Key factors in the emergence of combinatorial structure: An experimental and computational approach

University Of Pennsylvania, Philadelphia PA

Investigators

Abstract

Speech sounds are fundamental building blocks of spoken languages, but languages differ greatly in both the identity and number of speech sounds they include. For example, although languages with five vowels are particularly common, some languages have as few as two vowels and others have over ten. The sound characteristics of the vowels of a language can be compared, as can aspects of how they are produced (e.g., the position of the tongue and lips needed to produce the vowel). It turns out that for both the resulting sound space and production space, there is a strong tendency for the vowels to be much more dispersed and symmetrically arranged than would be expected by chance. That is, they do not tend to cluster together, but are spread out relatively evenly. This pattern occurs regardless of the number of vowels in the language. How is it that the sound systems of different languages can be so varied, yet pattern themselves in similar ways? The answer may lie in the needs of human communication. When people listen to others speak, they need to be able to distinguish one word from another, often in noisy environments. The speakers, in turn, need to be able to produce sounds without too much effort that listeners can distinguish reliably. It may be that dispersion and symmetry emerge naturally as a by-product of human attempts to meet these communicative needs. In testing this hypothesis, the investigators may shed light on how languages can be so varied while remaining fundamentally similar in other ways. The project will also offer research opportunities to university and high school students who will be directly involved in the research, broadening their understanding of how language works and how to study it. This work tests the hypothesis that the dispersion and symmetry observed in vowel systems are due to the needs of communication in general, rather than anything specific to language as such. To this end, the investigators explore whether similar effects occur when people communicate through a medium other than language. The project involves an innovative experimental approach in which pairs of people communicate using specially designed software that allows them to send each other series of colors by moving their fingers around on a trackpad. While this medium is clearly very different from ordinary linguistic communication, it is designed to mimic speech in key ways: It involves sending observable signals by performing a physical action in a constrained space (comparable with making different speech sounds by varying the position and shape of the tongue and lips). If the same kind of organizational properties (i.e., dispersion and symmetry) arise in this color communication medium in response to the kind of communicative pressures described above, then this will strongly suggest that those pressures help shape the organization of speech sounds in language. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.