LINGUIST List 5.525

Sat 07 May 1994

Sum: Origin of Case systems

Editor for this issue: <>


Directory

  • "William J. Griffiths", Sum: Origin of Case systems

    Message 1: Sum: Origin of Case systems

    Date: Fri, 06 May 94 18:30:21 CDSum: Origin of Case systems
    From: "William J. Griffiths" <WJGRIFFUKANVM.bitnet>
    Subject: Sum: Origin of Case systems




    In what seems like eons ago--enough time has elapsed for a language to go from analytic to synthetic and back to analytic again--I posted a query about the origin of case systems and the shift from analytic to synthetic. I apologize for the delay in posting this summary and I would like to thank the following who responded to my query:

    Anthony Rodrigues Aristar <aristartamuts.tamu.edu> Lunk Lagerwerf <lagerwefkub.nl> Fritz Heberlein <F.HeberleinKU-EICHSTAETT.D400.DE> Janne Bondi Johannessen <jannebjhedda.uio.no> Martin Haspelmath <martinhafub46.zedat.fu-berlin.de> Sheila <EMBLETONVM1.YorkU.CA> Geoffrey S. Nathan <ga3662SIUCVMB.SIU.EDU> John E. Koontz <koontzalpha.bldr.nist.gov> Harold Schiffman <haroldfsu.washington.edu> Scott C DeLancey <delanceydarkwing.uoregon.edu> Steven Schaufele <fcoswsux1.cso.uiuc.edu> Lynne Hewitt <hewittcs.Buffalo.EDU> Ian MacKay <IMACKAYACADVM1.UOTTAWA.CA> Gregory V. Gouzev <gouzevgvsun.mcs.clarkson.edu> STEVE SEEGMILLER <SEEGMILLERapollo.montclair.edu> Randy LaPolla <HSLAPOLLAccvax.sinica.edu.tw>

    In response to that part of my query dealing with shifts from analytic -> analytic and vice versa, it was brought to my attention that the issue of grammatical shift had already been addressed on LINGUIST (LINGUIST List: Vol-4-256 Sum: Grammar Shifts). On the origin of cases, it seems clear that and there is the most evidence for case systems arising from adpositions which become grammaticized--and to a lesser extent, adverbs and independent words. In the excerpts and the list of recommended works that follow, the origin of case endings in Turkish, Finno- Ugric languages, Proto Indo-European, Manda, Dravidian languages, Common Slavic, and Sino-Tibetan languages is discussed.

    =================================================================

    Anthony Rodrigues Aristar <aristartamuts.tamu.edu> writes:

    " . . There's a great deal which most linguists believe, but which no-one can--or has taken the trouble to--prove. One example: case-affixes arise when adpositions cliticize to their arguments, and these clitics through time become interpreted as bound. I can easily give examples of this occurring--the Turkish affix -le (roughly "with") can still appear as a postposition. But I know of no serious study that tries to understand this process well."

    Martin Haspelmath <martinhafub46.zedat.fu-berlin.de> writes:

    ". . The best introduction to grammaticalization (using case marking as prime example) is Christian Lehmann's paper in Lingua e Stile 1985. You really need the theoretical background and the comparative evidence to understand the grammaticalization of case markers. The best special study is still Joan Caspar Kahr's 1976 paper in the Stanford Working Papers on Language Universals, where she cites lots of examples of case markers arising from adpositions. The best-attested system is probably the rich Hungarian case system, which was only incipient in Old Hungarian. However, I know of no attested case where a whole language turns from soltating to (case-)inflecting."

    Geoffrey S. Nathan <ga3662SIUCVMB.SIU.EDU> writes:

    "Gordon Fairbanks, an Indo-Europeanist whom I studied with in the seventies, believed that PIE had only four original cases, and that the other ones (ablative, locative, instrumental etc.) were later developments. Unfortunately he only wrote one paper on the subject: Case Inflections in Indo-European. Journal of Indo-European Studies. 5.1-31. I found his arguments quite persuasive. He argued that case inflections arose from reanalysis of postpositions and postposed adverbs, at least in Sanskrit and a few other languages."

    John E. Koontz <koontzalpha.bldr.nist.gov> writes:

    "- Generally oblique case markers seem to originate from postpositions or (in the case of locatives) from old noun+poss_case+local_noun, i.e., in-house might be from house's-interior or, less abstractly, house's-chest. I think postpositions are essentially what are proposed for the origin of local cases in PIE, while in the Eskimo languages the local cases are transparently constructed from the ergative/possessive case + particle. Naturally, as a system of this or these sorts is morphologized, it is possible to get different systems prevailing in singular and plural, especially if the number category is itself (re)morphologizing at the same time, which accounts for the lack of transparent relationship between singular and non-singular case forms in, e.g., PIE, as well as syncretization of case-number forms. - The PIE nominative/accusative marking scheme, particularly the masculine and neuter genders of forms like o-stems, are considered by some to result from reanalysis of an original ergative/absolutive system. In general, the forms of case marking with ergative and accusative cases in ergative/absolutive and nominative/accusative systems can result from reanalysis of old obliques as these two types of systems are interconverted. For example, as if a nominative/accusative system develops by generalization of the antipassive construction in an ergative/absolutive system, the "new" accusative may be the "old" dative, even if the old dative has been replaced by a new one."

    Harold Schiffman <haroldfsu.washington.edu> writes:

    "I have spent a lot of time thinking about this issue with regard to the Dravidian languages, which are traditionally analyzed as having 7 cases, but most people admit that the 7 case framework was borrowed from Sanskrit along with its tradition of grammatical analysis. In the earliest grammar of Tamil, the Tolkaappiyam, the grammarian expressed his discomfort with the 7 cases and showed that Tamil either had fewer cases, more cases, or needed a different configuration altogether. Many of the original case suffices were free forms, but by the modern period have become bound, and the line between bound case markers and unbound (free form) postpositions, derived from nouns, verbs, or whatever, is in the modern language very fluid. If I take a very traditional approach to case I would have to say that Tamil, e.g. either has 20-some cases, or very few, maybe 4 cases but there are many problems and dilemmas with this analysis. I tried to write this up some years ago, and am stalemated; maybe I'll get back to this. What I would say is that there is a core of a case system, then some case-like morphemes that are free forms in the literary lg. but bound in the spoken lg. and then there are lots of postpositions that sometimes occur WITH case markers, sometimes WITHOUT, raising the total into the dozens. There are also lots of syntactic issues, e.g. if you use the instrumental with certain modals it implies being able but not nec. willing; using the dative with same modal means able and willing; this is a feature of spoken Tamil but not allowed in literary,etc. Anyway I think the system was once much smaller and more analytic, and has become more synthetic, and continues to get more synthetic all the time."

    Scott C DeLancey <delanceydarkwing.uoregon.edu> writes:

    "Maybe the Munda languages--within Austroasiatic, Mon-Khmer languages are isolating, Munda agglutinative. Patricia Donegan and David Stampe say something about this in a paper in the CLS Interplay volume (1983). Somewhere or other Sapir argues that Athabaskan is another example."

    Steven Schaufele fcoswsux1.cso.uiuc.edu writes:

    "There has been some speculation (going back to the 19th century) that the agglutinative case-markers in the Dravidian languages developed out of what were originally independent lexemes, in many cases not merely adpositions but full-fledged nouns with real-world referents."

    Lynne Hewitt <hewittcs.Buffalo.EDU> writes:

    "Joan Bybee has for a number of years been working on a project in quantitative cross-linguistics called the GRAMCATS project, whose purpose is to determine the lexical sources of grammatical material, using a stratified random sample of the world's languages."

    Gregory V. Gouzev <gouzevgvsun.mcs.clarkson.edu> writes:

    "Cases emerged when there was a need to mark off the different roles that words play in different narrative situations. The most common roles are subject, object, instrument, location, recipient, source, beneficiary, and so on. Not all roles have corresponding cases. Originally, the first cases to appear were the Nominative and Accusative when the subject and the object were separated in human's mind (NB! Indo-European languages only). Gradually Genitive, Dative and Locative developed over the centuries. The roles conveyed by these cases are so in demand that the expressors, originally being other words, ossified in endings, stress shifts, etc. thus forming cases. In fact, this is described in the literature pretty well, and I don't think you'll find many different points of view. As of a shift from analytic to synthetic, there is some weak evidence. First, when Chirch Slavonic from tri-case system (Nom., Gen.,(?) Dat.) developed into six-case system, which is well documented. Second, even today more roles occasionally get different forms, thought to be new case forms. E.g. in Russian there is second Genitive and Locative case forms, physically different from the canonical ones. Also, some people outline Inclusive case when Accusative plural takes the position of Nominative plural. Lastly, with numerals 1-4 same words must have different forms, e.g. 4 oficerA but 5 oficerOV (bad transliteration). This all broadens the family of the case forms, making some think that the shift from analytic to synthetic structure takes place. In fact, the other trend can be demonstrated as well, and it is not clear which is more powerful in the long run."

    Steve Seegmiller <SEEGMILLERapollo.montclair.edu> writes:

    "The Turkic languages contain some pretty clear cases of the evolution from (relatively) analytic to agglutinative structure. Turkish, for example, has six cases, and some or all of the suffixes clearly come from what used to be particles, prepositions, or things of that sort. It is just a step from agglutintive structure to synthetic, although I don't know offhand of any attested cases."

    Randy LaPolla <HSLAPOLLAccvax.sinica.edu.tw> writes:

    "I have been working on the origin of morphology in Sino-Tibetan, and so far have worked on verb agreement and two types of case marker (agentive and anti-agentive). On the latter, the origin can usually be traced back to a body part or other locational noun, with the former deriving in general from an ablative (which itself may derive from a locative plus directional verb) or a genitive plus ablative or locative marker. No form is reconstructable to Proto-Sino-Tibetan or even Proto-Tibeto-Burman."

    RECOMMENDED LITERATURE

    Anttila, Raimo & Eeva Uotila, (1984) "Finnish OVELA 'sly, cunning' and the Baltic & Finnic outer local cases". Ural-Altaische Jahrbuecher 56, 121-128.

    Antilla, Reimo. (1989) Historical and Comparative Linguistics. Amsterdam/Pjhiladelphia: John Benjamins Publishing co.

    Fairbanks, Gordon. (/) "Case Inflections in Indo-European." Journal of Indo-European Studies. 5. 1-31

    Heine, B., U. Claudi and F. H nnemeyer. 1991. Grammaticalization. A conceptual framework. The University of Chicago Press, Chicago.

    Hopper, P.J. and E.C. Traugott. 1993. Grammaticalization. Cambrudge Textbooks in Linguistics. Cambridge.

    Kahr, Joan. (1976) "The renewal of case morphology: Sources and constraints" Stanford Working Papers on Language Universals_ 20:107-151

    Lehmann, Chr. Latin case relations in typological perspective. in Touratier, Chr. (ed.), Syntaxe et Latin (actes du IIeme congres international de linguistique latine) Aix-en-Provence 1985, 81- 104.

    Odo Leys. (1993) "Reflections on the German Case System." in LEUVENSE BIJDRAGEN, volume 82.3, pp. 305-328.