To appear in Linguistics and Education, 10 (3) MODES OF MEANING IN A SCIENCE ACTIVITY Gordon Wells Ontario Institute for Studies in Education University of Toronto When I selected the videotape of the shoebox experiment as the basis for practical work in my graduate class, I did so for two reasons. The first was because of its completeness as the record of the various stages of an activity, from initial selection by the two children to the writing of the texts in which they report what they have done and learned; the second was that I thought that, as I had been the 'visiting teacher', there might be advantages in my being able to offer a participant's perspective. As it turned out, both of these expectations were fulfilled - though the second not exactly as I had anticipated. For as we viewed, reviewed and discussed the tape analytically, it was the nature of my own participation that became problematic for me; in particular, I was disturbed by the differential manner in which I interacted with the two children. What was it, I wondered, that had led me to give so much more of my attention to Jessica than to Alan? At first I thought it was just that Jessica was more interested than Alan in my attempt to get them to see their efforts to replicate the phenomenon of 'bending light' - an 'experiment' that they had found in an illustrated book on light - as an opportunity to engage in the more scientific activity of formulating and testing an explanatory hypothesis. Certainly, on a purely verbal level, Jessica was much more willing than Alan to engage in my reconstruing of the activity, and she seemed to be the main beneficiary of my efforts. In fact, from the transcript made of the recording, Alan hardly seemed to participate in the discussion at all. However, as became