Rules play a central role in human behavior. Social rules (e.g., do unto others as you would have them do unto you; don't judge a book by its cover) support the harmonious coexistence of people in a society. Rules of individual choice-making (e.g., exercise daily, beware of strangers bearing gifts) support the well-being of individuals. Although we are usually not consciously aware of using them, there are many rule-like behaviors associated with our use of language (e.g. add -s to form the plural of a noun). An important property of linguistic rules, like social rules and individual rules, is that they admit exceptions (e.g., not all nouns form a plural with -s). Knowing when to apply the rules and when to make exceptions is an important facet of being an effective member of a language community. This project investigates neural models of the formation of linguistic rules to understand how humans achieve an appropriate balance between rigid rule-following and flexibility. It explores human learning of simple invented languages in laboratory experiments and compares the results with mathematical models that approximate important features of neural systems. The goals are to identify two extremes---situations where people adopt rigid, universal rules and situations where they treat every instance as a special case---and then to examine how a balance between these can be achieved. By studying the neural underpinnings of such learning processes, we can gain insight into what kinds of human-environment relationships sustain appropriately systematic but flexible behavior.
This project has implications for the science of learning and for understanding the neural connectivity patterns that underlie complex thinking. Regarding learning, people sometimes fail to discover systematic principles, e.g. learning the spelling-sound correspondences that support reading. This project may shed light on why this happens. Regarding societal well-being, the tendency of people to form stereotypes appears to be a case of learning a rule "too well," that is failing to strike the right balance between generalization and exception making. By examining what brain/environment circumstances lead to such inaccurate learning, the research may help societies avoid the formation of rigid stereotypes.
