Hi, I'm Aaron, and I like ducklings!

I propose a theory of semantic change as concept development instantiated in a computational model.

I work with a computational model of how people learn and reason called DORA - Development Of Relations by Analogy (Doumas, Hummel, & Sandhofer, 2008). This model is a symbolic-connectionist model - it implements symbols through distributed connectionist representations, supporting the strengths of both symbolism and connectionism without resorting to a hybrid model design.

DORA extends Hummel & Holyoak's (1997) model of analogical reasoning LISA (Learning and Inference with Schemas and Analogies) through a developmental account for the representations both models depend on.

Another model in the same design lineage is BART (Bayesian Analogy with Relational Transformations; Lu, Chen, & Holyoak, 2012). Within our academic family we name our models after cartoon characters.

Why DORA? It's neurally plausible with strong correlates to key recent neuroimaging findings that fell out of computational constraints of the model (e.g., Knowlton, et al., 2012).

DORA accounts for 35+ findings surrounding the development of relational reasoning and 50+ findings from the adult analogy-making literature.

DORA is built on a small set of assumptions (three or four, depending on how you count). We're going to focus on the one that this theory hinges on - comparison-based learning.

This is built on simple Hebbian learning; what fires together wires together. This mechanism is exploited to discover the overlaps between things by firing the distributed representations for both simultaneously and learning the differential activation strengths across both. The resulting overlap is fired as the representation of the commonalities. This process supports concept discovery and refinement through the same mechanism!

A distributed representation is represented by a bunch of nodes (e.g., neurons) firing at the same time. The pattern of firing carries the information.

The representation for darker is highlighted here.

The representation for Kamil is highlighted here.

The representation for paler is highlighted here.

The representation for Alex is highlighted here.

Not that Alex - this Alex.

The representation for pale is highlighted here.

And the representation for me.

This comparison-based learning mechanism allows context-specific semantic information to wash out of the representations over time. As these representations become increasingly refined they apply to more situations because they carry less semantic stuff.

We posit that lexical entries are attached to these conceptual representations akin to semantic features (i.e., in the same way, using the same learning mechanisms). This isn't necessarily an isomorphic mapping - in fact, the example I used contains lexical items that map to more general concepts.

Consequently, as the concepts become more refined they apply to more situations, and the lexical items that are tied to them get to apply more broadly.

The same processes support narrowing as well - if a concept (and its corresponding lexical item) occur with a specific object a ton then the semantic features of that object will become increasingly associated with the concept.