Grant: Discovery Grants Program - Individual

The aim of my research program is to develop an integrated theory of semantic and episodic memory by articulating how semantic knowledge (and other kinds of linguistic knowledge) play out in episodic memory performance. In my recent work, I derived representations of semantic meaning in numerical form using distributional semantic modelling and imported these representations into an episodic memory model that simulates encoding, retrieval, and recognition decisions (MINERVA 2; Hintzman, 1986). The combined model, MINERVA S (S for "semantics"), provides a theoretical framework for the integration of semantic and episodic memory and explains how people remember the meanings of words, sentences, and even metaphorical expressions. I will extend my theory by developing an account of how perceptual features of words interact with semantic knowledge and episodic memory processes. Semantic and perceptual features interact in interesting ways to influence memory. For instance, after studying the words "bug", "gnat", and "fly", one may resist misrecognizing the semantically related word "butterfly" because it is a longer word compared to the studied words (a perceptual characteristic). To model the perceptual representations of words, I will draw on phonological and orthographic models from the field of computational linguistics. My vision is to leverage these techniques to create a complete model of memory that explains the interaction of semantic and perceptual features and episodic memory processes. I will also explore the interaction of semantic and episodic memory under conditions of intentional forgetting. The cognitive mechanism of intentional forgetting remains elusive, but a promising recent theory suggests that it involves the de-coupling of the item from its context, resulting in an inability to retrieve the item from the lack of contextual cues. I will conduct a coordinated experimental and computational analysis of this theory. I will conduct experiments in which the learning contexts of words are varied, and I will have participants make source judgments about these contexts. This will help tease apart memory for item information and context information. I will then use MINERVA S to computationally examine the relative contributions of semantic similarity and contextual cues to memory judgments. This two-pronged approach will illuminate the interactions between semantics, context, and episodic processes in dynamic circumstances where people try to remember as well as forget studied items. By developing an integrated theory of semantic and episodic memory, I will advance knowledge within the Natural Sciences on the fundamental cognitive processes of language and memory. This research will benefit Cognitive Scientists by providing a framework that brings together theory and data on semantic and episodic memory, research areas that have become increasingly disconnected despite the clear interconnectedness of these two memory systems.