Epistemic Transfer Between Linguistics and Neuroscience: Problems and Prospects

Abstract

This chapter analyzes instances of successful, partial, and failed unidirectional epistemic transfer between theoretical linguistics and neuroscience. I distinguish three types of transfer, depending on the nature of the linguistic knowledge involved: type-A knowledge, about language as such, essentially invariant across theories or formalisms; type-B knowledge, about alternative formal analyses of basic structures and operations in language; type-C knowledge, about the application of various computational methods to analyzing or modeling behavioral or neural data. I conclude that successful epistemic transfer may be achieved, under certain conditions, with type-A and type-C knowledge, and I present some examples of the strengths and limitations of each approach. Type-B transfer, however, in particular from theories of syntactic and semantic composition, so far has not led to new knowledge of the neural correlates and mechanisms of linguistic computation. I suggest that greater theoretical emphasis on algorithmic-level analyses, via a revised notion of linguistic competence and a new model of epistemic transfer, can bring formal linguistics and neuroscience closer together. Finally, I discuss the possible role of a computationalist psycholinguistics as a semi-autonomous 'bridge science', that could serve the aim of linking linguistics and neuroscience through algorithmic models of linguistic functions, aided by emerging methods and results in areas of computational linguistics, computer science, and AI.

Biolinguistics and biological systems: a complex systems analysis of language

“…2 Thus, research in neurolinguistics neither needs to include nor presuppose work in theoretical linguistics. Baggio (2020) explicitly considers the role linguistic theory can and does play in generating new knowledge the neuroscience of language or what he calls 'unidirectional epistemic transfer'. After surveying a number of candidate options such as constructions, parameters, syntactic and semantic composition, he concludes that despite convergence between linguistics and neuroscience being consider a desirable goal, "achieving it has proved exceedingly hard.…”

Three conceptual clarifications about syntax and the brain

“…Rather, it indicates that the underlying neural populations are responsible for the compositional combinations of specific lexical items (instead of abstract variables), and that these compositions are constrained by the syntactic properties of those lexical items. We speculate that such context dependence might be the reason that it has proven difficult to isolate syntactic combinatorics in neural datathat is, because syntax is to be found in the constraints on the combinatorial operations, not in the operations themselves (see also Pylkkänen, 2019;Baggio, 2020).…”

Cleaning up the Brickyard: How Theory and Methodology Shape Experiments in Cognitive Neuroscience of Language

“…The more agreement exists on a given construct or operation among different theoretical accounts, the more likely it is that the transfer from linguistic theory to neuroscience will be successful (Baggio, 2020). Nonetheless, the choice between the formal descriptions of language provided in the field of theoretical linguistics constitutes an additional auxiliary theory in itself.…”

“…We propose that the lack of agreement on the strict definition of "language" is also partially rooted in the fact that the translation from linguistic theory into observational terms through empirical hypotheses is not always straightforward. It is by now a widely known and established fact that linguistic theories and processing theories do not directly map onto neurobiological processes (for detailed discussions, see Baggio, 2020;Embick & Poeppel, 2015;Martorell, 2018;Poeppel & Embick, 2013). Using Marr's classic definition of the tripartite view of levels of analysis (Marr, 2010), the link between the computational, algorithmic, and implementational levels in our field has not been clearly defined yet (Figure 1).…”