Skip to main content

Advancing Behavioral and Cognitive Understanding of Speech

Final Report Summary - ABACUS (Advancing Behavioral and Cognitive Understanding of Speech)

The aim of the ABACUS project was to Advance the Behavioral And Cognitive Understanding of Speech. The project investigated speech from an evolutionary point of view and more precisely combinatorial speech: the way that human speech uses a limited set of building blocks (among others, syllables, vowels and consonants) to produce an unlimited number of utterances. The project investigated how this works, how this is used, how it evolved in biology and how it evolves culturally when language is transmitted from generation to generation. The idea behind this is that the ability to deal with combinatorial speech is a uniquely human trait (at least compared to other primates) and an essential one for language. Moreover, it is a somewhat neglected area: traditional linguists and cognitive scientists focus on discrete aspects of language, such as syntax, vocabulary learning or processing of written language. In order to do this, the ABACUS project used a mix of experimental research and computational modeling.
The project's main achievements are threefold: experimental achievements, computational achievements and methodological achievements.
Through experiments it has clarified how the structure of communicative signals depends on the nature of those signals and on the relation between the signals and the meanings they express. This helps to explain the difference between spoken and signed languages. It has also made clear that evolutionary adaptations to speech are subtle and hard to detect experimentally. Finally, it has shown that there are deep relations between animal vocalizations and spoken language.
Through computer modeling the ABACUS project has helped quantify the amount of combinatorial structure in speech, it has helped to show what mechanisms are needed to acquire speech (and that these mechanisms may at first sight appear unrelated to language) and computer models have also helped to show what kind of adaptations to speech and language can evolve and why they would be hard to detect experimentally.
Methodologically, the project has advanced the methods for studying evolution in a laboratory setting, for studying signals in different modalities (sound, vision and touch) and for combining animal data with human experiments. It has also advanced the way in which different disciplines (linguistics, cognitive science, artificial intelligence and biology) can interact, and also how computer models can interact with experimental work in order to check models, analyze data and determine what precise experiment would be most useful to do at each stage of the research.
The ABACUS project has achieved a considerable advance of our understanding of speech-related cognition and speech-related behavior. It has also considerably advanced the state of the art of our methods for conducting research on these topics. The project has therefore achieved the high gain that it aimed for.