Case revisted

Daniel Riaño danielrr at
Wed Nov 17 15:51:39 EST 1999

I wasn't planning to enter here, but since nobody answered... Besides, I
have a request to do concerning this issue.

Some time ago, Jay Adkins wrote (i.a):
>     c.   The newer 8 case system is based upon cases found in other
>primitive Indo-European languages, particularly Sanskrit.  "When Greek
>cases were studied in the light of these eight Sanskrit cases, it was
>discovered that the same general distinctions prevailed.  This sound method
>of comparative philology has brought the twentieth century Greek grammarian
>to recognize that there are eight cases in Greek instead of five.  In
>addition to the process of comparative investigation, this conclusion is
>also based upon the very obvious fact that case is a matter of function
>rather than form.... In the simplest typical sentence the noun is the
>subject, and therefore, in the nominative case.  It is absurd to think of
>turning this statement around, and saying that the noun is in the
>nominative case, and, therefore, the subject."  Dana & Mantey, page 65-67.
>     d.   Wallace says, "an appeal to an earlier usage.... may have little
>or no relevance to the present.... Further, the appeal to such older
>languages as Sanskrit is on the basis of forms, while the application to
>Greek is in terms of function.  We have few, if any, proto-Greek or early
>Greek remains that would suggest more than five forms.  Second, the "very
>obvious fact" that case is a matter of function rather than form is not as
>obvious to others as it is to eight-case proponents.  And it is not carried
>out far enough.  If case is truly a matter of function only, then there
>should be over one hundred cases in Greek.  The genitive alone has dozens
>of functions" (page 33-34).

	Dana & Mantey's view is wrong in many points, that I will not enter
here. Wallace's criticism is sound in some aspects, but equally wrong in
others, IMO. I think he is right in rejecting a synchronic analysis of data
based in diachronic categories alone. But on the other hand we have many
"remains" that not only suggest but prove the presence of many other
suffixes no longer productive. The problem is if they are to be considered
as belonging to the case system or not. And if we consider them a part of
the case system, what kind of case-system was that. Moreover, Wallace's use
of "function" is very idiosyncratic, methinks. What Wallace calls
"function" is more often and (again IMO) more properly called "uses" of a
case. Nobody, to the best of my knowledge, has ever dreamed to see "over
one hundred cases in Greek" prior to this author and, as a matter of fact,
no author, to the best of my knowledge, has recognised such a number of
"uses" of the cases, even when the number of "uses" of a case are always
greater than the number of its syntactic or semantic functions.
>     e.   Practicality speaking (IMHO), case is not primarily a matter of
>mere form or mere function.  Function can't be determined by form alone as
>our example above in 1. C. 1. I think proves that.  Yet, Wallace is
>absolutely correct when he suggests if case is a matter of function alone,
>8 is not enough.  .....
	Modern functional grammar uses "case" (with low case letters) to
design the morphological (and syntactical, in the structural sense of
"syntax") category, and CASE (with uppercase letters) to designate the
semantic functions. It is almost universally recognised that there is not a
one to one relation between case and CASE. It is really a pity that such a
confusion in terminology is being used, but it is usually clear that, e.g,
when you speak of the functions BENEFACTIVE, AGENT, PATIENT you are
speaking of CASE, and if you speak, I mean write, about the syntactic
function "direct objet", "agent", etc. you are talking about "case". The
article of Crespo mentioned in Wallace's bibliography (p.176) can be
illuminating in this aspect [besides: there are some other very interesting
articles about the accusative (specially) in the same volume where this
paper comes from]

>		ii.	Case is important, according to David Black says
>that, "About 60% of
>all words in the NT use case endings."  It's Still Greek to Me, page 45.
>(57.79% according to Wallace, page 31)

	Statistics vary if they express (a) the number of actual forms in
the context that do have a morphological mark of case or (b) whether a word
appearing in a given text has or not case mark. Black statistics refer to
the former, I think. My own data for other "registers" of Ancient Greek are
somehow different:

	Author		W case	W/o case mark
	Thucydides: 	58,20%	41,80%
	Xenophon: 	54,02%	45.98%
	Polybius		63,21	36.79%

[The data are based on a partial corpus (about 20000 words for every
author. Not every instance is checked yet). Fosilized forms of a case are
counted in column A: There is a variation of about 1,5% if you consider
such instances in column B.]

	Wallace makes some use of statistics from a real corpus, which is a
clear point to his favour in comparison with most of Greek grammars. One of
the most striking figures he uses (pp. 34-5) is about the uses of cases.
According to his data (I think that they come from the Accordance program)
the percentage of apparitions of a case is as follows:

	Nominative	31%
	Accusative		29%
	Genitive		25%
	Dative		15%
	Vocative		<1%

	This figures are very interesting indeed, specially the
predominance of the nominative (However, it is more usefull, to study the
percentages relative to the number of words with case ending, not with the
total of words). I have found no other corpus of Greek where nominative
exceed accusatives, and according to my own partially examined data, this
is not the case in Luke (where accusatives outnumber nominatives by more
then the 1%, i.e. about 2% of the words with case ending). I'd *really*
like to know if somebody is aware of another corpus of Greek literature
prior to the 6th century AD where nominatives are more abundant than
accusatives. And, since I don't have access to the morphologically tagged
NT module of Accordance (I found it an extraordinary program, but the
modules are expensive!) I'd like to ask any good Samaritan to tell me if
this predominance is general in the NT or only in some books. An
interesting review of Accordance 3.5, with special emphasis on the
reliability of the grammatical data can be read in Duhoux, Ives,
"Electronic Antiquity: Communicating the Classics" 5.1 (1999).

	A final note. There are *many* problems and dangers with
statistics, of course. There is nothing simple in counting linguistic
units. If you are studying case, different results are obtained if you
count every single instance of a word with any given case morpheme, or if
you count as one instance of nominative only the syntagm whose head is a
nominative (even when it can have an article and an adjective in
dependency). Depending of what kind of study you are doing you must use one
way of counting or the other. Another example may be "Lists". For my own
methodological purposes I define a "List" as a string of seven or more
words/syntagms coordinated in the same case form, or repeating the same
pattern (don't ask me why "seven"). An example of list may be the beginning
of Eu.Matt: It is useless , I think, counting this string of 14*3= 42
nominatives and accusatives to deduce the relative "weight" of the cases.

Daniel Riaño Rufilanchas
Madrid, España

Por favor, tomad nota de la nueva dirección de correo: danielrr at

More information about the B-Greek mailing list