Our previous examples have drawn on repetition and emulation of lexical and syntactic features to illustrate the individual's voice in asynchronous electronic discourse; our next examples analyze selected syntactic features and add an examination of discourse features for larger patterns of utterances. Two linked Conferences about Language-Learning involving L1 and L2 students asked them to present three types of text: summaries of writings by recognized authorities, support or refutation of other writings by recognized authorities, and responses to the writings by their peers, which gave students the opportunity to confer authority on others and to appropriate authority for themselves.

In general, the L2 writers, most of whom were graduate students in a master's-level TESOL program, wrote longer summaries and debates than the L1 writers, half of whom were undergraduates. Not surprisingly, both groups of students used more modals, more downtoners and hedges, as ways to shade the confidence of their assertions in the second conference, a staged debate, particularly if they were presenting a potential refutation.

Refutation, like any sort of critique, is often difficult for students; in a study of one group of student writers engaged in learning to write critiques, Mathison comments "It may be that for some students the very notion of texts as pliable and transformable was alien, and therefore, asking them to contribute to scholarly conversations was inconsistent with their views on knowledge" (Mathison, 1996, p. 345).

Given the different cultural preferences for whether, when or how to express disagreement, one might expect to find differences in, for example, tokens suggesting mitigation across L1 and L2 writers. These, however, were not always substantial; the use of modal verbs presented considerable variation both across text-types and across student cohorts.

Modals presented by both groups L1 and L2 were almost always epistemic in force; however, each group had decided preferences for which modals to choose, keyed to text type. The difference between summary and debate apparently required the writers to assume a different stance toward the authority being summarized as opposed to the authority being refuted. Asian L2 students, like L1 writers, preferred "can" to "may," "should" to "would" for summaries, where they presented far fewer modals than the class average (3.15 for all, v 2.2 for Asian L2). In debates, where each writing on average for all students presented nearly 4 modals per writing, L2 students rose to near that average (3.97 for all; 3.57 for Asian L2), and modal preferences shifted slightly, with Asian L2 students increasing their uses of "may" and, especially, "should." However, the usage for modals alone did not fully contextualize the emergence of individual voices in the Language-Learning conferences.

Examining the changing structure in the sense of rhetorical organization for "Wang," a Chinese student, proved especially interesting. Her instructor neither read nor spoke Mandarin, the language of Wang's education, though a second spoken Chinese language had been part of Wang's life since childhood. Deciding to become well-versed with the conferencing software, Wang chose to write a third set of replies (Additional Replies) and append it to the Summary conference a month after it ended. Her modals shifted ("could" and "will," for example, exceeded averages for the whole class and for the cohort of other L2 student writers), as did her use of I/you references.

More interestingly, she moved to vary her preferred rhetorical pattern, which when charted first by an adaptation from Labov (1972a) and then by the 4-part structure from classical Chinese rhetoric (Kirkpatrick 1995), showed that she was beginning to vary her pervasive use of compliments (see Ye 1995), the targets of those compliments, and the third part of her usual structure, typically a summary of a previous text. Table 4, which charts the successive parts of an early Reply followed by three of her 10 Additional Replies, shows Wang's changes in responses to other authorities, who were graduate students writing in the conference. These included one with whom she had conversed only in the conference discourse, two with whom she also had other classes, and a fourth, with whom she had collaborated in small-group work and dyadic writing.

We have boldfaced personal pronouns, and underlined modals to show how these correlate in successive entries with the "voice," the identity she constructed in her responses to each of these people. ------------------------------------------------------------------------- TABLE 4 GOES HERE -------------------------------------------------------------------------

Which of these entries best displays Wang's individual voice in electronic discourse? We suggest instead that they collectively display parts of her increasing repertoire, what we might call a "portfolio" of voice, emerging as she became more confident ("could," "will") in appropriating and reconstructing words from others in a new language and a new medium of expression.

We have but limited ways to explore in Wang's English writings what Cummins summarizes as five stages in the development of voice (Cummins 1994, p. 53). These stages seem to be way-stations as the person becomes aware of appropriation and reconfiguration. Powers and Gong (1994, p. 214). present an overview of some of the features in East Asian rhetorics, calling for the examination of "East Asian Voice and the Expression of Cultural Ethos." We propose that electronic discourse, whether synchronous or asynchronous, be part of such an examination.