This study compares telephone and face-to-face conversations in English and Japanese for the purpose of exploring what unit types constitute turns and at which places they occur.
Sacks, Schegloff and Jefferson (1974) state that various unit-types construct a turn and that usable turn constructional units have the feature of projectability that predict transition relevance places, or points where transfer of speakership is likely to take place. Ford and Thompson (1996) find that approximately three-fourths of turns take place at CTRPs, or Complex Transition Relevance Places, where grammatical and intonational completion points coincide.
The data for this study consist of a total of 80 minutes of conversation divided equally by languages, English and Japanese, and by channels, face-to-face and telephone conversations. The turns in the data are categorized into seven unit types; sound unit turn, laughter turn, word unit turn, phrase unit turn, single-clause unit turn, double-clause unit turn, and multiple-clause unit turn. The data analysis shows the following results: First, both languages have high percentages of single-clause units. This result indicates that clauses are basic turn units in both English and Japanese conversations. Second, the Japanese data have an extremely high percentage of sound units, while the English data have more frequent occurrences of longer turns. Third, the turns in telephone conversations co-occur with grammatical and intonational completion points more frequently than those in face-to-face conversations. Fourth, the larger the turn units are, the more frequently they coincide with CTRPs. This is because the larger unit types are more intrusive and thus less smooth if they occur at non-CTRPs.
Therefore, this study concludes that turns are more equally distributed between the interlocutors in the Japanese data than the English data, and that non-verbal cues play an important role in signaling turn transition relevance places.