Misconceived misconceptions?

Detailed user activity scripts from two previous studies of novice users working at a command language or a direct representation interface were submitted to independent expert judges for the justified ascription of misconceptions. Our initial hypothesis was that behavioral evidence for such misconceptions comes about as a result of well-articulated hypothetical reasoning. Although the evidence we obtained supports this view, it also suggests that for the direct representation case some activity normally attributed to misconceptions is non-reasoned in nature and governed by inherent powers of the representation.