This commentary is a critical appraisal of Gil-Pérez et al. ‘s (2002) conceptualization of constructivism. It is argued that the following aspects of their presentation are problematic: (a) Although the role of controversy is recognized, the authors implicitly subscribe to a Kuhnian perspective of `normal’ science; (b) Authors fail to recognize the importance of von Glasersfeld’s contribution to the understanding of constructivism in science education; (c) The fact that it is not possible to implement a constructivist pedagogy without a constructivist epistemology has been ignored; and (d) Failure to recognize that the metaphor of the `student as a developing scientist’ facilitates teaching strategies as students are confronted with alternative/rival/conflicting ideas. Finally, we have shown that constructivism in science education is going through a process of continual critical appraisals.

This paper argues that constructivist science education works with an unsatisfactory account of knowledge which affects both its account of the nature of science and of science education. The paper begins with a brief survey of realism and anti-realism in science and the varieties of constructivism that can be found. In the second section the important conception of knowledge and teaching that Plato develops in the Meno is contrasted with constructivism. The section ends with an account of the contribution that Vico (as understood by constructivists), Kant and Piaget have made to constructivist doctrines. Section three is devoted to a critique of the theory of knowledge and the anti-realism of von Glasersfeld. The final section considers the connection, or lack of it, between the constructivist view of science and knowledge and the teaching of science.

There is an old argument in philosophy that attempts to undermine any realist account of truth and thus knowledge in science and elsewhere. This is the claim that we can never compare either our experiences of, or our beliefs about, reality with how reality is because, in order to check whether our experiences or beliefs correspond with reality, always further experiences or beliefs must intervene. Thus we can never have direct knowledge, or more strongly any knowledge at all, of how reality is. Constructivists take this to heart and look for accounts of knowledge that are anti-realist and coherentist and/or constructivist in which the role of the external world in knowledge claims is thoroughly downplayed or non-existent. This argument is a central tenet of constructivism both with respect to scientific knowledge and learning. The argument, though widespread, is not without serious difficulties that leave the realist position untouched.