Four Yale investigators have been awarded a National Science Foundation (NSF) grant to use behavioral, brain scanning, and information theoretic methods to investigate how and why meaning changes in language.
The recipients of the four-year grant are Maria Piñango, linguistics and the Yale Interdepartmental Neuroscience Program (INP); Ashwini Deo, linguistics; Mokshay Madiman, statistics and probability; and Todd Constable, Yale Magnetic Resonance Research Center and INP. This project is funded under the NSF's new Interdisciplinary Research and Education (INSPIRE) program, established to support the application of innovative multidisciplinary research to complex scientific problems. The researchers will focus on those patterns of linguistic change that have been observed in many languages and try to identify the cognitive and neural mechanisms that underlie such patterns.
"One striking fact about human languages is that, regardless of their genetic affiliation, they exhibit surprisingly similar ‘trajectories’ of change over time," notes Deo. "For example, over a long period of time, if a language will innovate a new marker for expressing possession, it will likely recruit a location marker like ‘near’ or ‘beside,’ even when there are other sorts of markers available in the language."
Piñango points out that such trajectories are unidirectional (that is, linguistic forms that convey location meanings acquire possession meanings over time, never the reverse). It is this “unidirectionality” and the kinds of semantic fields within which the changes operate that has led the group to hypothesize that they are rooted in the larger non-linguistic neurocognitive system, outside of the linguistic system itself.
The project will explore three languages — English, Hindi, and Spanish — to understand how expressiveness pressures deriving from the cognitive system shape the changes in the linguistic system and the ways in which these changes will occur. The other element of the project will model, using information theoretic and game theoretic methods, how and why a given change, once introduced, gets disseminated through a community of speakers.
The proposed multilevel approach rests on the interaction of three fields: linguistics, cognitive neuroscience, and statistics/information theory. The expectation, note the researchers, is that this approach will provide a cleaner view of the functional pathways that organize the infrastructure of the human thought system (addressing specifically the functional organization of the prefrontal cortex). Thus, they note, this research will use patterns of change as a probe into the larger cognitive organization of specific cortical subregions.