Studies in dialogue and discourse; an exponential law of successive questioning

The structure of natural conversations in first-grade classrooms is the focus of this inquiry. Analyses of a particular type of discourse, namely, connected conversations initiated and sustained by questioning, suggest that the probability that a conversation will be continued may be expressed as a simple exponential function. The formula, p i = ar i −1 , generates a curve of theoretically-expected rates of successive questions in a series that closely matches observed rates. The formula is based on the application of a constant ratio, that is, the ratio of rates within each pair of adjacent questions is the same throughout the series: p 2: p 1= p 3: p 2 = p 4: p 3. … Thus, it appears that the probability of a ‘next’ question following an exchange that contains a previous question remains constant through the length of the discourse series. In other words, the probability of a question is independent of the temporal location of an utterance in this type of connected conversation. The analyses suggest further that the model of a finite Markov chain, that is, of a particular type of stochastic process, may be applicable to certain features of a discourse. (Conversational analysis, sequencing in exchanges, U.S. English in first-grade classrooms.)