The design requirements for language
Each of the world's 7,000 languages is a natural experiment in how to design a language, which are now being extended with synthetically evolved languages. These languages show that there are many possible configurations of the fundamental building blocks of language: languages convey different information in different ways. For example, many Papuan languages have complex grammatical systems requiring speakers to track the source of information through verbal affixes like 'evidentials' – 'saw with own eyes', 'hearsay', 'inferred' etc., while other languages (like English) leave this to the speaker's discretion.
The Centre will uncover the links between linguistic features within this design space, vastly increasing our understanding of how different systems in each language interact with each other. At the microlevel we will chart how new structures and systems arise within communities, or experimentally, while at the macrolevel we will analyse how these systems and combinations are co-evolving over time.
These approaches provide us with the exciting ability to identify frequently evolved combinations of configurations within design spaces ('attractor basins') – and to find those configurations that are improbable or impossible. We aim to understand the extent to which language reflects:
- the existence of distinct, but equally good, solutions to a single design problem
- the limits of evolution as a mechanism for building optimal systems
- the existence of distinct design problems in different environments
- different trade-offs or compromises in response to competing selective pressures.
These insights will not only help us understand language, but will be used to design effective processes for evolving human-like communication with robots and other complex technologies in collaboration with the Processing and Learning programs.