Description:

The Surprisingly Lexical Behavior of Lexical Episodes

For years, it has been known that human memory has massive capacity to retain detailed traces of visual objects (e.g., Shepard, 1967; Standing et al., 1970), findings that have recently been extended (Brady et al., 2008). Such results led to the development of powerful exemplar models in the visual domain (e.g., Shi, Griffiths, Feldman & Sanborn, 2010). Similar models have been suggested in the auditory domain (e.g., Goldinger, 1998; Johnson, 2007; Walsh et al., 2010), proposing that lexical access can often proceed by memory-based classification, rather than segment-based signal analysis. Although exemplar models are powerful, they are often considered poor candidates in our quest to understand language use, due to a combination of implausible assumptions and hypothetical dissociations between neural systems for episodic memory and language.