Professional Documents
Culture Documents
Review of Rocher, La Theorie Des Voix Du Verbe Dans L'cole Pdnineenne
Review of Rocher, La Theorie Des Voix Du Verbe Dans L'cole Pdnineenne
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Springer is collaborating with JSTOR to digitize, preserve and extend access to Foundations of Language.
http://www.jstor.org
REVIEWS
in the dhnika's
("day's work")
13 and 14.
familiar with
a Root-list
(Dhdtupdtha-)
to which
he makes
constant
REVIEWS
593
REVIEWS
594
a new sub-rule
of often
artificial
mentaries
for understanding
the Paiiinean
of
I
translation
the Kdsikd.
instance, Ojihara-Renou's
would have liked to see a short table outlining the relative chronology of the
so that the linguist to whom the book is addressed will know
commentaries,
to set by the expositions
what degree of confidence
of the various authors.
respect,
method
it presents
than, for
accessible
way
The Durghatavrtti
Its preoccupation
ture ismore
same method
should
By the same token, the Mahdbhdsya
with a great degree of reliability, although
listed all the argumentations
pro and con
in the face of
shortcomings
REVIEWS
595
indicates,
deals with
of special verbs and their voice. There is a short index of the principle roots
discussed. Preceding.the discussion is a short chapter, "Preliminaires", that
summarizes the definitions of Middle, Active, Agent, Goal, and State. The
book is uniformly of high quality and a reviewer can do littlemore than to
a number
emphasize
of features
and that
is the notion
that a sentence
is an expression
of a verbal
action
too clear
of the book:
"La
contains
a number
of useful
diagrams
and
tables. On p. 29,
REVIEWS
596
'after a root'.
It must
be noted
that most
the
ablative suffix and can be interpreted as specifying the word dhdtoh, e.g.
jeh 'afterji-', kri.da 'afterkrid-'.Otherwise (if Imay indulge in the type of
reasoning displayed by theKdsikd) one may erroneously conclude that the
rule is valid every time ji- appears in the language, so also in dji- 'race'.
Hence we need the specificationjeh, 1 dhdtoh 'after root ji-'.Moreover, the
Kdsikd
supplies
the word
this anuvrtti,
intended
Nor is theword neh dittoed into rule 89. The Kdsika supplies it, and in fact
the author argues later (par. 33) that the anuvrtti should continue. One
wonders why it did get left out.
From the point of view of grammatical strategy, the diagram on page 38
is interesting. The author there presents a table showing the plan of the
chapter on Voice. We see that there are threegeneral conditions underwhich
a rootmay getMiddle endings, namely if it belongs to the columns anuddtta
nit, svaritanit, and a class of miscellaneous conditions (column 3). From
Liebich's investigations (Einf. III 1920) and my own counts, we learn that
the first column of Mme Rocher's diagram comprises a total of 477 roots,
49 of which
are marked
about
the circumflex
and 43 the symbol fn. It is obvious, therefore, that themajority, well over
1000 roots, would fall in the third,miscellaneous column, and themajority
REVIEWS
597
of them in the lower right hand cell of the diagram; these are the roots that
by default take active endings. It is clear that Panini's methodology of
description in principle resembles that of any modern, logical system of
description: he startswith the smaller units and finally comes to a point
where he can state explicitly, "in the remaining roots the active endings are
used" (78).But the superficialresemblance stops here, for the actual arrange
ment and ordering of the rules in the treatise is peculiar. They fall into the
following groups: 12, 15-29, 30-37, 38-55, 56, 57-61, 64, 65-67, 68-71, 72,
74-77, 78, 79-82, 83-85, 86-87, 91-93. We note that one of the general
statements (12) is not followed immediatelyby its exceptions, but that the
exceptions are given after all the remaining theory has been taught, rules
83-85, 91-93.We also note that the next general rule (72) is not surrounded
by its exceptions, but that they occur in 30-37, 56, and 64, and afterwards in
79-82. Similarly, rule 74 (in turn an exception to 78) is excepted by 68-71
and by 86-87; finally the third general rule (78) has numerous exceptions in
15-29, 38-55, 57-61, and 65-67. The ordering of the rules seems to fall into
a vaguely discernible pattern of concentric shells. In the outer shell is rule
12 followed by its exceptions (83-85, 91-93). In the next shell is 78 preceded
by its exceptions including 74, which in turn is preceded and followed by its
exceptions and by other exceptions in the outer shell. In the innermost shell
is 72 preceded by its exceptions and followed by a set of exceptions in the
second shell. In diagrammatical form we would get the following scheme
where + means the statement of the rule, - means its exception:
+12 [-78 {-74
(-72
-84
-12.
598
REVIEWS
and a verbal
forms with an augment but without a verbal suffix (avid-).The same rule
brings about both augments and so the rule is by the standardsof the theory
of meta-rules not constant. Hence the problem is not solved yet. The form
ni-a-vis- still does not present theCause necessary to restrict the endings to
theMiddle.
The author then continues to show how theMbhMapplies the same criteria
to another context, ni-vi&-L-suffixas opposed to ni-vis-sap-L-suffix. To
generate the augment, the L-suffix must be replaced by an historical tense
(lai, lun, of Irn).But to apply the restriction to theMiddle endings, all that
is necessary is to replaceL by these endings.Kaiyyata finds this a sufficient
reason to declare the ruleofMiddle ending restriction"nuclear" (antarafga-)
in respect to the augment rule, but nuclear only because it required fewer
causes. This criterion of nuclearity is not the normal one, andNdgesa indeed
The discussion then takes another turn
rejects it in his Paribhdaendusekhara.
and becomes concerned with the establishment of the correct generative
cycle fromRoot - La-suffix to Root - Verbal Suffix - Conjugational Suffix,
amatter which the author describeswith her usual thoroughness.The inter
REVIEWS
599
esting point that emerges from these discussions is that themeta-rules are
quoted by Patafijali beginningwith theweakest: firstparatva- 'posteriority',
then nityatva- 'constancy',next antarafigatva- 'nuclearity',and presumably,
if therewere scope for it, he would have used the Exception criterion.This
order of application seems to imply a principle of economy whereby one
should not use a meta-rule stronger than necessary to argue a point, or else
he would have startedwith the strongest criterion first.
The solution to the present problem is indeed ingenious as the author
remarks, and is the one presented by thePradipa (p. 142, par. 374). Itmakes
an appeal toMR 41, a meta-rule with an adhoc character, that states that a
particular form, the context Root-La, is first affected by theMiddle restric
tion rule (1.3.12) and only afterwards by the rule of verbal suffixes sap, etc.
Ndgesa has two other examples of its applicability, but yet it has an unsatis
factory quality of artificiality about it. The development of ni- visL-suffix
would
i-
lamidde
ni
- vi a lami
iddle
...>
*nyavisat.
rule.
REVIEWS
600
comes up on the
REVIEWS
601
B. A. VANNOOTEN
BIBLIOGRAPHY
Ojihara, Y. and Renou, L.: 1960, La KdSikdvrtti, Publications de l'Ecole Franmaise d'
Extreme-Orient, vol. 18, Paris.
Thieme, Paul: 1935, Panini and the Veda, Allahabad.