The technique of type-token ratio (TTR) provides a useful way to identify the individual's voice in asynchronous electronic discourse. TTR indexes lexical repetition and variation within comparable segments of text (Carpenter, 1990).

Using the software program WordCruncher, a concordancer which indexes all words and all unique words, we hand-tagged syntactic features and calculated the type-token ratio for each successive 50-word segment in every writing that was 150 words or more in length.

Next, we illustrate the kinds of findings obtained with concordancers for the writings by two students in a single conference; then we focus on features in their writings from a single topical thread within the conference.

The two students each wrote well above both the average number of words and the average number of unique words for the total conference:

   Total, all words:         40644
   Average for 67 students:  606.62
   Dutton:                   1334
   Kendra:                   1174
   
   Total, unique words:      3803
   Average for 67 students:  56.76
   Dutton:                   427
   Kendra:                   420

In terms of total output of unique words, Dutton wrote 2.20 times as much as the student average, and Kendra wrote 1.93 times as much. Both students presented complex sentences filled with multiple complement clauses, participle constructions, and prepositional phrases; what distinguishes each from the other is their use of mitigation.

Kendra, as illustrated in Table 3 for a single entry, presented nearly twice the amount of negation, modals, downtoners, and emphatics as Dutton, whose assertions, keyed to both self-disclosure and to information or opinions from other writings, were both more confident and more likely to be elaborated with prepositional phrases. Of additional interest is their use of public and private verbs, which, occurring in complementary distribution, underscored individual shifts in stance. --------------------------------------------------------------------------------------------- Table 3 GOES HERE ---------------------------------------------------------------------------------------------

Kendras's shifts to greater repetition and less variety were accompanied by self-disclosure of feelings or personal reactions; Dutton's shifts to greater repetition were generally cued to his elaboration of opinions. Their individual patterns in these two writings held true across other postings to the conference: that is, instead of repeating specific single words, e.g. "arrest," each presented lexical variation with "arrested," "arrests," "arresting" to a greater extent than their peers, and both consistently presented their own individual patterns of style shifts, keyed to individual preferences for the presence or absence of specific lexical and syntactic features.

Dutton's writings typically began with a reference to text, either in the newspaper story initiating a topical thread (paraphrase), or to writings by other students in the thread (direct address and paraphrase). He noted his area of agreement with one or more student interpretations, presented it as a claim, and elaborated his position with other references to text and to self-disclosure of his opinions. Kendra's writings typically began with reflexive writing about the language used in the newspaper story (unmarked single-word quotation and paraphrase), which scaffolded her series of narrative illustrations from personal knowledge of similar events or issues, and self-disclosure of feelings about the similarities.

Figures 1 and 2 display the full pattern of shifts in the same writings whose public and private verbs were displayed in Table 3. --------------------------------------------------------------------------------------------- FIGURES 1 AND 2 GO HERE ---------------------------------------------------------------------------------------------