irregular plurals

Mon Feb 12 09:43:02 UTC 2001

Oscar Wilde once put it very felicitously: "I asked the question for the best reason possible, for the only reason, indeed, that excuses one for asking any question –-simple curiosity." 
I guess curiosity is always simple and never idle. It is _lack of curiosity_ that is idle.

Turning now from philosophy to linguistics :-), I think that the key word in Gilbert's original question, despite the emphasis on 'nouns', (Does anyone know of any language where the singular and plural of a significant number of NOUNS are formed from historically unrelated stems?) was the word 'significant'. What makes me think so is the fact that suppletion is a rather common phenomenon cross-linguistically and it should not be a problem to find a great deal of languages with quite a few examples.

I am not sure what Gilbert meant by "significant number", so I will interpret it my own way: "strikingly/unusually many". 

I believe one is unlikely to find a language with, say, some 200 or more suppletive plural stems. Seems to me it would be uneconomic for the speakers of the language to afford such a luxury merely for grammatical purposes. We do know of numerous cases of people actually speaking or understanding two or even more "parallel" languages. I am not talking of multilingualism. I mean, for example, the honorific terms that are so common in Japanese, Austronesian, and others. (I found one example from Samoan. The equivalents for 'speak' in it are: tautala (polite), felalai (orator), fofonga (chief), malele (high-rank chief). 

Other instances are male vs. female parlance (not uncommon in Amerindian) or secret languages (in Africa) or avoidance styles (in Australia). They may seem to be "uneconomic" for the speakers, but they fulfill a social function. In other words it makes sense and is worth while to invent/develop/ and learn new unrelated words in order to emphasise that you're a male/female or in order to show due respect to the chief, etc. And it does not seem to be worth any efforts to develop and learn new words for the sole purpose of marking plurality (purely grammatical function with no social implications).

Most Russian examples mentioned earlier are just cases of one form falling into misuse and some other filling the gap. That's a pretty normal way of developing suppletion, of course, but where NOUNs are concerned, Russian, I am sorry to say, does not have that much suppletion and we do retain most of the "lost" forms in non-neutral styles. We have sg. reb'onok – pl. reb'ata and we have sg. dit'a/dit'o – pl. deti. Both couples can be used for "child(ren)/kid(s)/etc." But it just happened so that the NEUTRAL pair would be sg. reb'onok  – pl. deti. Reb'ata being reserved mostly for "guys (with or without girls)". Dit'a sounds too literary, lofty and old-fashioned, while dit'o sounds too informal and even uneducated (and may in fact be a relatively recent back?formation from deti). But all the forms are still in use and are perfectly well understandable.

As for chelovek – l'udi (which seems, by the way, to be related to the German Leute), and other cases of suppletion, like

idu – shol                     go – went            vado – andavo (Italian.Very much the same inFrench, +  fut. 'ira')
xoroshiy – luchshe        good – better       buono – meglio 
ploxoy – xuzhe             bad – worse         male - peggiore 

and others, it looks to me very much like an areal feature diffused at some time among (Indo)European languages. That is why I think one shouldn’t expect to find some strikingly deviating cases in the said languages. 

Better candidates would be Amerindian, Australian, Caucasian, Papuan languages. Caucasian languages are ever so "bizarre" in many respects. Some of them have, for example, temporals  like 'in three/four/five (!) days' that are not derived from respective numerals. And there are of course lots of instances of suppletion. Given that many common words in Australian languages were used as names and got tabooed after the person died (and often in such cases equivalents from neighbouring languages were taken) suppletion may be widespread in Australian languages. I skimmed through "The Languages of Australia" (Dixon) but didn’t find anything to that effect, though. I just didn’t have much time for that.

Suppletion is rather common in Papuan languages. One example is Yimas (William Foley, The Yimas Language of New Guinea, Stanford UP, 1991: p.120ff):

kalakn – kumpwi                                    'son, child, boy'

away – ngawanyct                                'mother's brother'

narmang – ngaykum                                  'woman, wife'

kaywi – ngaykumpn kumpwi                        'daughter, girl'

tnum – tpwi (suppletive?)                'canonical sago palm'

But there are, of course, better experts in Papuan languages on this list who might say more.

>Russian has a real suppletive example: _celovek_ "person, human
>being" (_c_ is like unaspirated English _ch_) vs. _ljudi_ "people"
>(_j_ is palatal glide). The respective unsuppletive complementary
>forms do not exist in the modern language, as far as I know not even
>as archaisms.

It is not quite so. There is the singular form l'ud 'people, rabble, mob' (with the connotation of uneducated/lower-class/common people). It is a common word. Everybody will understand it. The plural of chelovek also exists absolutely legally – only in "oblique" cases, i.e. all cases except for the nominative: shesti chelovek – of six people, shesti chelovekam – to six people, shestyu chelovekami – with/by 6 persons, etc (I am sorry to wage on this discussion here, after all it's not a Slavonic list. I just mean to say that Russian is a bad candidature for a "suppletive language", not better than many other European languages, at least).  Moreover, even the nominative plural exists (being of course a recent coinage). It's quite normal to hear the following phrase anywhere (except in formal speech): Vse my l'udi, vse my cheloveki – All we (are) humans.

The situation with suppletive leta – "years" is not so simple either. First, it is not conditioned by combination with numerals. The lexeme sg. god – pl. gody simply has a defective paradigm in that the plural genitive is missing. Mnogie gody (nom./accusative) – mnogix let (genitive). In formal style only this would be possible. But, secondly, it is quite possible to say mnogix godov informally. That would sound like a countryside way of speaking, but it still is fairly common. 

The numerals ending in 2, 3, and 4 (2, 3, 4, 22, 23, 104, etc) take singular genitive nouns and the numerals ending in 5-0 take plural genitive nouns. Since the plural genitive for god "year" is suppletive, naturally those numerals combine with the suppletive form. Other instances of pl. gen. would be suppletive too: moix let – of-my of-years. 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/an-lang/attachments/20010212/2fcb31fa/attachment.htm>