Musicians have always been fascinated by the patterns of human speech. The composer Leos Janácek (1854-1928), for example, would jot down 'speech-melodies', which he heard around him, in a process rather similar to the one described in this project. More recently, Steve Reich has used recordings of spoken text to generate short melodic phrases in a number of works, including Different Trains (1988) and The Cave (1994). The word-phrase-sentence structure of speech is replicated in much music, particularly where the human voice is the basis of the melodic construction. This is even true in some electronic music, where the limitations of the instrument are not normally a factor in determining the lengths of phrases. There is something fundamental to musical communication in the patterns of speech, even if direct comparisons with language are not always helpful.
Sit in a crowded café or restaurant and listen carefully to the patterns of human speech. Try not to understand what people are saying, just the sound they make when they say it. Just the same, ask if you can hear meaning in the sound? Are there recognizable rhythms and shapes? How long do these go on? In what ways does the overall level of activity vary? Document the sessions in writing and do several sessions in different locations on different days.
The same exercise can be tried using recording. It is interesting to note the difference between what the ear hears and what the microphone 'hears': the two are not necessarily the same.