Doctor Alec Marantz
Lecture Abstract
Structure and Memorization in Words: When to Look at the Brain
The linguistic account of language relies on a particular notion of the "morpheme" as the building block of language: Tokens of morphemes are identical to each other as far as the computational structure of language is concerned, not merely similar along dimensions of sound and meaning, and words that are morphologically related, such as "acidic" and "acidity," share (decompose into) identical morphemes (here the root "acid"). Competing connectionist (neobehaviorist) accounts treat the relevant notion of "identity" as similarity at the limit. MEG experiments supplement behavioral results and reveal the crucial distinction between similarity and identity in language, while supporting linguistic decompositional theories of word structure. Strictly behavioral experiments mask distinct stages of linguistic processing and fail to disentangle priming and competition effects in word recognition.
Neuromagnetic Evidence for the Timing of Lexical Activation: An MEG component seneitive to phonotactic probability but not to neighborhood density
Neural Mechanisms of Spoken Word Recognition
