The University of Texas at Austin

abstract

The "Frame/Content" theory is a Neo-Darwinian account of the evolution of speech by means of descent with modification from pre-speech capacities. According to this theory, the Frame provided by mandibular oscillation (paired with phonation) results in the mouth close-open alternation associated with the syllable. This capacity may have evolved first, and then the capacity to program Frames, with internal segmental "Content" elements may have evolved later. Beyond mandibular oscillation itself, the presence of three patterns common to infants’ first words, words of modern languages, and a proposed proto-language corpus is taken as evidence of the structure of original word forms. (1) The tendency of labial consonants to co-occur with central vowels (e.g. "ba") is taken as evidence of an evolutionary role of "Pure Frames" produced by mandibular oscillation alone; (2) The tendency for coronal consonants to co-occur with front vowels (e.g. "di") and dorsal consonants to co-occur with back vowels (e.g. "go") is evidence for a lingual inertial effect produced by a static non-resting tongue position throughout an utterance. (3) The favoring of a labial consonant - vowel - coronal consonant (LC) sequence in first words of infants (e.g. "bado" for "bottle") is considered to be the result of a self organizational increase in serial output complexity made both by earlier hominids and infants in response to the demand to increase the size of the lexical message set. The first two patterns are considered characteristic of the initial "Frame" stage and the third pattern may initiate/have initiated the Frame/Content stage in infants and in early hominids.

Generative Phonology is suggested by many to provide an alternative approach to the evolution of speech. However, there is no evidence for the most basic claim from this perspective - namely that a genetic specification for the phonological component of speech evolved de novo as part of universal grammar. In addition, no evidence is available from the biological sciences to suggest that the genes contain sufficient information to play an underlying role in the production of skilled actions (such as speech, which occurs at the rate of 15 segments per second). As in other essentialist conceptions, no time domain is specified (either historically or in contemporary acquisition of speech production) by generative phonology. Consequently generative phonology has no ability to cope with the evident fact that speech has evolved from simpler initial forms to present forms historically, and develops from simpler forms to more complex forms in infants. In addition, there is presently nothing approaching a comprehensive set of abstract distinctive features - the basic elements of generative phonology - though it is often asserted that the set of features is finite. There is good reason to be skeptical that cross-language continua in values of even the most common phonetic parameters such as those associated with consonant place of articulation and the vowel space could ever be susceptible to discrete parcelling into features. In addition while the core of phonology is considered to be abstract, the currently favored set of distinctive features is defined in concrete (articulatory) terms, which results in difficulties in characterizing aspects of sound patterns that are perceptually motivated. Moreover, to the extent that such features are considered to "account for" or "explain" the existence of speech phenomena, they presumably do so in terms of phonetic substance, and therefore cannot be taken as evidence of an independent abstract level of phonological form with explanatory value. The concept of markedness, considered to give a unitary "explanation" of the relative prominence of certain sounds and patterns, is defined in terms of a set of properties that are not always correlated (frequency of occurrence, complexity, time of acquisition by infants). Explanatory claims are based on circular reasoning; e.g. frequency reflects markedness, therefore markedness explains frequency. One specific problem for a supposedly unitary concept of markedness is that universal and therefore unmarked phenomena are sometimes different in infants and languages; e.g. a predominance of inter-syllabic repetition in infants versus a relative prohibition of inter-syllabic repetition in languages (the Obligatory Contour Principle). Specific differences between the Neo-Darwinian and Generative Phonological approaches will be illustrated by comparing their treatments of the basic consonant-vowel alternation of speech, the three intra-syllabic CV co-occurrence patterns and the labial-coronal sequence.

 Conference site: http://www.infres.enst.fr/confs/evolang/