A Unified Model Of The Origins Of Phonemically Coded Syllable Systems

Abstract

A Unified Model Of The Origins Of Phonemically Coded Syllable Systems Pierre—yves Oudeyer Sony Computer Science Lab, Paris e—mail : py@csl.sony.fr Abstract Human sound systems are invariably phonemically coded, which means that there are parts of syl- lables that are re-used in other syllables. It is one of the most primitive compositional system in language. To explain this phenomenon, there ex- isted so far three kinds of approaches : “Chom- skyan”/cognitive innatism, morpho-perceptual in- natism and the more recent approach of “language as a complex cultural system which adapts under the pressure of efficient communication”. We pro- posed in (Oudeyer 2002) a new hypothesis based on a low-level model of sensory-motor interactions, characterized by the absence of functional pressure and the use of very generic neural devices. This paper presents a unified model of the origins of syl- lable systems which does allow a comparison of the different hypothesis on the same ground. We show that our hypothesis is the only one to be suflicient, and that all others are not necesary. Moreover, the model we present the first that shows how a popula- tion of agents can build culturally a complex sound systems without the assumption that they already share a phonemic repertoire. What does explain phonemically coded syllable systems ‘.7 Human sound systems have very particular proper- ties. Perhaps the most basic is that they are phone- mically coded. This means that syllables are com- posed of re-usable parts. These are called phonemes. Thus, syllables of a language may look rather like la, li, na, ni, bla, bli, etc than like la, ze, fri, won, etc This might seem unavoidable for us who have a phonetic writing alphabet, but in fact our vocal tract allows to produce syllable systems in which each syl- lable is holistically coded and has no parts which is also used in another syllable. Yet, as opposed to writing systems for which there exists both “pho- netic” coding and holistic/pictographic coding (for e.g. Chinese), all human languages are invariably phonemically coded. The question is then : Why is this so ? How did it appear ? What are the genetic, glosso- genetic / cultural, and ontogenetic components of this formation process ? These questions are of particu- lar interest and generality since phonemic coding is a form of primitive compositionality. Compositional- ity is thought to be the keystone of syntax, and thus understanding how it appeared might help a lot to understand syntactic languages which make humans unique. Several approaches have already been pro- posed in the literature. The first one, known as the “post—structuralist” Chomskian view, defends the idea that our genome contains some sort of program which is supposed to grow a language specific neural device (the so—called Language Acquisition Device) which knows a priori all the algebraic structures of language. This con- cerns all aspects of language, ranging from syntax to phonetics (Chomsky and Halle, 1968). In particular this neural device is supposed to know that sylla- bles are composed of phonemes which are made up by the combination of a few binary features like the nasality or the roundedness. Learning a particular language only amounts to the tuning of a few param- eters like the on or off state of these features. It is important to note that in this approach, the innate knowledge is completely cognitive, and no reference to morpho-perceptual properties of the human artic- ulatory and perceptual apparatuses appears. This view is becoming more and more incompatible with neuro-biological findings (which have basically failed to find a LAD), and genetics / embryology which tend to show that the genome can not contain specific and detailed information for the growth of so complex neural devices. Another approach is that of “morpho-perceptual” innatists. They argue (Stevens 1972) that the prop- erties of human articulatory and perceptual sys- tems explain totally the properties of sound systems. More precisely, their theory relies on the fact that the mapping between the articulatory space and the acoustic and then perceptual spaces is highly non- linear : there are a number of “plateaus” separated by sharp boundaries. Each plateau is supposed to naturally define a category. Hence in this view, phonemic coding and phoneme inventories are di- rect consequences of the physical properties of the body. Yet, it seems that there are flaws to this view : first of all, it gives a poor account of the great diversity that characterize human languages. All humans have approximately the same articu-