Skip to main content
eScholarship
Open Access Publications from the University of California

UC Santa Cruz

UC Santa Cruz Electronic Theses and Dissertations bannerUC Santa Cruz

The Features We Use and the Features We Lose: Encoding, Maintenance, and Retrieval

Creative Commons 'BY' version 4.0 license
Abstract

This dissertation explores the role of memory in language processing, and specifically how interference during lexical encoding can result in downstream interference during retrieval. The dissertation merges insights from both the sentence processing literature as well as the study of memory in non-sentential contexts and focuses on two factors which have been shown to contribute to interference: semantic similarity and proximity during encoding.

Experiments 1-2 in Chapter 3 examine interference on the basis of semantic similarity between a subject and intervener at a later retrieval site that initially does not provide any semantically selective information. Differential reading times prior to semantically selective information provide evidence for interference not just during retrieval, but during encoding. Experiments 3-5 in Chapter 4 examine proximity during encoding, and demonstrate in both recall and recognition that linguistic boundaries (e.g. clause boundaries) can serve to delineate encoding contexts in memory: reactivation of one word prior to recall boosts recall of surrounding words only within a clause, and sensitivity to changes in noun order is greater if a change occurs across a clause boundary. Experiments 6-9 in Chapter 5 explore lexical encoding in filler-gap dependency processing. Experiment 6 investigates the filled gap effect as a moment of overlapping activation between the filler and the encountered word, which may increase the chances of encoding interference. Experiments 7-9 examine whether a highlighted thematic role in the first clause of a biclausal sentence may influence the position of predictively postulated gaps. This is discussed in terms of the initial featural representation of a wh-filler word.

The dissertation concludes by connecting the themes discussed throughout the work to other areas of interest, such as predictive processing more generally.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View