An operational definition of abstractness in nouns was constructed by using the human discriminative response to identify two points on a scale of abstractness. This scale, consisting of 490 'abstract' and 571 'concrete' nouns, was found to have adequate reliability. When the scale was manipulated as an independent variable, the effect of abstractness on short-term recognition memory was highly significant, 'abstract' nouns being less well remembered than 'concrete' nouns. Frequency was found to be pertinent variable, independent of abstractness, very frequent nouns being less well remembered than some-what rarer nouns.
{"source_snippet": "The purpose is to distinguish two classes of nouns: (a) \"concrete\"\u2014being those whose reference to objects, to material, to sources of sensation is relatively direct; (b) \"abstract\"\u2014being those whose reference to objects, to material, to sources of sensation is relatively indirect."}