Prosody-pragmatics rich language- the case of Piraha~?

jess tauber phonosemantics at
Thu May 25 07:08:47 UTC 2006

Dan Everett's recent challenges to orthodoxy regarding the interpretation of grammatico-lexical oddities in Piraha~ has had me thinking quite a lot about the language's place in typology. A recent posting in LanguageLog brought this to mind once again (the post links to this paper: emphasis on prosody, singing, humming, whistling, its de-emphasis on segments (having the lowest phoneme count of any language on earth, with huge variation of what there is), focus on the here-and-now aspects of reality, etc., lack of many grammatical distinctions, all seem to point the way towards a very different language type than those usually encountered.

Some of the grammatical findings- no numbers, no number contrasts on lexical items. Individuable versus mass contrast exists- but not actual count. No quantifiers, but there are forms for 'whole' or 'part', which are not used in the abstract. Simplest pronoun inventory in the world, forms may be borrowings- and a much reduced set of roles for them in any case. No nominal morphology. No embedding. Verbal complexity high. Only two tense-like morphemes (remote versus proximate), no perfect.

Lexical- no color terms, only 90 verb roots in entire language (less than Kalam/Kobon family!), very few words for time at all. Perhaps the simplest kinship system known (reduced gender distinctions), and only to known relatives.

Cultural/discourse- no indigenous creation myths or fiction- everything is and always was like it is hear and now, no individual or collective memory more than two generations long. No art (at least drawing for decoration), but stick figures made representing spirit world directly experienced here, now.

Everett's treatment appears aimed at challenging assumptions (especially Chomskyan) about what needs to be considered necessary for language, and he posits cultural reasons to explain the above facts.

However, I have found little or no discussion of typology or typological change by the author. I cannot believe that the language has always been as it is today (despite the probability that speakers themselves might believe it). Thus it must have arrived in this state from some other. The question that I'm concerned with here is whether such change is somehow unusual in in type, or just in its extremity.

As many of you are aware (though getting constructive criticism seems tantamount to extracting blood from stone- what, worried the men in black will come for you??), my later work on correlating language type with ideophone numbers leads to some interesting conclusions with regard to cyclic change and where iconicity versus symbolicity is expressed in the entire linguistic system.

We are all aware of grammaticalization out of the lexicon, the processes evolving grams undergo, their functions, etc. I've hypothesized that ideophones and other more pragmatically oriented forms constitute a kind of 'antigrammar', which stands in diametric opposition to closed-class grammatical morphology on many fronts. It is becoming increasingly apparent (among those linguists who actually care to look further than purely left-hemisphere effects, thank you very much) that ideophone roots can, with the same sort of attrition and selection we see in the evolution of lexical items canalized towards morphology, move towards lexical status.

We thus have three stations in a chain relation. Ideophones to lexemes to grams (likely multiple hops are not much sanctioned- reduced numbers limit such  in any case). Grams then wear away to Cheshire Cat smiles (but not without first affecting prosody which has effects on syntax). Straight interjections may be the sources of many ideophone roots. There is implicit in all of this a re-prioritization of prosody versus segmentality in a Yin/Yang mandala-style cycle. Where languages are 'at' in this never-ending game is strongly linked to language type, as many people are slowly learning (though I doubt the interjection/ideophone connection has been made by them).

OK, then, what does this have to do with Piraha~? It is obviously very heavy on the prosody, light on segmentality. And although Everett hasn't looked I'll bet a lot of the variation of the existing segments is pragmatically oriented (and not merely automatic, though it might be hard to tell if both go hand in hand). You see such variation in ideophones in many languages- in Africa, in Korean, etc., but there is much more of it in interjections in many other languages- one of the reasons they are often so hard to phonetically render in a consistent and satisfying way.

If one thinks of the type cycle as a circle, and arbitrarily label the antigram-heavy state as zero degrees, then lexeme heavy types are at 90, gram-heavy types at 180, and syntax-heavy types at 270. Segmentalism is (roughly) more pronounced in the upper half of the figure (0 to 180), while prosody is in the lower half (180 to 360/0). It may not be as simple as this, but this is I feel a good first approximation.

Piraha~ would be a 'syntax' heavy sort of language- where all sorts of nonsegmental phenomena take center stage and pragmatics is king. I believe there to be a good possible link, in general, between prosody and pragmatics, just as there may be between segmentality and grammar (at least the morphological kind). Certainly studies of the brain demonstrate that segments prefer the left (home of the lexicon and morphology) and nonsegments prefer the right (home of much that is pragmatic, interjections, and possibly also ideophones, though nobody has looked for the latter relation).

Cyclic changes in language type would correlate with changes in how speakers view and interact with their world. Strong Whorfian argument (though I doubt he would have agreed with my particular take on the matter). Also changing might be WHICH hemisphere is in charge of language.

As hinted at above, things might not be so simple or cut-and-dried. For instance the DEGREE to which languages change may depend on other factors, such as how fast or slow the change happens (thus either dragging the past with it and muddling the synchronic picture, or allowing slow 'recrystallization' to use geological/chemical metaphor that allows one to purify the system, and work out the bugs). According to Everett, Piraha~ speakers are for the most part monolingual in the language and have been for hundreds of years. Time enough?

Anyway, it is my hope that my above tirade will generate some interest at the least, and discussion (but I won't bet that the heavy-hitters will want to dirty their fingers here...). I feel that Dan Everett, if his observations hold up, has found some VERY interesting data that pertain far more to the diachrony of language, typology, grammaticalization (versus antigrammaticalization!), etc. than they do to beating particular (un)dead Cambridge horses (even if such continue to rise every night to suck the blood of young and unsuspecting linguists). A horse is a horse, of course- eh Count Ed? BTHUH!

Jess Tauber

More information about the Funknet mailing list