Recall of Connotative Meaning

The term connotative meaning is here used to describe the type of response called for in the ratings of words on the GOOD-BAD scale as specified by Osgood's semantic differential (1952). This study is a brief report of an experiment in which the method of paired associates learning was employed in testing the following hypothesis: Subsequent to the learning of a giuen meaningful verbal response presented as a translation of a paralog, there should be a tendency for the connotatiue meaning acquired by the pamlog to be retained when the verbal response itself can no longer be recalled. The paralogs selected for the experiment were the following 18 two-syllable Turkish words: aslan, bardak, bayram, demet, filiz, ilim, hasret, kalem, lale, merkez, mzkil, ortak, sanut, texat, vatan, yalan, zarif, and xahmet. The responses learned to these paralogs were selected from the semantic profile atlas of Jenkins, Russell, and Suci ( 1957). These words were equated in so far as possible on the basis of the geometric means of the L and S counts of frequencies of usage in the Thorndike-Lorge tables ( 1944). Three groups of six words each were chosen so as to represent a sampling from connotatively good, neatral, and bad meanings according to the values given in the atlas. They were as follows: (a) good, with values of 1.40 to 1.67: faith, joy, justice, nice, success, sunlight; ( b ) neutral, with values of 4.00 to 4.30 : dim, heavy, knife, slow, stout, winter; (c ) bad, with values of 5.17 to 6.87 : devil, hate, hungry, mad, troable, steal. These words were presented as translations of the Turkish words which Ss were to learn. A different randomized assignment of the supposed translations to the paralogs was used for each of 25 Ss. The paired-associate learning was continued to the criterion of three correct recalls of all the supposed translations. The test for recall was given one week after the initial learning. The paralogs were presented one at a time in the exposure apparatus. Ss' responses to these stimuli were labelled correct when they were the originally learned translation words. All other responses and failures to respond were labelled as non-recalls. Regardless of how Ss responded, they were asked to rate the meaning of each paralog on the GOODBAD semantic differentia1 scale. In analyzing the data the semantic differential values ascribed to the paralogs were assumed to be a function of the translation words with which they were paired in the initial learning. On this basis the means of the