Consonant-vowel cohesiveness in speech production as revealed by initial and final consonant exchanges

Two experiments use a procedure developed by Carter and Bradshaw (1984) to examine the role of syllable structure in speech production. In the procedure, subjects exchange phonological segments in corresponding positions of a pair of visuallypresented words or nonwords and to produce the resulting words or nonwords as quickly as possible. Carter and Bradshaw had shown that the pattern of latencies mirror that of frequencies of exchange errors in natural speech. The first experiment of the present study shows that initial consonant exchanges are promoted by phonetic similarity of the exchanging consonants and they reflect a bias for producing real words. With these influences controlled, Experiment 2 replicates and extends the finding of Carter and Bradshaw that initial consonant exchanges are made more rapidly than final consonant exchanges. The discussion relates the latency difference between these conditions to a difference in the "cohesiveness" of initial and final consonants with their vowel. In particular, in Experiment 2, more than one-third of errors made on final-consonant or vowel exchanges are exchanges of the whole syllable rhyme (VC), whereas just 10% of errors made on initial consonant or vowel exchanges are exchanges of the initial CV of the word. Various explanations for the difference in cohesiveness are examined in post hoc analyses.