JOURNAL OF VERBAL LEARNING AND VERBAL BEHAVIOR 9, 126-135 (1970) Meaningfulness and Trigram Recognition1

low, medium, and high meaningfulness (M) was studied in the Shepard-Teghtsoonian paradigm. Correct recognition varied irectly with M and inversely with number of interven-ing events. False recognition varied inversely with M and increased with total number of presentations. Confidence incorrect recognitions varied directly with M and inversely with number of intervening events, but remained stable over the experimental session. Confidence in"new'judgments of new trigrams declined sharply over the experimental session. Evidence is presented to the effect hat false recognition is largely item specific and not a matter of general decision criterion. Shepard (1963) has argued that a characteris-tic of verbal units low in meaningfulness (M) is that they are sequentially unintegrated. Of the two trigrams RZQ and MOP, the former may be viewed as more fractionable than the latter from the point of view of Ss who are faced either with reproducing it in recall or with using it as a stimulus in paired-associate learning. Melton's (1963) intra-unit interference in-terpretation of forgetting in the short-term memory situation is consistent with Shepard's interpretation of M. A low-M unit is seen to be received in memory as a set of more or less independent elements that in recall must be assembled in the presence of interelement interference. A high-M unit, however, is seen to be received as approximately a single element without internal interference. This interpretation is supported by Murdock's (1961) finding that short-term recall of wor