heri tina
aitian ("from whatever cause"), and also in the plural di'hous tinas [logons] ("because
of which [reasons]"), a n d di'has tinas [aitias] ("because of which [causes]"), a n d
hence also in di'ho ti. So it's clear that [the word dihoti] consists of three words, first 321
preposition dia in its causal sense which governs the accusative, and then two
parallel (i.e. appositional) nominals,ho ("which") and ti ("something"), in what can
only be the accusative case. T h e same proof also applies to kathoti [i.e., that it
consists of kata, ho, a n d ti.].
3 1 . Similarly it is indubitable that in expressions [with articular infinitive] like dia
to hemeran einai ("because it is day" "because of its being day"), dia to phos einai
("because it is light"), the article to ("the" n e u t e r sg.) cannot be any case but
accusative, as we have just noted. This can also be shown from correlative (an-
thupagomenos) clauses, [clauses where the subordinate clause contains a—usually
indefinite—relative and the main clause contains a demonstrative expression that
corresponds to it, "what . . . that . . . ", "whenever . . . then . . . ", etc.]. Given a
protasis (subordinate) clause [with masculine pronouns] di'hon tina logon tauta
egeneto ("For whatever reason that h a p p e n e d " ) , the corresponding apodosis (main
clause) is dia touton kai tauta sumbesetai("for that [reason] this also will happen"), or,
[with feminines], di'hen tina aitian . . . ("For whatever reason . . . ") dia tauten ("for
that [reason] . . . " ) . It must be the same also with neuters, dihoti hemera estin, dia
touto phos estin ("Because it is day, because of that it is light"), and the correlative
(corresponding, antapodidomenos) forms must be in the accusative case [as they
clearly are in masculine and feminine].
32. (VI) T h a t should be enough about the constructions of prepositions with
nominals in prepositional phrases, a n d about the c o m p o u n d i n g of prepositions
with nouns. As for prepositions with verbs, the combination is always c o m p o u n d
ing. This is the case with katagrapho ("I write down" from kata- "down" and grapho
"I write") etc., a n d also with [multiple prepositions as in] exhupaneste ("[the welt]
started u p [under the blow of the whip]"—Iliad 2.267—from ex "out", hupo
238 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
"beneath", ana "up", plus este "it got up") or parakatatithemetha ("we entrust . . . "
from para "with someone", kata "down" and tithemetha "we put"—form used by
Aeschines in Against Timarchus 9), etc. But perhaps someone will think that this is
not right, or at least needs a r g u m e n t to prove that it's composition [and that kata
etc. in the above examples are not i n d e p e n d e n t adverbs, like "down" in English].
O.K., suppose someone claims that these are verb phrases (parakeimenos) r a t h e r
than compounds (sunenomenos), using the following reasoning:
33. "No kind of c o m p o u n d word allows any change at the point of j u n c t u r e ,
whereas in phrases sometimes [the first word] may be varied. [The unchangeabil- 322
ity of compounds] is especially clear in cases where the first m e m b e r is phonologi-
cally less than a whole word [i.e. is a bare stem], such as leontophonos ("lion-voiced",
from leon "lion"), Menodoros ("Moon-gift" from men "moon" and doron "gift"),
kunodektos ("dog-bitten" from kuon "dog"), cheirographo ("I hand-write" from cheir
"hand" and grapho "I write"),paidagōgō ("I mind children" f r o m pais "child"). I n
these words, no matter how you change the inflection (klisis), there is never any
change at the c o m p o u n d boundary (sunapheia). Even in cases where whole words
(phone) [not just stems] are c o m p o u n d e d , this is clearly shown by the u n c h a n g i n g
character of the first word, as well as by the accent, e.g. kerasphoros ("horn-wearing"
from keras"hom"), heosphoros ("dawn-bringing", from heōs"dawn"), Astuanax
("City-king" from astu "city") [[in these three, the stem is identical with the
nominative]], tera-skopos ("portent-watcher" from teras "portent"; the form tera is
nom. p\.),Dioskoros ("Zeus-son", i.e. Castor or Pollux; Dios is the gen. of Zeus) [[this
last example is evidently a fused phrase "Son of Zeus"]].
34. "But real phrases do allow inflection at the boundary, e.g. Nea polis ("New
City", "Naples") Neas poleos (gen.), [Agathos Daimo'n ("Good Spirit")], Agathou
Daimonos (gen.), [Areios pagos ("Hill of Ares" "Areopagus")], Areiou Pagou(gen.);
the presence of two accents helps mark t h e m [as two-word phrases]. A n d in the
case of those prepositional sequences [which are ambiguous, see Ch. 13 above] like
paranomou(either "of an illegality" as one words, or "from a law" as two, though in
this use nomou "law" would have to mean a person) and periklutou ("of a very
famous one" or "about a famous one"), as we showed, if the article can be inserted
(epentithetai) between preposition and n o u n , as in para tou nomou, this proves a
phrasal construction, but if the preposition adheres (henotai) tightly so the article
must precede it, then the construction is a compound—tou paranomou ("of the
illegality"), tou periklutou("of the very famous one").
35. "And so in the case of [compound verbs like] katagrapho', if they are truly
compounded then they must be inseparable (ametabletos) [i.e. no conjugational
morphemes should come between] according to the above a r g u m e n t . Otherwise,
if they do take affixes (klisis) between [at the beginning of the second element] like
the simple verb, clearly then the preposition is syntactic, not c o m p o u n d e d . J u s t as 323
acceptance of the article [in the middle] was a mark of phrasal construction
(parathesis) in nominals, so accepting m o r p h e m e s (klisis) between the two parts in
verbs will show the same thing, and they do accept, e.g., reduplication (diplasias-
mos) [and syllabic augment], as in katagegrapha ("I have written down") from
katagrapho [where ge- is the reduplication m o r p h e m e marking the perfect aspect
stem], and katalelaleka ("I have spoken against") from katalalo [here le- is the
reduplication].
B O O K IV: PREPOSITIONS; A D V E R B S 239
37. "But if someone rejects these cases as being improperly inflected (keklimenos)
(though indeed they follow a very natural rule (akolouthia), as we have explained
elsewhere), then there is still the following a r g u m e n t , that only when prepositions
have their p r o p e r constructional characteristics d o they deserve their name "pre- 324
position", but when they are c o m p o u n d e d with any part of speech, they are no
longer really prepositions, but a part of the new c o m p o u n d word (onoma), and in
this case they don't have the characteristic features {idiomata) of prepositions [such
as having nothing before them].
38. "It is possible to get confirmation from what we said above (Ch. 12). When
ana ("up") in anoiktes ("opener", "one who opens up" an-oik-tes) has become a part
of the noun, it can then come in second place in a new c o m p o u n d thuranoiktes
("door-opener" thur-an-oik-tes), although the rule is that prepositions stand first in
c o m p o u n d s . A n d when kata in katadromos ("athletic field" kata-dromos) is com
p o u n d e d into a single n o u n , it takes the article out in front, no longer being first on
the g r o u n d of "pre-position", but second to the article. It was shown (Ch. 13) that
in prepositional phrases the article cannot stand before the preposition, as in kata
Ktesiphontos=kata tou Ktesiphontos ("Against Ctesiphon"), or ek Lesbou=ek tes Lesbou
("from Lesbos").
39. "So if preposition a n d verb are united into a single word in katagrapho, then it
would be quite natural for it to take prefixes (prosklinetai) [augment or r e d u p
lication] before the kata, which, as we just said, is no longer a preposition, but
merely the first syllables of a verb. But if it does not allow initial a u g m e n t or
reduplication, it must be conceded that it has the characteristic feature (idioma) to
mark it as a phrasal construction [i.e. kata grapho is two words]. T h e facts of accent
offer contributory evidence. It is characteristic of two-word phrases that the
240 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
second word preserves its accent unchanged; and that's what h a p p e n s in kateichon
("I held back", kata+ eichon, though *kdteichon should be possible by the rules
"verbs have recessive accent" and "when the ultima is short, the accent may be
acute on the antepenult"), kathepsa ("I fastened" from kata + hepsa), proelthon ("I
went forth" from pro + elthon), and so on with other c o m p o u n d s [[by the rule
which is now given as "the accent may not precede the augment"; the p h e n o m
enon obviously is limited to monosyllabic a n d disyllabic aorist and imperfect
indicatives, mainly from vowel-initial verbs plus a few Homeric or Doric third
plural second aorists and irregular imperfects like estan "they stood", ephan "they
said", edon "they gave".]]" [[This is the end of the devil's advocate a r g u m e n t which
began in chapter 33. A.D. now proceeds to answer it.]]
40. If you are influenced by these arguments you may believe that it has been
proved in detail that prepositions are not c o m p o u n d e d with verbs, but merely 325
combined in a phrase (parathesis). But it is possible to dispose of every one of the
arguments as follows. You shouldn't imagine that past-tense c o m p o u n d forms are
inflected (keklitai) from present compounds, but realize rather that each tense is
separately m a d e by c o m p o u n d i n g from the simplex tense. So just as katagrapho is
made from grapho, so kategrapsa is m a d e by c o m p o u n d i n g from egrapsa [and not by
inflection from katagrapho]. Each one of the verb forms, inflected also for voice,
accepts the prefix in the same sense in parallel prefixing operations on these
inflections because the underlying verbal idea (pragma) is the same. [[I.e. the
meanings of c o m p o u n d verb forms are as if all forms were derived from the same
base: e.g. from grapho "write" comes anagraphs "record", from egrapsa "I wrote"
comes anegrapsa "I recorded", etc., always with the sense "record".]]
Iliad 14.393:
4.17 xunisan megaloi alaletoi
("They came together with a loud shout")
etc., whose explanation I will give later [not in extant works]. So also in the case of
nouns, some, when c o m p o u n d e d , shift the accent back, while others preserve the 329
accent [of the u n c o m p o u n d e d second element].
48. (D) Most convincing is the fact that all [second singular imperfective or
thematic aorist] imperatives of active transitive (energetika) verbs, if they are
disyllabic, retract (anabibazei) the accent [onto the prefix]—kátelthe("come down!",
from kata + elthe), katdlabe ("Comprehend!" from kata + labe "take"), perigraphe
("Circumscribe!" from peri " a r o u n d " + grdphe "write"), and so there cannot be the
slightest doubt of their being c o m p o u n d s [[In I.E. these verbs were probably
enclitic to the prepositions, a possibility that A.D. doesn't consider.]] A n d if it is
conceded that these forms are c o m p o u n d e d , how can it help being granted that
the indicatives also are c o m p o u n d e d ? For they're both in the same voice, have the
same prepositional combination sundrome) a n d the same relationship (schesis)
between preposition and verb.
49. (E) Finally, in the subjunctive mood we find the special subjunctive conjunc
tions [such as ean " i f ] placed before the prepositions: ean katalabō ("if I com
p r e h e n d " ) , ean katagraphō ("if I record"). This wouldn't h a p p e n [if the forms
weren't compounds] since prepositions always take first place [in the phrase].
Similarly in the other moods, the special adverbs applied to them are a d d e d
initially [in front of the prepositions], eithe katalaboimi ("I hope I c o m p r e h e n d " ) ,
age katalabe ("Come on, comprehend!").
T h a t should be e n o u g h on the union of prepositions with verbs.
50. Obviously, the participles corresponding to each verb system will follow the
same pattern (schesis) as the underlying verb form, parastds ("standing by") from
pareste ("he stood by"),periphus ("clinging to") f r o m periephun ("I clung to"). A n d so
in the case oiparapherontos or para pherontos, if it is derived from the verb paraphero
("I carry past") it will be a single word [meaning "of one who carries past"]; but if it
is from the simplexphero ("I carry"), yielding pheron ("carrying"), gen.pherontos ("of 330
one carrying"), then, as we said [in Ch. 13], it will be two words para pherontos
("from one who carries"), and an a d d e d article will go between preposition and
participle [para tou pherontos "from the one who carries"].
5 1 . I n "On Participles" we r e m a r k e d that it is m o r e correct to call such forms
(schema) [i.e. participles of c o m p o u n d verbs] " d e c o m p o u n d s " [parasuntheta, i.e.
derivatives of compounds] rather than c o m p o u n d s , since when any part of speech
is c o m p o u n d e d , and then a different part of speech is derived (metapiptei) from it,
this derivative (metapeptōkōs) is no longer a c o m p o u n d but a d e c o m p o u n d . Hence
the accents of simple words and p s e u d o c o m p o u n d s [i.e. "decompounds"] follow
the same rules, as we showed in "On Forms". T r u e c o m p o u n d s show a shift from
ultima accent to recessive accent, as in the 2d declension c o m p o u n d s pánsophos
("all-wise" from sophos "wise") and nedoidos("new-singer" from aoidos "singer"). But
d e c o m p o u n d s do not retract (anabibazei) the accent, because they are not com-
244 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
57. A better account is to suppose that eisho or eis ho gets a temporal (chronikos)
structure from the construction (parathesis) of a preposition with a relative pro
n o u n [eis "to" plus ho "which"("postpositive article")], in which these words exhibit
their p r o p e r functions and senses appropriate to this construction. T h e preposi
tion [eis] loses its acute accent before a following word, as is the case with all
prepositions. [[Later the Byzantine custom, which we still follow, became estab
lished, of putting no accent on monosyllabic function words with initial vowels
unless one is needed to distinguish h o m o g r a p h s , so we write eis without accent.
But A.D. wrote eis before pause and eis elsewhere. In fact, most such words,
whether they began with consonants or vowels, were probably proclitics, pro
nounced in close linkage with the following word.]] T h e n the word ho with initial h
takes the acute (oxunetai), as is normal for an oxytone which is not followed by
another word. Of course, all oxytones change to barytones [i.e. change ' on the last
syllable to ] in construction with a following word [which is not enclitic], as is
illustrated by passages like Iliad 5.389
60. Just because we say these expressions (morion) [eis ho, ex hou, aph'hou, en hoi]
are built as syntactic phrases, don't suppose that therefore they must be entirely
outside the scheme of one-word parts of the sentence [hence cannot be adverbs],
just as other parts of the sentence are not rejected from consideration merely
because of irregularity in their internal construction {phono). [[It's just as vague in
Greek.]] Each of these expressions consists of two words, all right, but in the
combination (sunodos) they function like one. That's why we accept them as
united—I mean eis ho, en hoi and ex hou—just as we do the above-mentioned
adverbials in -de-oikonde, agrbnde ("to the country"). [[He seems to be arguing for a
borderline case, items which are externally single words, but internally two-word
phrases, but without any hint as to what such a status might mean.]]
6 1 . I will add an account of Heraclides' proposal that the temporal adverb epei
("when") arose from the same pattern. Starting with the local [relative] hou 335
("where"), he says, (1) apply the Doric transformation (metalepsis) ou.→ ei to it, as in
pou Doric pei ("where?"), autou, Doric autei ("right here"), then (2) prepose the
preposition epi ("on"), here equivalent to apo ("from, after"), as is attested in other
words, e.g. apomenisas (Iliad 7.230) used in the same sense as epimenisas ("being very
angry"). His assumed pattern of coalescence [of epi with hex to give epei rather than
*ephei] also is no obstacle to the etymology. For in Doric it happens very frequently
that unaspirated consonants are retained in elision (sunaliphe) before h- as in these
fragments of Alcman (fr. 87, a-e, Page):
4.20 ko toxoids Herakleees
("And the bowman Heracles" with ko from kai ho instead of cho as
expected),
kallisth' hupaulen
("to accompany most beautifully on the flute"; here kallisth' is
expected)
ka megasthenes Asanaia
("And the mighty Athena" with ka for kai ha instead of chd)
Melampoda t' Harpolukon te
("Melampous and Harpolycus" with t* where th' is expected)
and finally:
archoi men gar kothrasion
(possibly "may he rule over the Hothrasians also" or, reading
archoimen, "may we rule", but the name is unknown; here, presum
ably kōth- . . . is for kai hoth-).
62. But even if we grant the possibility of a substitution (enallage) of prepositions
[epi for apo] and loss of aspiration [epei for *ephei], the accentuation (ta tes taseos)
argues against this etymology. T h e [Doric] adverbs in -ei are circumflexed on the
ending, just like their equivalents in the common dialect—pou, pei ("where?"),
autou, autei ("right here")—which is not the case with epei. T h e meaning is also
against this. Forms which change phonologically (phone) from one dialect to
another do not change in meaning (deloumenon) as well. Now epei ("when, after,
since") does not signify place, as [Doric hei does] in this fragment of Sophron [fr.
98 Kaibel]:
248 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
65. In some of the above chapters [esp. 12-26] we have shown that prepositions
are used in their p r o p e r sense, in prepositional phrases, only when they are
assigned to their appropriate oblique cases (since they can't all be used with every
case, as we explained in o u r book called "Introduction") but in combination with
caseless words they are all only c o m p o u n d i n g elements. Hence, for this reason, as
well as the others which we have presented, they do not form phrases with the
nominative, which is caseless [[as in Aristotle's doctrine]]. T h e r e f o r e it is absolutely
necessary that the same principle apply to adverbs, which have no oblique cases by
which they could be recognized as governed by a preposition. A n d don't suppose
that dihoti is a counter-example; we proved [26-31] that this contains an accusative
[ho ti] governed by the preposition dia, not hoti the conjunction; but if it is hoti the
conjunction, then dihoti is c o m p o u n d e d into a single word just as epei ["since"] the
causal (parasunaptikos) conjunction is c o m p o u n d e d of the conjunction ei ("if) and
the preposition epi ("on"), as Posidonius says.
noun, and it's also essential to add the article in o r d e r to indicate the understood
(hupakouomenos) [i.e. deleted] nominal. It is a criterial feature of construction with
articles to presuppose nominals, since otherwise semeron ("today") has no right to
an article tes ("the"), genitive singular, feminine, since adverbs have no such
features. Of course there are adverbs where no u n d e r s t o o d nominal is possible,
and which therefore do not form prepositional phrases, e.g. apodis ("twice each"?),
apotris ("thrice each"). [[Both words otherwise u n k n o w n ; see Chapter 64 above.]]
70. A n d the adverb cannot take a preposition in construction before it unless
first the article is added [which is appropriate to the understood n o u n ] , a n d the
preposition and the article are to be taken as going with this u n d e r s t o o d n o u n , as
in en tei aurion("during tomorrow"), en teiechthes ("during yesterday"). [[Here, as in
69, the feminine article can only be explained as agreeing with the u n d e r s t o o d
word hemera "day", which is picsent in some kind of deep structure and has been
deleted.]] These phrases could not be said without the presence of the articles,
since prepositions govern (pheretai epi) only nominals, and cannot be construed
syntactically with adverbs. It is impossible to say *en semeron ("in today") or *ex
aurion ("from tomorrow").
7 1 . T h e same explanation applies also to cases like en tdi hupokatd ("in the
underneath") a n d en toi epand ("in the above"). T h e article [tdi] goes with the
implicit n o u n of place [or, perhaps, the implicit n o u n topos "place"], a n d is in 340
syntactic construction with the preceding preposition [en]. T h e r e f o r e , in parsing
(merismos) these phrases, one must say that the first preposition [en] applies in
phrase construction to the implicit dative n o u n and its accompanying article [tdi],
while the second [hupo- or epi-] is in composition with the local adverb [-katd, -ano].
72. This makes it clear why apechthes (otherwise u n k n o w n , probably "day before
yesterday" or "three days ago") cannot take (proslambanei) the article, because h e r e
the apo modifies a temporal adverb, and together they make one word [a com
p o u n d ] . But apo tes echthes ("from yesterday") can be said, with its own p r o p e r
construction, where the apo governs the [understood] n o u n [i.e. hemeras, "day"],
according to the same pattern as ek tes semeron ("from today"), which requires the
insertion (epenthesis) of the article. You can't say *ek semeron [without any article].
This is the proof, then, that we said above [Ch. 64] was required by the current
problem to make this classification (epimerismos) [of words like epand etc. as com
pounds] fully convincing.
73. (XII) Next we must discuss the word exaiphnes ("suddenly"; c o m p o u n d e d of
ex "out" and aiphnes "suddenly", according to us) whether it is a single c o m p o u n d
word of adverbial form (tupos) or is a prepositional (prothetikos) phrase (suntaxis)
with adverbial sense as is the case with countless others which also have synony
mous adverbs; eis oikon aperchomai = oikade aperchomai ("I'm going back home"), ex
oikou paregenomen = oikothen paregenomen ("I arrived from home"), en oikdise mend —
oikothi se mend ("I'm waiting for you at home"). We also have a phrase in oikon de, as
in Iliad 6.365: 341
4.24 oikon d'eseleusomai
("I will go home").
B O O K IV: PREPOSITIONS; A D V E R B S 251
opposed to oikade [which is a single word], showing that oikon de is two separate
words. T h e word u n d e r discussion [exaiphnes], because of the internal phonologi
cal changes (pathos) it has gone through, as we shall see, does not show clearly its
underlying structure (ta tes suntaxeos), t h o u g h all other words of this type leave no
doubt as to their formation.[[See Chapter 58 above.]]
74. F r o m adjectives of the type of aphanes ("uncertain") [i.e. s-stems], there are
derived synonymous feminine [abstract] nouns of two patterns (schematismos):
aphanid a n d aphdneia ("uncertainty"). Similarly from eusebes("pious") comes
eusebeia ("piety"), from which in t u r n come eusebia as in this fragment [Cal-
limachus 519 Schn.]:
4.25 eusebie tethneken
("Piety is d e a d " )
[and from eumathes "intelligent" comes] eumdtheia ("intelligence") but also eumathid
as in [Callimachus epigr. 48,1 Cahen.]:
4.26 eumathien eiteito didous erne
("In giving me [to the Muses] he prayed for intelligence.")
a n d another similar case is [Callimachus fr. 520 Schn.]:
4.27 adranie tode pollon
("much this (is) weakness"?? [hard to u n d e r s t a n d without context,
but adranie, i.e. adrania, is an alternate for adrdneia, from adranes
"weak"]).
So from the above-mentioned aphanes ("invisible, uncertain"), then comes
[aphanid "uncertainty, invisibility"], and with its genitive is made the phrase ex
aphanids ("out of uncertainty") according to the same adverbial pattern as appears
in ex antes ("immediately"), also feminine, as in [comic fragment?]:
4.28 ex antes idmen pros Dionusion
("Let us at once go to Dionysius").
T h e n the following phonological changes apply: (1) syncope [i.e. deletion] of the
medial a, yielding ex aphnids, (2) metathesis (huperbibasmos) of penultimate i into the
first syllable, ex aiphnds—this i is what conditioned the a [from e] in the ultima. (3)
Necessarily, then, there is change of a in the last syllable back to e, since the i which 342
conditioned the change of e to a is now gone, yielding ex aiphnes or exaiphnes. [[Once
again, only 2 is an implausible change in Greek phonology.]]
75. A n d there's nothing abnormal about the deletion ["syncope"] of a, since the
same thing occurs in sparganon ("swaddling-clothes" assumed to be from sparassd
"rip", hence underlyingly *sparaganon), hednon ("bride-gift, dowry" assumed to be
from hedd "please" hence *hedanon and *hedanon), teknon ("child" from tekd "give
birth to", hence *tekanon) [[The suffix of all three words is assumed to be underly
ingly -anon, not -non.]], daknd ("I bite" assumed to be syncopated from dankand "I
bite"), Erichthonidai ("sons of Erichthonius'" assumed to be s h o r t e n e d from
Erichthoniadai), and many more. Nor should the transposition (huperbibazetai) of i
bother anyone; a good parallel is apeiresia and apereisia ("boundlessness"). A n o t h e r
parallel is perhaps the genitive of doru ("spear"), which is first *doruos, and then, by
252 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
metathesis, douros; the same for gounos from gonuos, genitive oigonu ("knee"). T h e
reason I say "perhaps" is that it is also possible that gounos and douros are by
syncope [deletion of -at-] from gounatos a n d douratos.
76. We can throw light on the changes (pathos) leading to [exaiphnes] also by
comparing the corresponding adverb [aphno "suddenly, unexpectedly"]. From
the adjective aphanes ("invisible, uncertain") comes the regularly formed adverb
aphanos("secretly, obscurely") accented like (homotonos) the genitive plural apha-
nōn, just as eusebos ("piously") comes from eusebon ("pious" gen. pl.); we have
elsewhere ["On Adverbs" 169] pointed out the relation in form and accent of these
adverbs to plural genitives. (1) T h e n the aforesaid change (pathos) [deletion of a]
applies here, too [yielding *aphnos]. (2) And since, because of the change (pathos), it
no longer agrees in shape with the genitive plural, it no longer takes the same
accent, either (since, as we said, parallelism with genitives in form goes along with
parallelism in accent). So the accent moves back to the penult [giving dphnos], (3)
which then, necessarily, d r o p s its final -s [giving dphno]. For whenever the adver
bial ending (katalexis) agrees with the genitive plural in accent, it also keeps final -s,
but when it changes the accent, then it is liable to optional loss of -s, since it then
falls into the accentual pattern of adverbs in -ō, which are never derived from 343
regular adjectives (trigenes onoma), such as eso ("inside"), kato ("down") [[these two
are derived from prepositions, es ("into") a n d kata ("down along")]], kukloi
("around"), tonoi ("intensely"—attested only in Hesychius) [[these two are datives
of the nouns kuklos "circle" and tonos "tension".]]
77. Elsewhere we have observed that very often words (lexis) shift their accents
coincidentally with losing phonemes (stoicheion). Adverbs ending in -x take acute
on ultima, odax ("with the teeth"), hupodrax ("scowling"), okláx ("squatting"); but
when the final -x is deleted in hupodra ("scowling"), the accent also moves back to
resemble that of other adverbs in -a, like rimpha("nimbly"), mala ("very"), aip-
sa ("quickly"), with recessive accent (barunomenos). Similarly with the adverbs in -is
with long penult; as long as they keep -s they are oxytone—choris ("apart"), amphis
("about"), amoibedis ("alternately"), but when they lose their -s, they become
barytone [i.e. shift the accent to the left] like polldki ("often" for pollakis), and dekdki
("ten times" for dekdkis); a n d choris, losing -s, becomes chdri, on the model of authi
("again", from authis) a n d its class. A n d we also showed that when *despostes
[hypothetical underlying agent n o u n from despozo "I am lord over"] and ergastes
[rare regular agent n o u n from ergazomai "I work"] lose the medial -s-, they become
accented like elates ("driver"), arotes ("plowman"), hence despotes ("master") and
ergates ("workman") with recessive (bareia) accent. Elsewhere I have cited a mul-
titide of such examples. Those, then, are the rules that apply in the derivation of
dphno [and their justification].
78. We remarked above that some prepositional phrases with adverbial mean
ings have synonymous adverbs. So ex autes ("at once") is synonymous with autothen
("at once" and "from right here"): iomen autothen ("let's go right away"), apelthomen
ex autes ("let's go back at once"). T h e same sense appears in autodion in Odyssey
8.449:
4.29 autodion d'ara min
("At once [she bade] him [wash . . . ]")
B O O K IV: PREPOSITIONS; A D V E R B S 253
614 [79] [XIII] Adverbs of place occure in three varieties: place in which (en topoi) 201
place to which (eis topon), place from which (ek topou). A n d obviously place in which
is [temporally and logically] prior; (a) motion from a place presupposes being in a
place [before leaving it]; (b) and motion to a place presupposes that that motion
started from some place. This fact is clear also from the forms (phone) in use. For
the adverb pou ("where?") asks about place in which; from it is derived (paragogos) a
place-to-which adverb pose ("whither") [normal in epic; later poi is the normal
form] as well as pothi ("whither" or "where"), and for place from which pothen
615 ("whence"). So from hupsos ("height") comes hupsou ("on high"), a n d from it come
hupsothen ("from on high") and hupsothi [[A.D. appears to believe it means "to the
heights", but it is only attested meaning place where, like hupsou]].
[80] Just as in regard to the feature of gender there are nouns with only one
gender form (prophora), e.g. masculine, or only feminine, or neuter—so called
one-gender (monogenes) nouns [i.e. ordinary substantive nouns like "chair",
"river", "house", etc.], a n d as there are also three-gender (trigenes) nouns with a
distinct form for each gender [i.e. adjectives], e.g. kalos(m.), kalē (f.), kalon (n.)
("beautiful" or "fine"), a n d still others [also adjectives] in which the masculine a n d
feminine forms are not distinct, e.g. pseudes (m. or f.), pseudes (n.) [i.e. so-called
two-termination adjectives], and likewise some nouns with two genders [but only
one form—mostly names for animate beings whose sex is distinct] like hippos (m. or
f.) ("horse", either "mare" or "stallion") and its class, so also with adverbs of place it
is possible to find some that express only a single local relation, so that there is no
possibility of deriving adverbs for the other relations, such as the adverb pei ("by
what way") [[We might consider the adverbs in -ei as expressing a fourth local
relation, that of "path" or "direction"; they are formed mainly from demonstra
tive, relative, interrogative a n d indefinite stems, however, not from nouns like
hupsos or oikos. ]] and the others we mentioned, and likewise to find others that have
three distinct forms like the three-termination adjectives, e.g. from oikou ("home"
gen.; the genitive is often taken as base for derivatives) we get oikothi ("at home";
more often oikoi), oikothen "("from home") and oikade ("home(ward)"), where each
adverb takes its appropriate distinct form (phone).
[81] T h e r e are also some with partial syncretism (epikoinonei), showing identity of
form (charaktera) for place to which and place where, but not for place from
which—e.g. and ("up", "upward") and the rest, about which we will speak [Ch. 88].
It is rare for a single form (phone) to express all three local relations, as is the case
with prosthen ("in front", "forward", "from in front") and any others there may be
like it. T h e above local adverbs show a tendency toward syncretism (sunekdrome)
B O O K IVa: A D V E R B S (continued) 255
according to a kind of semantic kinship between adverbs of place where and 202
adverbs of place to which, just as similar tendencies appear in other parts of
speech, e.g. a tendency in cases toward syncretism of nominative a n d accusative or
vocative, and of genitive toward dative, and for this reason there are n u m e r o u s
examples of such syncretism and coincidence (sunemptosis) of form, e.g. in the
dual, where this is always so [[i.e. nom-acc-voc are always alike and gen-dat are
always alike]], in the neuter gender, where nominative, accusative and vocative
always coincide, and in many masculines where nominative and vocative are alike
[this is always so for feminine singular nouns], so as to encourage enallage {schema)
of these cases in poets [e.g. use of nominative for vocative], as in Iliad 3.277:
"thus", later "here") acquired the place-to-which sense (schesis), as in hōde elthen ho
deina ("Here came so-and-so"), and then the place-where sense according to the
semantic kinship (sumpatheia) discussed above between place-to-which a n d place
where: hdde se mend ("I'm waiting for you here"). Perhaps the Homeric usage {ethos)
may seem to be against this construction of hdde as an adverb of place, since in the
epic it is never local, but only indicates m a n n e r (poiotes) and is equivalent to houtds
("thus"). We have discussed this matter fully and decisively in "On Adverbs"
[178-9?] showing that while, from the point of view of the form, hōde ought to
mean "thus", its frequent juxtaposition (parathesis) with adverbs of place caused it
617 to acquire a local meaning itself.
[84] T h e r e are, however, some rather rare adverbs in -de where this suffix 203
(paragdge) does not have a local meaning, but one shared with other parts of speech
[namely, demonstratives]; words with this suffix a d d e d don't actually mean any
thing different from the same words without it. T h e forms which admit this
e n d i n g belong to the class of correlative (antapodotikos) words (morion) beginning
with t- -tososde ("so much"), toiosde ("such"), telikosde ("so big"), [[and, of course, hode
"this", which does not begin with t- in the nominative masculine or feminine,
though all the other forms do]]. Parallel to these we also have henika("when"),
tenika ("at this time"), tenikdde ("at this time"), emos ("when"), temos ("then"), temosde
("at this time"). We must, however, observe that while the extended forms are
made from correlatives in t-, it is not the case that all correlatives in t- may take the
-de extension (epektasis). For instance, ophra ("while") has a correlative tophra
("meanwhile") and heds ("as long as") has teds ("so long"), but there is no Hophrdde
or *tedsde. Perhaps it's because these forms are deficient (ellipes) in other respects:
for one characteristic of the extendable words is that they have a corresponding
interrogative (pusma) [e.g.posos "how much",poios "what sort",penika "when", etc.],
but these deficient [words] [teds and tophra] have no matching interrogative forms
(prophora).
[85] T h e r e are also some counterarguments to this, e.g. that hode ("thus") from
hds ("thus") doesn't begin with t-, and neither does enthade ("here") from entha
("here"), and yet it is obvious that these are not instances of the local adverb
f o r m a t i o n [-de m e a n i n g " t o " ] , for t h e r e a s o n t h a t t h e local derivatives
(parachthenta) regularly accent the third syllable back from the end, as indgrade ("to
the country") or oikade ("home"), but enthdde does not. T h e local adverbs do not
have a meaningful r e m a i n d e r when -de is removed, unless they are not really
derivative adverbs but phrases formed by the construction of the "conjunction"
[postposition] de with an accusative case, e.g. Oulumpon de ("to Olympus") etc. This
test makes hdlade ["to the sea"] dubious as adverb [because hdla is the accusative of
the word for "sea"]; at any rate, it is not like enthdde ["here"]. So, since entha
("here") remains after the de is removed from enthdde ("here"), this is a distinctive
mark of words formed with the de extension, namely their having the same
meaning after deletion (apobole) of de, as in tenikdde ("at that time") and tenika ("at
that time"), temosde ("then") and temos ("then"), hode ("thus") and hō ("thus"), the
Doric form, as in the verse [of u n k n o w n source, quoted above in 2.114 as well as in
"On Adverbs", with minor variations in spelling]:
4.34 ho te chernatis gune
(perhaps "and thus a poor woman")
B O O K IVa: A D V E R B S (continued) 257
and, in the case of nouns (onoma) toiosde ("such") a n d toios ("such"), tososde ("so
much") and tosos ("so much"). This is the -de which is suffixed in enthade [i.e. one
which is attached to demonstratives to add the deictic notion "near me"]. [[These
demonstratives, which we would class as p r o n o u n s , are nouns for A.D.]]
618 [86] So if it is a mark of words derived by the demonstrative extension -de to
accent the p e n u l t (pro telous) with an acute, w h e n it is short, like temosde
("then"), tenikdde ("at that time"); then enthade shares the same derivation. A n d
obviously enthende ("from there", "thence") has the same extension (parolke), since
it can never signify place to which, obviously. Just as these forms differ [from
place-to-which adverbs], one set being made with the demonstrative extension de,
and the other consisting of two words in each case, each one with its own accents,
one on the [head word], the other on the de, as in Odyssey 14.36:
4.35 ho de proseeipen anakta
("And he addressed the king"—here the de is the conjunction
"and");
so also enthade ("here") differs from entha de ("and here") as in Odyssey 12.85: 204
4.36 entha d' eni Skulle naiei
("And inside here [in the cave] Scylla lives")
T h e word enthade is accented-on-the-penult, but entha de is accented on the
antepenult (proparoxunetai) and has the coordinating conjunction de as its second
word, as again in Odyssey 10.1-2:
4.37 entha d' enaienlAiolos Hippotades
("And h e r e [on the island of Aeolia] lived Aeolus son of Hippotas").
T h e adverbs in -ze signify only the place-to-which relation—eraze ("to the
g r o u n d " ) , thuraze ("to the door"), Aphidnaze ("to A p h i d n a e " ) , Thebaze ("to
Thebes"), etc.; and this is natural, since z and d are almost interchangeable
(isodunamei) phonologically, e.g. in kluzd ("surge") andkluddn ("surf"),knizo ("itch")
and knide ("nettle"), hezō("to seat"), hedos ("a seat"). [[In fact, z is nearly always the
final form of some d or g cluster. I n eraze etc., the underlying forms are eras-de,
thuras-de, Aphidnas-de and Thebas-de, which are all instances that A.D. would classify
as accusative plus postposition de (except for eras, which does not occur as an
i n d e p e n d e n t noun). In the verbs just cited, the z results from d plus y: kludyd,
knidyd, hedyō.]]
[87] Adverbs ending in 6 are always local when derived from prepositions, as in
and ("up" from ana "up along") etc.; but when [actually ending in -di, the dative
case form] from nouns [including adjectives], they present the relation inherent in
the n o u n with an adverbial sense. So from the verb teinein(" stretch, strain") comes
the n o u n tonos ("tension, accent, intensity"), whose sense in retained in tondi
("intensely"); it gives intensity (epitasis) to the action of the verb. Similarly the
adjective aneōs means "mute", "silent", and from it is derived the adverb anedi with
the same sense "mutely", "silently". [[All supposed occurrences of such an adverb
are apparently simply the nominative plural of the adjective.]] And obviously the
adverb kukldi ("around" "in a circle") doesn't acquire a local meaning by its suffix,
but simply because it comes from kuklos ("circle") which has a local reference.
258 T H E SYNTAX O F A P O L L O N I U S DYSCOLUS
Similarly porrō ("far") has an Attic extension [i.e. the doubling of r] and is derived
fromporos ("way", "passage"), which again has a local sense. [[This etymology is
incorrect; porrō is in fact from porsd which is a variant oiprosd "forward", which he
619 mentions in the next section.]]
[88] But the adverbs in o" which are derived from prepositions regularly indicate
a local relation, perhaps because the prepositions themselves, along with their
other meanings, have local senses (schesis), especially in composition with verbs
which, themselves, signify motion in space, e.g. anabainei ("he goes u p " : ana " u p "
+ bainei " goes"), katabainei ("he goes down": kata "down"). W h e n prepositions are
used with nominal [complements] they have many other, non-local meanings as
well, as in kata Ktesiphontos ("against Ctesiphon"), dia Truphona ("because of
Tryphon"),peri emou ("about me"). [The pattern in -o] is represented by ana-dnd
("up"); kata-kdtd ("down"), prósō ("toward—forward"),es-esd ("into—inside"),
ex-exō ("out of—outside"). Each one of these forms in 6 signifies both the place to
which relation a n d the place where relation: and erchomai ("I am going up"), and
mene ("stay up!")—Odyssey 21.369:
4.38 atta, prosō phere toxa
("Nurse, bring forward the bows!")
and similarly with the others. So anyone who thinks that Odyssey7.13:
4.39 eisd dorpon ekosmei
("She fixed a meal [for him] inside.")
is incorrect [because eis is a place-to-which preposition only, "to" or "into"], cannot
be right, because eiso, like all the others, allows both the place where use a n d the
place to which, as in Euripides Phoenissae 262:
4.40 di' eupeteias teichedn eisd molein 205
("to go inside the walls with ease").
[89] But people claim we o u g h t to say endon ("inside") [for the place where use,
because it comes from en "in", the place where preposition]. But that itself may be
thought anomalous (alogos), since all the other adverbs in -don are qualitative
(poiotes) in sense, not local: botrudon ("in clusters"), oiakedon ("like a r u d d e r " ) ,
ageledon ("in herds"), donakedon ("like a reed"). [[Two of these, oiakedon and
donakedon are known only from A.D., a n d are probably taken from some lost
poems.]] "Surely it is unreasonable to reject what is regular and justifiable [the use
of eiso for place where] and prefer as m o r e correct what is irregular [endon] both in
phonology, since it is not accented on the ending, and in meaning, since it does not
agree in sense with any other adverbs in -don. In defense of the usage [of endon] we
can argue that, [except for eiso], all the other adverbs derived from prepositions
[i.e. those in -d] quite naturally accept the double usage, [for place where and place
to which, both], because they have no related (parakeimenon) adverb to express the
place where relation; but eiso, since it has endon to contrast (antiparakeisthai) with it,
would appear to be unjustifiably used in the sense of endon. [[But it is so used from
H o m e r on.]] For any part of speech the test of ungrammaticality (akatallelotes) lies
in the existence of another usage (chresis) [i.e. word or suffix] which is specialized
in the opposing sense. So graphei ("he writes") can't be used with first or second
B O O K IVa: A D V E R B S (continued) 259
[92] [There is one passage where] the reading with thi must be rejected, namely
Iliad 3.3
4.41 ouranothi pro
("in front of heaven").
For the context here does not call for a place-where construction (schesis), but a
genitive case, i.e. pro ouranou, since the flight [of the cranes] is not actually in
heaven [but before it]. So the correct reading must be either ouranophin or
ouranothen; by this I d o not mean to claim that forms in then [always] indicate the
place-from-which relation which normally is expressed in the genitive, or that they
are as a rule syntactically adverbs. We will show [Ch. 98] that forms in then do not
[necessarily] signify place from which, but may merely have the sense of the base
form [which is the genitive of nouns a n d the place-where form of adverbs].
Forms ending in phi have only the air of being adverbial derivatives, but are
not basically adverbs, but rather like nouns which often accept an adverbial syntax
and are then called adverbs. T h e explanation is as follows. Every adverbial
derivative expressing a local function can be resolved by one single case, e.g. oikothi
into en oikoi (with dative, "at home"), oikothen into ex oikou(genitive, "from home"),
and oikade into eis oikon (accusative, "(to) home"). So we must apply this a r g u m e n t
to words in -phi.
[93] T h e form [in phi] is genitive in passalophi (occurs with ek in Odyssey 8.67,105
and with apo in Iliad 24.268—"from a peg"), and dative in Iliad 2.363:
4.42 hōs phrētrē phrētrēphin
("so brotherhood [will help] brotherhood")
and accusative in Iliad 13.308:
4.43 epi dexiophin
("on the right"; m o d e r n editors would say "genitive" rather than
"accusative" here)
and therefore it will be conceded that this form [in -phi] syntactically belongs with
the nominal cases rather than with adverbs, so that it occurs even when preposi
tions are placed before it [[this translates Uhlig's emendation]]. This is not true of
adverbs. Not even forms like toiosde ("such") a n d tososde ("so much"), though they
take [what seems to be] the adverbial e n d i n g [-de] were called adverbs, (a) because
they still have case inflection [before the -de: toiosde, toioude, toioide, toionde, etc.],
and (b) because of the u n c h a n g e d meaning [i.e. toiosde = toios]. [[A.D.'s instincts are
sound, whatever one may think of his arguments; the evidence of Linear B shows
even more clearly than H o m e r that forms in -phi are case-forms, not adverbs. See
Householder 1972.]]
[94] Sometimes a phrase m a d e by placing a preposition in construction with [a 207
form in -phi] will express a local relation (topike schesis), but to be an adverb, a form
622 should indicate a place relation by its formation alone. So local adverbs can be
transformed (antimetalambanetai) into prepositional phrases: oikothi-en oikoi ("at
home"),—but you would never say *en oikothi-, oikade-eis oikon ("(to) home"), never
*eis oikade. T h e r e f o r e it is also ungrammatical to say ex oikothen ("from home"), a n d
such phrases as ex ouranothen ("from heaven", e.g. in Iliad 8.19) are instances of
B O O K IVa: A D V E R B S (continued) 261
poetic pleonasm (parolkē). We'll discuss this later (Ch. 98). So if this rule holds, then
we don't have adverbial constructions [but true prepositional phrases] in examples
like ek passalophin ("from the peg", e.g. Odyssey 8.67), kaf oresphin("down from the
mountains"),etc. This is why we would prefer the form with -phi in ouranothi pro,
[rather than that with -then], because the form with -then is a self-sufficient adverb
derivative. [[Modern editors disagree.]]
[95] Adverbs ending in -ou signify the place where relation a n d are regularly
circumflexed on the e n d i n g [-ou], like autou, This is a p u r e adverb, too [like forms
in-thi], as in Aratus, Phaenomena 225:
4.44 autou kai krioio
("here also [are the swiftest routes] of the Ram")
or Odyssey 4.785:
4.45 hupsou d'en notioi
("And far out on the water . . . ")
anchou ("nearby"), telou ("far away"). From this it can be proved that pou ("where")
[basically] indicates place where, as is regular in H o m e r , b u t by extension is also
used for place to which, d u e to the tendency toward syncretism (sumpatheia)
mentioned above, as inpou aperchei ("Where are you off to?"). We should add that
the Doric transformation of such [adverbs is with el for -ou], as in S o p h r o n (fr. 98
Kaibel):
4.46 (=4.21) hei ta ton choiranchan
("Where the . . of the hog-chokers?"),
("from down under"). In the Doric dialect, too, they have this derivative from
pronouns, e.g. toutothen ("from here") which they customarily alter by deleting
(apokoptei) the last syllable (-then) and lengthening (ekteinei) the o to ō, e.g. S o p h r o n
(fr. 85 Kaibel):
623 4.48 toutō thametha
("We watch from here.")
just as for autothen we have (fr. 23 Kaibel): 208
4.49 auto hore'is, Phuska
("Do you see from here, Physka?").
T h e same formation is present also in (fr. 125 Kaibel):
4.50 po tis onon ōnaseitai
("From where will one buy a donkey?").
T h a t these are special Doric forms is evident from the fact that there are many
place-from-which adverbs in -then, but [except in Doric] no place-from-which
adverbs ending in -6. [But these Doric adverbs in -ō do mean place from which, and
are derived from adverbs in -othen.] For it is well-established that in the changes
(heteriosis) made by a dialect it is only the sounds [i.e. endings or stems] that are
altered, never the meanings (deloumenon). So the meaning (semainomenon) of pothen
("from where?") is preserved u n c h a n g e d in [Doric] pō and that of autothen ("from
right here") in auto. [{The doctrine that dialects differ only in phonology a n d
lexicon, i.e. that a direct m o r p h e m e - b y - m o r p h e m e r e n d e r i n g is always possible
and semantically identical, is almost correct for many dialect areas. See above, Ch.
62. T h e suggestion that the underlying form of all other Greek dialects is Attic or
koine", however, is something which must have developed first at Athens and later
at Alexandria.]]
[97] This suffix, when a d d e d to personal p r o n o u n s , preserves the meaning of
the base, e.g. frome emeu ("of me") emethen("of me"),seu ("of you") sethen ("of you").
[[As often, the genitive case is here taken as the base.]] As a rule, only p r o n o u n s
[not adverbs] can be used with personal reference [i.e. refer to definite entities].
Now you may object, "But what of the fact that adverbial forms (schema) are indeed
derived from pronouns?" [I.e. emethen and sethen]. T h e answer is that they have the
form (schesis) of adverbs, but not the construction of real adverbs derived from
genitives, [but rather that of genitivies]. Adverbs of any derivation at all char
acteristically express a single sense (ennoia), and so don't need to have prepositions
added. But genitives, because they need the prepositional construction [to make
the meaning clear], necessarily take prepositions—apo Aristarchou ("away from
Aristarchus"), ex Aristarchou ("out of Aristarchus"), ex oikou("out of a hall"), ap*
oikou ("away from a hall"), and so on with other nouns. You cannot say *apo
Delothen ("away from Delos") or *ap' oikothen ("away from home"). W h e n n o u n s are
involved [with prepositions], you must use the genitive case form, whereas
[constructions like ex emethen "out of me" or apo sethen "away from you"] are
peculiar to p r o n o u n s .
[98] But they do occur occasionally with nouns. I n Iliad 3.276, instead of the
genitive Ides we have:
B O O K IVa: A D V E R B S (continued) 263
[100] T h e local adverb heichi ("where") doesn't signify place to which by virtue of
its final derivative suffix; that syllable is rather a r e d u n d a n t syllable (parolke) as in
naichi ("yes!") [from nai] or ouchi ("no!")fromow. T h e phonology also shows this; in
a normal single word the initial syllable could not begin with h [i.e. h would be
deleted by what we now call Grassmann's law], since n o syllable preceding ch
begins with h (dasunetai). [[He could have stated the law more generally: "no
syllable preceding an aspirated stop (ch, ph, th) begins with h or an aspirated
stop.]] In the same way as nai [in naichi] is an i n d e p e n d e n t word (free form), with
chi an optional addition, and likewise ou, so also hei, signifying place to which
[[rather route by which ]], is a free form (enteles). This is also clear from the
correlative (antapodidomenos) forms, which do not share this optional extension,
625 namely tei ("this way"), the interrogative pei? ("which way?"), a n d the indefinite
[relative] hopei ("whichever way"), all with iota subscript [[i.e. tei, pei, hopei, not te, pe,
hope; this final -i was silent by A.D.'s time]] according to the tradition, so that they
rather resemble [feminine] dative singular forms, e.g. as in tautei poreuthomen
("let's march this way"; in tautei, unlike pēi and hopei, there is actual homonymy
with the dative singular feminine of houtos "this"). T h e Doric forms pei ("where?")
and hei ("where") etc. should not be r e g a r d e d as equivalent topei and correlative tei
etc., since the Doric forms signify place where rather than place to which and so
correspond rather to pou ("where?") and hopou ("where", "wherever"). So in the
Sophron fragments [see above Ch. 95], in hei ta ton choiranchan (fr. 98 Kaibel) hei is 210
equivalent to hopou ("where", relative) a n d in pei gar ha asphaltos (fr. 5 Kaibel) pei is
equivalent to pou ("where?"). And anyhow Attic adverbs in -ei are matched [in
Doric] by forms m-ai rather than-ei, e.g. Atticpantei = Doricpantdi ("every way"),
allei = allai ("another way").
[[This is all we have of the Syntax, though references elsewhere show that
Book 4 must have been longer.]]
REFERENCES
A. Apollonius Dyscolus: Editions
Works, ed. by Schneider and Uhlig (Grammatici Graecill, 1-3). (Leipzig, 1878-1910; reprinted 1965.)
Syntax, translated into German by A. Buttmann. (Berlin, 1877.)
Syntax, ed. by Sylburg and trans. into Latin by F. Portus. (Frankfurt, 1590.)
References to other Classical authors are to standard editions of their works. Some, but not all, of these are listed
under B.
B. Secondary Literature
Adams, D.Q. (1972): "Relative Clauses in Ancient Greek", in: Peranteau, Levi and Phares (eds.):
The Chicago Which Hunt (Chicago), pp. 9-22.
Barwick, K. (1957): Probleme der stoischen Sprachlehre und Rhetorik. (Berlin.)
Barwick, K. (1922): Remmius Palaemon und die rbmische Ars Grammatica. (Leipzig: Hildesheim.)
Bloomfield, Leonard (1933): Language. (New York: Holt, Rinehart & Winston.)
Camerer, R. (1965): "Die Behandlung der Partikel an in den Schriften des Apollonios Dyskolos",
Hermes 93:168-204.
Cooper, Guy L., III (1972): "In defense of the special dual feminine forms of the article and
pronouns td, tain tauta, tautain, ktl., in Attic Greek" ,Transactions and Proceedings of the
American Philological Association, ed. J.J. Keaney, Vol. 103:97-125.
Dionysius Thrax: Techne Grammatike, ed. by G. Uhlig. (Leipzig, 1883.)
Donnet, D. (1967a): "La place de la syntaxe", LAntiquite Classique 36:22-48.
(1967b): "Le traite peri suntaxeos logou de Gregoire de Corinthe", in: Etudes de
philologie, d'archeologie, et d'historie anciennes, publiees par L'Institut Historique Beige de
Rome, Vol. X. (Brussels & Rome.)
Dronke, G. (1857): "De Apollonii Dyscoli techne grammatike', Rheinisches Museum filr Philologie XI,
549-85.
Egger, E. (1854): Apollonius Dyscole, essai sur l'histoire des theories grammaticales dans l'antiquite.
(1877): "Observations sur le vocabulaire technique des grammariens etc.", in:
Annuaire de l'association pour l'encouragement des etudes grecques, pp. 138-46.
Egli, Urs (1970): Zwei Aufsatze "Zum Vergleich der stoischen Sprachtheorie mit modernen
Theorien". in Arbeitspapier Nr. 2, Universitat Bern, Inst. fur Sprachwissenschaft.
Hahn, E. A. (1951): "Apollonius Dyscolus on Mood", Transactions of the American Philological
Association, 82:29-48.
Hoerschelmann, W. (1880): "Kritische Bemerkungen zu A.D. de pronomine", in: Rheinisches
Museum 35:373-89.
Householder, F.W. (1972): "Variation in Early Greek", in Greek, A survey of Recent Work by F.W.
Householder and G. Nagy (The Hague: Mouton).
Jakobson, R. (1936): "Beitrag zur allgemeinen Kasuslehre", TCLP 6,240-88.
Kemp, A. (1978): "Apollonius Dyscolus , a pioneer of western grammar", in Work in Progress of the
Dept. of Linguistics, U. of Edinburgh, No. 11.
Lakoff, G. (1972): "Linguistics and Natural Logic", in: Davidson & Harman (eds): Semantics and
Natural Language (Dordrecht: Reidel.)
Lakoff, R. (1968): Abstract Syntax and Latin Complementation (MIT Press.)
Matthias, (1887): "Zu alten Grammatikern", in: Fleckeisen's Jahrbucher, Suppl. XV, p. 591.
McCawley, James (1968): "Lexical insertion in a transformational grammar without deep structure"
in Papers from the 4th Regional Meeting of the Chicago Linguistic Society, pp. 71-80.
Postal, P. M. (1966): "On so-called pronouns in English", in: F.P. Dinneen, S.J. (ed.): Monograph
Series on Language and Linguistics No. 19 (Washington, D.C.: Georgetown University Press).
(1971): Cross-Over Phenomena (New York: Holt, Rinehart and Winston).
Priscian: Institutiones Grammaticae, ed. by Hertz (Grammatici Latini, ed. Keil, II-III). Leipzig, 1857-
80).
Robins, R. H. (1967): A Short History of Linguistics (London: Longman).
Ross, J.R. (1970): "On Declarative Sentences", in: R.A. Jacobs & P.S. Rosenbaum (eds.):Readings
in English Transformational Grammar (Ginn & Co.).
266 REFERENCES
Schneider, Otto (1843): "A.D. de synthesi et parathesi placita," In Zeitschrift fur die Altertumswis-
senschaft, pp. 641-54.
Scholia on Dionysius Thrax, ed. by A. Hilgard (Grammatici Graeci I, 3). (Leipzig, 1901).
Skrzeczka, R.F.L. (1853): Die Lehre des Apollonius Dyscolus von den Redetheilen. Program of the
Kneiphof Gymnasium (Konigsberg).
(1855-9): Die Lehre des Apollonius Dyscolus vom Verbum. 4 programs of the Kneiphof
Gymnasium (Konigsberg).
Smyth, H.W. (1956): Greek Grammar (Cambridge: Harvard.)
Steinthal, H. (1890-1): Geschichte der Sprachwissenschaft bei den Griechen und Romern. 2nd Edition.
(Berlin).
Thompson, Sandra Annear (1971): "The deep structure of relative clauses", in: Studies in linguistic
semantics ed. by Fillmore and Langendoen (New York).
Thorne, J.P. (1969): "English Imperative Sentences", Journal of Linguistics 2, pp. 69-78.
Tryphon: Fragments, collected by A. de Velsen. (Berlin, 1853.)
Versteegh, C.H.M. (1977): Greek Elements in Arabic Linguistic Thinking (Leiden: E.J. Brill.)
Von Arnim, J. (ed.) (1903-24): Stoicorum veterum fragmenta. 4 Vols. (Leipzig.)
Wackernagel, J. (1892): "Uber ein Gesetz der indogermanischen Wortstellung", Indogermanische
Forschungen 1:333-431.
INDEX
[This does not include the introduction, or the notes, except for names in brackets.]
A
ABSOLUTE (apolelumenos, apolutos): 26, 89, ACTOR, (INFLUENCER) (energon, diatitheis):
103, 110 29, 110, 143, 185; "The Actor versus the Acted-
ACCENT [ACCENTUATION, ACCENTED] upon" (energon [kai] energoumenos): 185;
(tasis, tonos): 52, 54, 89, 90, 91, 104, 105, Actor's relation (energetike): 186 [See also
108-118, 120, 122, 123-126, 128, 131, 133, AGENT, INFLUENCER, SUBJECT]
134, 136-139, 146, 147, 150, 167, 172, 182, ACUTE (ACCENT) (oxeia, oxunetai): 52, 90,
196, 199, 207, 210, 225-231, 238, 239, 242- 115, 245, 246, 252; Acute on the antepenult
247, 252, 257, 259, 261; (orthotonoumenos): (proparoxunetai): 225; Acute on the penult
196, 228; Accented like (homotonos): 252; Ac (paroxunetai): 225; Acute on the ultima (he epi
cent on the penult (en bareiai tasei): 259; telous oxeia): 245 [See also ACCENT, UL
(barunomenos): 199, 229; Accent(ed) on the ul TIMA, ANTEPENULT]
tima, Ultima or Final accent (oxeia, oxunetai, ADD, ADDED, ADDITION (prosginetai,
oxunomenos): 199, 228, 229, 243, 245, 258, parathesis): 235, 239, 241; (prostithetai): 32,
259; Draw back the Accent (barunei): 229; Re 33, 40, 41, 45-47, 56, 70, 76, 112, 232, 241
cessive (Accent) (bareia, barunomenos): 199, [See also INSERTION]
227, 243, 252 [See also RECESSIVE, RE ADDITIVE: [See under CONJUNCTION]
TRACTION, SHIFT] ADDRESS (proskaleitai, prosphoneitai): 170,
ACCEPTABLE EXPRESSION (phrasis 171, 194, 209; Addressee (prosphonoumenos):
euparadektos): 44, 65, 78, 130, 136 [See also 170, 172, 196 L 197
GRAMMATICAL] ADHERES (henotai): 238
ACCIDENTAL PROPERTIES OR FEATURES ADJECTIVE [onoma], epithetikon [onoma]),
(parepomena): 69, 96, 116, 170 [See also ADJECTIVAL (epithetikos): 29-31, 34, 39,45,
FEATURE, PROPERTY, QUALITY] 46, 51, 55, 61, 62, 65, 67, 69, 75, 89, 93, 112,
ACCUSATIVE (CASE) (aitiatike): 35, 38, 104, 150, 151, 160, 162-164, 251, 252; (trigenes
128, 159, 175, 184-186, 210-217, 219, 221, onoma): 252, 254; Possessive Adjectives
223, 234, 235, 237, 248, 249, 255, 256, 260; (ktitika): 132-134, 138, 140, 218; Three-
Accusative singular; 172; Two-accusative con termination Adjectives (trigeneia): 163, 167;
struction: 185, 186; Accusative with Infinitive: Two-termination Adjectives (koina): 150; Com
186 parison of Adjectives: 89
ACQUISITION (peripoiesis): 21.9 [See also ADJOINING, ADJUNCTION (parathesis): 45,
BENEFIT, DATIVE] 50, 53, 56, 73, 164, 227
ACT, ACTION (pragma): 177, 183 189, 217, ADJUNCTIVE (epizeuktikos): [See u n d e r
223, 235; Action (energeia, energetike diath CONJUNCTION]
esis): 22, 216; Actions (drasis): 212 [See also ADVERB, ADVERBIAL (epirrema, epir-
EVENT] rematikos): 21, 22, 28, 30, 39, 40, 53, 54, 72,
ACTANT (hupokeimenon): 129; (prosopon): 74, 97, 160, 164, 174, 176, 177, 180-183,
125, 141, 232 [See also NOUN (NP), PER 186-190, 228, 231, 243-264j Optative Adverb,
SON] Adverb of Wish (to tes euches e.): 188; Time/
ACTED-UPON, INFLUENCED (energoume- Temporal Adverb (chronikon e.): 160, 247
non, diatithemenon): 142, 143, 185 [See also ADVERSATIVENESS (enantiotes): 202
OBJECT] AEOLIC (Aiolikos): 52, 85, 95, 104, 112, 116,
ACTIVE (VOICE) (energetike): 126, 151, 165, 229
175-178, 186, 190, 208, 209, 211, 212, 219, AFFECTED (diatithemenos): 186; (paschein):
223, 243 [See also VOICE] 218 [See also ACTED-UPON, OBJECT]
ACTIVITY (energeia, energetike): 91, 142, 208, AFFIRMATION, AFFIRMATIVE(NESS), AF
209, 217, 218, 222 FIRMING (kataphasis, kataphasketai): 102,
103, 187, 188; Affirmative (diabebaiotikos):
236 [See also ASSEVERATIVE]
268 INDEX
BOEOTIAN (FIGURE) (Boiotios): 174 COMBINATION, COMBINE: 19, 27, 30, 53, 97,
BOOK TITLES: 55 161, 163, 243, 247 [See also COMPOSITION,
BOUNDARY (sunapheia): 238 [See also LINK] CONSTRUED, TAKE, CONJOINING,
(ROUGH) BREATHING (daseia): 53, 112, 236, STRUCTURE, PHRASE]
237 [See also 'H', ASPIRATION] COMMAND (prostaxis, enkeleusis, prostassei):
[BUTTMANN: 94, 112, 175, 186, 196] 66,189, 191, 192, 194, 196-198; Near-future
Command (hupoguos prostaxis): 191 [See also
c IMPERATIVE (MOOD)]
COMPARATIVE (sunkritika): 57, 178, 196
[CAMERER: 161]
COMPLEMENT (OF VERB): 110, 128, 159,
CASE (ptosis): 23, 40, 43, 48, 49, 82, 83, 88, 89, 176, 211, 222; (Infinitive) Complement
92,95,98, 101, 104, 108, 126, 135, 140, 141, (aparemphaton suntaxis): 79 [See also
156, 159 160, 163, 167-169, 171, 183, 216, NOUN(NP), OBLIQUE, PERSON, GOVERN,
218, 223, 230, 232, 234, 255, 260; Case- OBJECT, REQUIRE, TAKE]
(inflected) word (ptotikon): 31, 35, 40, 52, 60
COMPLETE, COMPLETENESS, COMPLE
[See also NOMINAL]; Caseless word (aptoton):
TION: Complete (Grammatical) Sentence (au-
249 [See also DECLINABLE, INDECLINA
toteles [logos], autoteleia): 19, 23, 25, 102,
BLE, ACCUSATIVE, DATIVE, GENITIVE,
156, 161, 180, 211, 223 [See also SENTENCE,
NOMINATIVE, OBLIQUE]
PREDICATE, UTTERANCE]
CATEGORY, CATEGORIAL, CATEGORIZA
COMPOSITION, COMPOUND (suntithetai,
TION: 49, 159, 163, 174, 177; (merismos):
sunkeitai, sunthesis): 22, 28, 54, 93, 110, 119,
181, 182 [See also PART-OF-SPEECH,
120, 139, 140, 143-146, 148-152, 155, 157,
FORM CLASS]
158, 181, 203, 225,_226, 230-235, 237-245,
CAUSAL (aitiologikos, aitiodes): 186, 187, 201, 248, 250, 258; (sunenomenos): 237-8; Com
202, 236, 237; (parasunaptikos, parasunem- pound [Reflexive] Pronoun (sunthetos): 117,
menon): 22, 249; Causal Conjunction: [See 119, 128, 140, 141, 143-146, 148, 155; Com
under CONJUNCTION] pound Sentence (sumpeplegmenos): 22, 80
CHANGE (heteroiosis): 262; (summetabainei): [See also COMBINATION, CONJOINING,
181; (pathos): 2 1 , 11_5, 252, 263; LINKING, UNITED]
(meterchetai): 135; (metaptosis): 51, 155 [See
CONCORD(ANCE) (akolouthia, to akolouthon):
also PHONOLOGICAL CHANGE, SHIFT,
59, 134, 159, 162 [See also AGREEMENT]
TRANSFORMATION]
CONDITIONAL (sunaptikos): 200, .201, 203,
CHARACTERISTIC, CHARACTERISTIC 226; Conditional Sentence (sunemmenos): 21,
FEATURE OR VALUE (idioma, idios, 22 [See also CONJUNCTION, SYNAPTIC,
idiotes): 50, 70, 81, 88-90, 97, 104, 105, 112, HYPOTHETICAL]
113, 134, 135, 139, 148, 151, 163, 170, 203,
CONFUSED, CONFUSION (sunkechutai, sun-
212, 225, 239, 245 [See also ESSENCE, FEA
chusis): 32, 33, 88 [See also SYNCRETISM,
TURE, UNIQUENESS, VALUE]
HOMOPHONY, AMBIGUITY]
CIRCUMFLEX (perispatai): 52, 115, 182, 207,
CONJOIN(ING), CONJOINED (sunartatai, sun-
225, 228, 247, 261 [See also (CONTRACT)
dedetai, sumpleketai/sumploke): 22, 23, 27,
VERBS, ACCENT]
28, 61, 80, 85, 106-109, 148, 169, 194 [See
CLASSIFY, CLASSIFICATION (epimerismos):
also LINK, CONJUNCTION, COORDINATE]
250; (merismos): 97, 197 [See also CATE
CONJUGATION: 168; (suzugia, suzugos): 39,91,
GORY, PARSING, FORM CLASS, PART-
101, 204, 206
OF-SPEECH]
CONJUNCTION, CONJUNCTIONAL, CON
CLAUSE (logos): 103, 107, 123, 181 [See also
JUNCTIVE (sundesmos): 22, 23, 28, 53, 62,
MAIN CLAUSE, SENTENCE]
81, 105-107, 123, 159-161, 169, 174, 181,
[CLEFT SENTENCE: 76, 83]
186, 187, 200-207, 226, 227, 235, 237, 243,
CLUSTER: 21, 56
246, 248, 249, 255, 256, 257; Causal Conjunc
COINCIDENCE (ek p a r e p o m e n o u ) : 32; tion (aitiologikos): 186, 205; Conditional
(sunemptosis, sunempiptei): 48, 52, 163-166, (Synaptic) Conjunction (sunaptikos): 114, 200,
168, 199, 255 [See also AMBIGUITY, 201; Coordinating Conjunction, Copulative Con
HOMOPHONY, HOMONYMY, NEU junction (athroistikos, sumplektikos): 80, 81,
TRALIZATION] 106, 226, 257; Disjunctive Conjunction
COLLECTIVE (athroistikos): 46 (diazeuktikos) 23, 200, 201; Expletive Con
COLLOCATION (suntaxis, suntassetai): 19, 187 junction (parapleromatikos): 20, 26, 201, 202,
COMANUS: 124 228; Final (Purposive) Conjunction (apoteles-
270 INDEX
176, 178, 181, 224, 225, 239, 245; (Distinguish 141, 143, 145, 156, 158, 167, 173, 178, 184,
ing) Features (parepomena): 94, 98, 154; 185, 211, 212, 214, 216-219, 222, 224, 225,
Semantic Feature (enklisis): 160 [See also 251, 252, 255, 260, 262, 263; Partitive Genitive
PROPERTY, MARK, QUALITY, SPECIAL, (genike epimerizomene): 41, 42, 43, 156;Pos
DISTINCTIVE] sessive Genitive (ktetike): 47, 59, 67, 132, 134,
FIGURE, FIGURATIVE(LY) (schema): 42, 49, 135, 136, 140, 141
52,86, 102, 164, 167, 174, 179; (hupailage): GOVERN(ING), GOVERNMENT: 157, 218;
164; Figurative (tropikos): 189 [See also AT- (aitios): 184; (paralambanetai meta): 180;
TIC, ENALLAGE, HYPALLAGE, HYPERBA- (suntassetai): 110, 111, 183; (charizetai): 83;
TON, ELLIPSIS, ANASTROPHE, TRANS (sunuparchon): 180; (pheretai epi): 78, 84,
FORMATION] 101, 108, 184, 214, 215, 222, 224, 250;
FINAL (PURPOSIVE) (apotelestikos): 201, 202, (apaitei): 83, 108 [See also CONSTRUED,
205 [See also CONJUNCTION] TAKE, REQUIRE, VERB]
FINITE FORMS (enklisis): 162 [See also MOOD, GRAMMAR (logos): 156; (suntaxis): 42
VERB] GRAMMATICAL,_ GRAMMATICALITY (katal-
FORCE (dunamis): 23, 96, 161, 187, 188, 200, lelos, katallelotes): 19, 27, 33, 37, 43, 58, 62,
201, 205, 226, 233, [See also MEANING, 65, 76, 82, 83, 84, 94, 101, 106, 134, 135, 143,
SENSE, SEMANTIC] 146, 155, 157, 160, 164, 166, 173, 175, 179,
FORM (phone): 48,49, 56, 79, 91, 92, 93, 95, 97, 184,214,226,234;(sustaton):205;Grammati-
98, 134, 149, 150, 163, 175, 202, 254, 255; cal correctness (akolouthia): 162; Grammatical
(graphe): 66, 85, 89; (ekphora): 39, 89, 137, error: 161, 165, 167, 169 [See also AGREE
203; (enklisis): 38, 163, 166, 177, 192, 206, MENT, ANALOGY, CORRECT, ERROR]
207, (thesis): 39; (klisis): 52, 59, 94, 112, 140,
189, 208, 233, 241; (prophora): 38, 165, 174, H
199, 211, 233, 254, 256; (schesis): 242;
(schema): 26, 42, 121, 132, 138, 141, 155, 'H' (daseia, dasunetai): 21, 44, 51-54, 99, 112,
158, 190, 198, 203, 204, 241, 243, 262; 236; (pneuma (dasu)): 245 [See also ASPI
(schematismos): 40, 159, 206, 230; (tupos): RATION, (ROUGH) BREATHING]
250; (charakter): 29, 255; Form-class: 151; HABRON: 58, 91, 99, 103, 109, 146, 147, 171
(merismos): 49, 50; Full (Underlying, Free, [HERODIAN: 195]
Base) Form (enteleia, enteles [phone holo- HETEROCLITES (heteroklitoi): 92
klera): 114-115, 121, 263, 264 [See also AF HOMER: 73, 86, 92, 132, 136, 138, 172, 180,
FIX, WORD, PHONOLOGY, SOUND, MOR 214, 261 [Quoted on almost every page];
PHEME, SPELLING, PARADIGM, INFLEC Homeric: 119, 131, 135, 136, 146, 173, 179,
TION, DERIVE] 256
FORMULA: 179 HOMONYMOUS, HOMONYMY (homonumia):
FRONTING: [See under TRANSPOSITION] 45, 69, 93, 132, 135, 163, 228, 236; (sunem-
FUNCTION: 139, 140, 141; (enklisis): 190 pesonta): 163 [See also AMBIGUITY, CON
FUTURE (TENSE) (mellon): 66, 160, 187, 200, FUSION, HOMOPHONOUS, SYNCRETISM]
202, 205-7, 223, 240; Future, Futurity HOMOPHONOUS, HOMOPHONY (homo-
(esomenos): 66, 116, 202, [See also POTEN phonia, -os): 36, 40, 48, 94, 163, 166, 167,
TIAL, TENSE] 170, 172, 173, 175, 178, 192-194, 197-199,
205, 206, 237; (sunempiptei, sunemptosis):
135, 166, 167, 172, 173, 178, 194, 197 [See
G
also COINCIDE, HOMONYMOUS, NEU
GENDER (genos): 23, 27, 31, 32, 33, 39, 49, TRALIZATION]
51-54, 56, 59, 93, 94, 98, 99, 116, 118, 133, HORTATORY: 196
140, 156-160, 162, 163, 164, 167, 172, 174, HYPALLAGE (hupailage): 164-168, 173, 179
175, 190, 207, 208, 237, 250, 254, 255; Com [See also FIGURE, ENALLAGE, SUBSTITU
mon gender (ta koina en genei): 94; One-gender TION, ANTIPTOSIS]
(monogenes): 254; Surface gender (charakter): HYPERBATON (huperbibasmos, huperbaton):
174; Three-gender (trigenes, trigeneia): 39, 113, 124, 186, 231, 233, 234, 245, 251 [See
163, 167 also ANASTROPHE]
GENERIC (genikos): 29, 38, 249; (aoristodes): HYPOSTASIS (MENTION): 31 [See also MEN
34, 64, 65, 101 [See also INDEFINITE] TION]
GENITIVE (CASE) (genike): 23, 35, 36, 41-43, HYPOTHESIS, HYPOTHETICAL (=Subjunc
47, 51, 59, 67, 70-72, 75, 77, 86, 92, 94, 98, tive) (distagmos, distaktikos): 200, 201 [See
102, 108, 116, 126, 127, 131-136, 138, 140, also CONDITIONAL, DOUBT, DUBITA-
TIVE, UNCERTAINTY, SUBJUNCTIVE]
INDEX 273
NON-FINITE (aparemphatos): 180 [See also OPTATIVE (euktike): 38, 160, 166, 180, 181,
INFINITIVE] 187-191, 206, 223; Optative Construction
NORMAL (kurios): 138 (euktike suntaxis): 188 [See also MOOD, IL-
NOUN (onoma): 22-32, 34, 37, 38, 40-50, 52, LOCUTIONARY FORCE, WISH, VERB]
54, 69, 74, 78, 88-90, 92-97, 100, 101, 158, OPTIONAL/OBLIGATORY (ou deontos/anan-
159, 167-169, 173, 178, 198, 202, 218, 228- kaios): 170
230, 233, 239, 243, 246, 249-251, 254, 257, ORDER, ORDERING (taxis): 23, 67, 111; (katas-
259, 260-262; (onomatikon): 40, 41; Common tasis): 113; (schema): 123 [See also POSI
Noun (onoma prosegorikon): 29, 36, 63, 64, TION, UNDERLYING (FORM), WORD OR
65, 69, 70, 76, 78, 89; Head Noun: 76, 133, DER]
134; Proper Nouns, Names: 29, 45, 58, 61, 63, ORDINAL NUMBERS (taxeos onomata): 147
64, 65, 68, 69, 71, 77, 78, 99, 100, 101, 141, [See also NUMBER]
149, 150, 170, 179; Subject Noun (onomatike ORTHOTONIC (PRONOUNS)(orthoto-
thesis): 232; Noun Phrase, NP (prosopon): 27, noumenos): 89 [See also ACCENT(ED), UL
56, 85, 89, 96, 108, 142; (hupokeimenon): TIMA, OXYTONE]
147 [See also NOMINAL, ENTITY, RE OXYTONE (oxutonos): 199, 245, 246, 252 [See
FERENT, ACTANT] also ACCENT(ED), ACCENTUATION, UL
NUMBER, NUMERICAL (arithmos): 29, 38,42, TIMA]
92, 103, 146-149, 152, 155, 158-160, 162,
163, 167, 171, 174-177, 181, 183, 197, 259; P
(Cardinal) Number (arithmos): 47, 158 [See
also DUAL, PLURAL, SINGULAR, QUAN PARACONDITIONAL: [See u n d e r CON
TITY] JUNCTION]
PARADIGM (parakeitai): 195; (ekphora): 209;
(paradeigma): 166; (prophora): 198 [See also
o INFLECTION, CONJUGATION, FORM]
OATH: 164; Oath Particle (omotikon): 53 PARAPHRASE (metaphrazetai) 244; (sum-
OBJECT (heteron prosopon): 124; (plagia phrazei): 46 [See also EQUIVALENT, SYN
ptosis): 211; (to pathos to anadechomenon): ONYMOUS]
212; (to diatithemenon, energoumenon): PARSING (merismos): 250 [See also CLASS,
110, 208; (hupokeimenon): 208, 219, 222; FORM CLASS, PART-OF-SPEECH]
Direct Object (aitiatike): 211; (to en metabasei PARTICIPANT: 100, 232 [See also PERSON,
prosopon): 175; Indirect Object (to kata doti- ACTANT]
ken): 179 [See also ACTED-UPON, COM PARTICIPLE,PARTICIPIAL (metoche): 26, 31,
PLEMENT, INFLUENCED, OBLIQUE] 61, 63-65, 67, 69, 70, 76, 77, 85, 97, 135, 163,
OBLIQUE (CASES) (plagiai [ptoseis]): 25, 26, 164, 176, 193, 198,210, 224, 242-245; Partic
47, 51, 71, 76, 82, 84, 93, 95, 99, 101, 105, ipial Construction (suntaxis metochike): 78, 79
110, 119, 125-127, 142-145, 163, 168, 176, PARTICLE (epiphthegma): 53; (morion): 49;
210, 211, 221-223, 234, 244, 249 [See also Emphatic (Particle) (parapleromatikos): 201,
COMPLEMENT, CASE, OBJECT, GENI 202 [See also ADVERB, CONJUNCTION,
TIVE, DATIVE, ACCUSATIVE] WORD]
OMISSION, OMIT (ekstenai): 35; (apobole): 51 PARTITIVE APPOSITION (epimerizomena ek
[See also DELETION, LOSS, ELLIPSIS, plethous): 86, 106; Partitive Genitive: 41, 86
CONTRACT, CRASIS] PART OF THE SENTENCE (meros tou logou):
'ON ACCENT (Peri Tonon): 112 28, 32, 247
VN ADVERBS' (Peri Epirrematon): 160, 180 PART-OF-SPEECH (lexis): 202; (meros (tou)
'ON ARTICLES' (Peri Arthron): 37_ logou): 22-25, 28, 30, 32, 37, 49, 50, 53, 56,
'ON BREATHINGS' (Peri Pneumaton):_236 81, 82, 89, 93, 97, 99, 102, 150, 158, 174, 180,
'ON CONJUNCTIONS' (Peri Sundesmon): 80, 201, 202, 218, 228, 239, 256 [See also WORD,
201, 236 FORM CLASS]
'ON GENDERS' (Peri Genon): 32, 33 PASSIVE, PASSIVITY, PASSIVIZ(ATION):
'ON FORMS'(Peri Schematon): 243 (pathos): 70, 126, 142, 144, 151, 178, 186,
'ON IMPERATIVES' (Peri Prostaktikon): 66 204,_207-210, 212,_217, 219, 221, 222;
VN SPELLING' (Peri Orthographias): 205 (pathetikon, pathetike): 126, 165, 175, 176,
'ON THE PARTICIPLE' (Peri Metoches): 26, 178, 186, 190, 208, 211, 212, 222 [See also
62, 63, 79, 224, 242, 243 _ EXPERIENCE, VOICE]
'ON PERSONS' (Peri Prosopon): 168 PAST (TENSE) (paroichemenos): 34, 66, 160,
'ON THE PRONOUN' (Peri Antonumias): 58 161, 181, 182, 187, 190-192, 202, 204-206,
'ON VERBS' (Peri Rematon): 67 223, 240, 241 [See also TIME, TENSE,
TENSE-ASPECT]
276 INDEX
PATRONYMIC (patronumikon): 57, 178 PHYSICAL (somatikos): 208, 212; Physical Object
PATTERN (suntaxis): 235, 236; (schesis): 243; (soma): 25, 28, 221; Physical Relation
(schematismos): 183, 251, 252; (akolouthia): (somatike diathesis): 212
49, 52; (analogos, analogia): 52 [See also RE PLACE (topos): 189, 247, 250, 254, 255, 256;
LATION, REGULARITY, STRUCTURE, (taxis): 227; Place From Which(ek topou): 254,
CONSTRUCTION] 260-263; Place Where, Place in Which (en to-
PENULT (pro telous, paredreuousa, para- poi): 254, 258-261, 263, 264; Place to Which
legousa): 195, 206, 207, 252, 257 [See also (eis topon): 254, 255, 257, 258, 261, 263, 264
SYLLABLE, ACCENT(ED), ACCENTUA [See also LOCAL, LOCATIVE, NOUN, POSI
TION, ULTIMA] TION]
PERFECT (Tense, Aspect) (parakeimenos): 161, PLEONASM, PLEONASTIC (pleonazei,
240, 255 pleonasmos): 19, 20, 43, 99, 102, 114, 115,
PERFECTIVE (i.e. Aorist) (anusis): 205; (sun- 185, 201, 202; (ho ton parallelon logos): 189;
telike, sunteleiosis): 66, 192, 205 [See also (parolke): 261 [See also INSERT, EXCESS,
TENSE, TENSE-ASPECT] _ REDUNDANCY]
PERSON, PERSONAL: (prosopon): 25, 26, 31, PLUPERFECT (hupersuntelikos): 210, 255
33, 34, 38, 47, 49, 50, 56, 57, 59, 88, 91, 99, [See also PERFECT, TENSE]
100, 101, 103, 104, 118, 122-4, 127-130, PLURAL, PLURALITY, PLURALIZATION
138-140, 144-156, 158-160, 162, 163, 166- (plethos, plethuntikos): 31, 34, 42, 46, 64, 71,
172, 174-177, 179, 181, 183, 190, 192, 194- 74, 103, 146-149, 155, 156, 167, 170, 171,
197, 208, 210, 213, 219, 222-224, 231, 232, 174, 193, 194, 197, 255; Pluralization
242,258, 259; First Person: 154, 155, 162, 166, (arithmos): 148, 174; (plethunetai): 38 [See
167, 172, 175, 189, 192-195, 205, 209, 223; also NUMBER, QUANTITY]
First Person Plural: 169, 171, 195, 199; First & POEM, POETIC(AL), POETRY: 19, 55, 66, 98,
Second Person: 154, 175; Second Person: 154, 99, 102, 119, 148, 179; Poetic Figure (schema):
155, 166, 168, 171, 172, 175, 180, 192-197, 52, 123, 167; Poetic License (adeia): 8,43, 102,
204, 205, 209, 223, 258, 259; Second Person 215 [See also HOMER, USE, FIGURE]
Plural, Second Person Dual: 170-172; Third Per POSITION (PLACEMENT) (thesis): 25, 79, 95,
son: 88, 90, 91, 100, 101, 104, 122, 124, 125, 111, 113, 161, 227; (suntaxis): 28, 96, 111,
126, 139, 146, 153-155, 160, 162, 168, 170, 228; (taxis): 28, 56, 79, 84, 111, 226; (topos):
171, 173, 179, 192, 194-197, 199, 205, 210, 28, 112, 116, 117 [See also WORD ORDER]
222, 223, 259; Third Person Plural: 204; In the POSSESSION, POSSESSIVE, POSSESSOR,
Same Person (en toi autoi prosopoi): 118, 119; POSSESSED (ktema, ktesis, ktetike): 67, 68,
Person (Participant, Actant): 232; Definite Per- 72, 74, 75, 77, 78, 126-136, 138, 140,_141,
sons (prosopa horismena): 25, 123; Double Per 171, 178, 218; Double Possession (disse ktesis):
son (disson prosopon): 196; Indefinite Person 59, 141; Unique Possession (monadike ktesis):
(aoriston prosopon): 123; One-person: 196; 34, 67, 74; Possessor (ktetor): 74_, 78, 127, 129,
Two-person: 196; Personal: 262 [See also 130, 133,142, 172; Possessed (ktema) 127, 129,
ENTITY, PARTICIPANT, ACTANT, PRO 133, 135, 140-142, 172 [See also GENITIVE,
NOUN] PRONOUN]
PHONEME (stoicheion): 19, 22, 56, 91, 232, [POSTAL,P.: 96, 143]
252 [See also LETTER, SOUND, ELEMENT, POSTPOSITIVE (hupotaktikos): 21, 56, 111,
ALPHABET] 227, 228, 232; Postpositive Article [= relative
PHONOLOGY, PHONOLOGICAL(LY): Phonol pronoun] (hupotaktikon arthron): 21, 51,
ogy (phone): 39, 54, 111, 112, 138, 229, 258, 79-86, 116, 173, 227, 246 [See also SUB-
264; Phonological: 89, 105, 183, 238, 247; fUNCTlVE, PREPOSITIVE, POSITION,
Phonological Change (pathos): 115, 178, 182, RELATIVE PRONOUN]
251; Phonological Form (phone): 49,_91, 98, POTENTIAL(ITY) (dunetikos, dunasthai): 161,
204; Phonological Rule (logos tes phones): 145; 200
Phonological shape (phone): 99, 104, 132, 166 PRAYER (euche): 188, 190, 191, 200 [See also
[See also FORM, MORPHEME] OPTATIVE, WISH]
PHRASE: 34, 140, 231, 261; (lexis): 227; PREDICATE, PREDICATION, PREDICATOR:
(parathesis): 110, 120, 144, 158, 203, 205, 60-2, 77; [Stoic terms] (sumbana/kategorema):
225, 230, 232, 233, 235, 237-240, 244, 247-9, " 37, 223; Predication (proscharizetai): 101 [See
252, 260; (parakeimenos): 238, 239, (ptosis): also VERB, SUB-PREDICATION, SUB-
58; (suntaxis): 41, 235, 244, 245, 250, 257 PSE UDO-PREDICATION]
[See also ADDITION, AD)UNCTION, CON PREFIX (prothesis): 20, 2 3 1 , 2 4 1 ;
STRUCTION, EXPRESSION, SYNTACTIC] (prosklinetai): 239
INDEX 277
PREMISE (sunaphe tou logou): 21 [See also 122-126, 130, 132-136, 140-143, 148, 149,
CONDITION, PROTASIS] 150, 154, 155, 170, 172, 179, 181, 196, 227,
PREPOSITION (prothesis): 20, 21, 24, 27, 28, 228, 230, 262; First Person Pronoun: 25, 56, 91,
53,70, 102, 110, 111, 113, 136, 137, 150, 159, 104, 128, 135, 139, 148, 149, 150, 152,
181, 211, 225-235, 237-239, 241-250, 257, 169,170,171; Second Person Pronoun: 56, 88,
258, 260-262;Prepositional (prothetikos): 245, 91, 104, 128, 135, 139, 148, 149, 150, 152,
250; [See also under ADDITION, ADJUNC 168-172; Third Person Pronoun: 89, 91-93, 95,
TION, CONSTRUCTION, COMPOSITION, 100, 115-117, 121, 124, 125, 127, 130, 139,
COMPOUND, PREFIX] 146-148, 152, 169, 170, 171, 173; Plural Pro
PREPOSITIVE (protaktikos, prothetikos): 56, noun: 112, 152, 171; Possessive Pronoun/ Adjec
114, 226, 231, 232; Prepositive Article (protak- tive (ktetike): 59, 74, 75, 124, 126, 127, 128,
tikon arthron): 21, 231 [See also ADDITION, 130, 131, 132, 134, 135, 136, 138, 140, 141,
CONSTRUCTION, ADJUNCTION, ARTI 146, 147, 152, 170-173, 197'; Reciprocal Pro
CLE, ARTICULAR, CONJUNCTION, OR noun: 47, 145; Reflexive Pronoun (antana-
DER, POSITION, POSTPOSITIVE] klomene): 142; (autopathes): 125, 143, 213;
PRESENT (TENSE) (enestos): 66, 160, 187, (sunthetos): 117, 119-125, 127-131, 139-148,
191, 202, 205, 207, 209, 240, 255; Present 155, 213 [See also ANAPHORA, ARTICLE,
Indicative: 199, 204; Present Progressive (enes COMPOUND, DEICTIC, DEMONSTRA
tos parateinomenos): 191, 192 [See also TIVE, EMPHASIS, CONTRAST, ENCLISIS,
TENSE, TENSE-ASPECT, IMPERFECT- PERSON, POSSESSION,_REFLEXIVE]
(IVE)] PRONUNCIATION (ekphonesis): 22, 25 [See
PRESUPPOSITION, PRESUPPOSE (oide): also PHONOLOGY, SONORITY, SOUND]
184; (paralambanetai): 42; (paremphainei, PROOF (apodeixis): 28, 102, 134, 144, 155, 240,
paremphasis): 36, 64, 84, 232,; (hupa- 248 [See also ARGUMENT]
goreuei): 48, 196; (proiiphestos): 36, 42 [See PROPER: [See NAME, NOUN] _
also UNDERSTOOD, MEAN, IMPLICIT] PROPERTY (parepomenon, idioma): 19, 21,
PRIMITIVE (prototupon): 178 [See also BASE, 35, 50, 81, 90, 170 [See also CHARACTERIS
FORM, UNDERLYING] TIC, FEATURE, DISTINCTION, VALUE]
[PRISCIAN: 28, 88, 100] PROSE-WRITERS (sungrapheis): 102
PROGRESSIVE (parateinei, paratasis): 191, PROTASIS (SUBORDINATE CLAUSE): 237
192, 206 [See also IMPERFECTIVE] [See also CONDITION, CLAUSE]
PROHIBITION, PROHIBITIVE (apagoreusis): PSEUDO-PREDICATE (parasumbama): 222
187,211 [See also STOICS, SUB-PSEUDO-
PRONOMINAL (antonumikos): 59, 69, 99, 244 PREDICATE, PREDICATE, PREDICATOR,
PRONOUN (antonumia, antonumikos): 22, 23, EVENT]
25-31, 51, 52, 55-9, 68, 69, 71, 74-7, 81, 88- ■PUNCTUATION (stigme): 106, 117
105, 107-155, 157-159, 162, 163, 168-172, PURPOSE(IVE) (apotelestikos): 201, 202, 205,
179, 189, 196, 213, 218, 226, 228, 244, 245, 248
262; Anaphoric Pronoun [ = Referential] (an
tonumia anaphorike): 27, 56, 66, 81, 90, 118,
170, 246; Appositive Pronoun [= Intensive Pro
Q
noun, q.v.] (a. epitagmatike): 154; Articular QUALITY, QUALITATIVE (poiotes, to poion):
Pronoun: 55, 57, 172; Compound (Reflexive) 29, 30, 50, 53, 60, 61, 62, 89, 93, 96, 101, 189,
Pronoun (sunthetos): 119, 120, 121, 128, 258; Individual Quality [=Stoic term] idia
140-142, 144-146, 14_8, 155; Contrastive Pro poiotes): 50, 101, 189; Secondary Quality
noun (prosdiastellouse ant.): 213; Deictic (De (sumparepomenon): 30 [See also ATTRI
monstrative) Pronoun (deiktike ant.): 56, 57, BUTES, FEATURE, CHARACTERISTIC,
59,75,81,84,90,91,116, 118, 170; Emphatic INHERENT, ESSENCE]
(Contrastive, Accented) Pronoun: 89, 90, 104, QUANTITY QUANTITATIVE: Quantity =
105, 108, 110, 111, 118, 119, 124, 125, 134, Amount (pelikotes, plethos): 29, 42, 96, 106;
137, 147, 228; Enclitic (Unemphatic) Pronoun: (to poson): 29; (chronos) = Length of Vowels:
89,90, 104, 110, 111, 115, 116, 124, 128, 133, 91, 245; Same Quantity (isochronei): 195 [See
227, 228; Ethnic Pronoun: 152, 153; Intensive also LENGTH, MORA]
Pronoun (epitagmatike): 59, 89, 117, 143, 154, QUESTION_(peusis): 28, 29, 68, 69, 168, 214,
170; Non-articular Pronoun (asunarthros): 31, 235; (erotesis): 103; (eperotesis): 168, 188
55, 91; (Simple) Personal Pronoun: 29, 56, 59, [See also INQUIRING, INTERROGATIVE]
81, 89, 90, 92-94, 99, 100, 101, 104, 117, 118,
278 INDEX
TEXT (anagnosis): 42, 172 [See also READING] UNDERSTANDING (sunienai): 220
THEORY, THEORETICAL (logos, ta tou UNDERSTOOD, UNDERSTAND (hupa-
logou): 42, 102, 107, 141, 171, 184, 212, 237; kouetai): 31, 179, 180, 184, 185, 234, 235,
Theoretical Defect (kakia): 119 250; (noeitai): 26, 47, 78, 108, 223, 248 [See
THESSALIAN'(DIALECT) (Thessalikos): 51, also IMPLICIT, DELETION, ELLIPSIS]
132, 168 UNGRAMMATICAL(ITY) (akatallelia, -os,
[THOMPSON, S.A.: 80] -otes): 21, 36, 37, 47, 51, 63, 79, 84, 101, 120,
[THORNE, J.P.: 168] 127, 131, 135, 138, 142, 154-158, 162, 163,
THOUGHT (dianoia): 114, 198, 210 165, 166, 167, 169, 174, 175, 180, 202, 205,
TIME (chronos): 60, 66, 160, 190, 235, 246, 207, 258; (alogos): 154; (hamartanei): 163;
249; (katastema): 221 [See also ASPECT, FU (aparadektos): 155, 218; (asunetos, asus-
TURE, PAST, TEMPORAL, TENSE, and tatos): 45, 72, 205; (kakia): 102, 162 [See also
under ADVERB] ANOMALOUS, GRAMMATICAL(ITY), SOL
TMESIS: 203,231 ECISM, BARBARISM, DEFECT(IVE), ER
TOTALITY (holon plethos): 44 ROR, AGREEMENT]
TRADITION (paradosis): 42-44, 144. 212, 264; UNINTELLIGIBLE (asunetos): 44
(chresis): 171 [See also USAGE] UNINTERPRETABLE (anephiktos): 62 [See
TRANSFORM, TRANSFORMATION (meta- also UNACCEPTABLE]
lambanetai, metalepsis): 95, 96, 128, 130, UNIQUENESS (idiotes): 65, 69
136, 152, 155, 163, 164, 176, 179, 184, 185, UNITED, UNION (henomenos): 169, 235, 239,
194, 204, 247; (metabatikos): 231; (metapip- 243, 247, 249 [See also COMPOUND]
tei, metaptosis): 26, 207; (metapoiei): 203, UNMARKED ([aparemphatos]): 178
205; (antimetalambanetai): 206; (aphistesi): USAGE, USE("'ordinary" or "everyday usage")
203 [See also CHANGE, DERIVE] (ethos, ethimos): 73, 135, 148, 167, 256; (the-
TRANSITIVE, TRANSITIVITY (metabasis, sis):_36, 42; (sunethes): 20, 73, 144, 215;
metabatikos): 35, 79, 125, 126, 159, 185, 212, (chresis): 51, 57, 104, 119, 122, 130, 141, 146,
214; (diabatikos, diabibazetai): 47, 125, 208, 155, 167, 170, 171, 172, 184, 193, 258; (koine
209, 213, 217 [See also VERB] phrasis): 102; Usage & Tradition (paradosis):
TRANSPOSED, TRANSPOSITION (anas- 212; (suntaxis): 180 [See also HOMERIC,
trephei): 74, 75, 111; (metatithetai, metathe POETICAL, EVERYDAY, DIALECT, TRADI
sis): 58, 112, 226, 227; (huperbibasmos): 233, TION]
234, 251; (hupertithetai): 22 [See also ME UTTERANCE (eiremenon): 19; (prophora): 38;
TATHESIS, ANASTROPHE, HYPER- (autoteleia): 156; Utterance (context, situation):
BATON] 34, 60 [See also SENTENCE, STRUCTURE]
TRYPHON: 29,37,49,50,61,76, 140, 145, 168,
227, 239 V
VALUE (idioma): 50, 104, 233 [See also CHAR
U ACTERISTIC, FEATURE, UNIQUE]
[UHLIG: 143, 161, 165, 168, 183] [VARRO: 159]
ULTIMA: 245 [See also ENDING, SYLLABLE, VERB, VERBAL (rema): 22-26, 28, 30, 37, 38,
ACCENT, PENULT, ANTEPENULT, 60-62, 65, 69, 70, 74-77, 79-81, 83-86, 88-91,
OXYTONE] 101-106, 110-112, 118, 119, 125-127, 141-
UNACCEPTABLE (anephiktos): 47; (asunetos): 146, 156, 158, 159, 162-167, 169, 173-185,
45; (aparadektos): 57 [See also UNINTER 189, 198, 203, 208, 209, 218, 221, 223, 228,
PRETABLE] 237-240, 242, 243, 245, 249, 258, 259; Con
UNAMBIGUOUS: 124, 198 [See also AMBIGU tract Verb (perispomenon rema): 85, 204-206,
ITY, AMBIGUOUS] Governing Verb (sunuparchon rema): 180;
UNASPIRATED (psilos): 54, 99, 112, 247 [See Implicit [Performative] Verb (enkeimenon
also ASPIRATION, 'H'] rema): 179; [Intransitive] Self-Sufficient Verb
UNCERTAINTY (distagmos): 200, 205 [See also (autoteles): 101; Simplex Verb (haplous): 241;
DOUBT, DUBITATIVE, HYPOTHETICAL, Transitive Verb: 185, 213; [Transitive] Deficient
SUBJUNCTIVE] Verb (remata elleiponta): 101; Verb "is": 68,
UNDERLYING: Underlying Form (to hexes, 70, 71; Verbs of Being (Existence) (huparxis):
logos): 74 , 75, 109, 137, 173; Underlying: 21, 60,61,63,74,75,76, 101,127, 142, 169; Verbs
86, 99, 113. 114, 121, 124, 137, 143 [See also of Naming-or-Being (Existence): 69, 76; Verb of
HYPERBATON, TRANSFORMATION, Wishing (proairetikon rema): 78; Verb Basel
TRANSPOSITION, ENALLAGE, SUB Root (auto to rema): 38; Verb Form (rematike
STITUTION, DELETION, ELLIPSIS] lexis): 125; Primary Verb Form (proteuon) 178:
INDEX 281
z
ZENODOTUS: 20, 43, 135, 136, 138, 172