Glad to see Kurt jumping into the fray. Here are a few thoughts in reaction to Ed Kleinman's [12 Nov 94] questions.
Semiotic ISD provides a focus on learning, not just on the learner. This means the signs used by the designer and those present in the learning environment are just as important to the instructional design process as the signs of the learner. To this point, there has never been much focus on the assumptions made by instructional designers when we design instruction, and we have a lot of implicit assumptions that affect how we do our work. A semiotic perspective requires a critical look at ourselves to hold those assumptions up for question. There are several folks looking a critical theory as a means of doing this (e.g., Denis Hlynka, John Belland and the Paradigms Regained book).
Gary Shank (in his ITForum comments [9 Nov 94] and other works) has done a great job of distinguishing between semiotics and constructivism, which are, in his words, worlds apart. In my view, constructivism is currently an epistemology without a corresponding learning theory, and as such, does not provide an effective guide yet to designing instruction. Semiotics, with its well developed theory base, has the potential of offering a theory of learning and a theory of communication that can provide effective guidelines for instructional design.One thing that Kurt [Rowley, 11 Nov 94] alluded to in his comment is the dynamic nature of learners' interpretations and the need to accommodate to those in instruction. Current paradigms seem to take for granted that once knowledge has been acquired, it is in some way static in representation. For example, some of the current research on concept mapping talks about experts mapping their structural knowledge so that learners have something to learn with. That presupposes that the expert's knowledge is stable, static, and capable of being represented in that form at any moment in time. From a semiotic view, I argue that not even an expert's knowledge is stable, much less a novice's. What we measure with a concept map is only a slice of knowledge. Despite its dynamic nature (and perhaps because of it), it can still be useful.