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

The role of co-occurrences on conceptual tokening: Accessing object concepts by their category labels

Creative Commons 'BY' version 4.0 license
Abstract

It is often said that we know the meaning of a word by the company it keeps. But what happens when the “company” is an object? How does the rate of co-occurrence between a word (bark) and a picture (dog) influence how fast the concept is accessed? We investigated this question by modeling data from three sources: (1) norming data obtained from a dataset of 78,000 features to a set of pictures; (2) accuracy and response time data from a picture-word masked congruency task with stimuli presented for 60 and 200 milliseconds; and (3) cosine-similarities between an object (dog) and lexical items labeling superordinate (animal), high-salient (bark), and low-salient (fur) properties of the referent object using word2vec. Results show a superordinate effect: words labeling object categories were most affected by cosine similarity. Our results suggest that the co-occurrence between an object and a superordinate label facilitate object concept tokening.

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