The high reputation Lacan's thought still enjoys in Humanities departments, both in Britain and elsewhere, is, so Tallis argues, in this reading from Not Saussure (1988, 1995), to be attributed to the fact that those trained in the interpretation of literature (and, one might add, of cinema) have little or no knowledge of scientific method and a less than rudimentary grasp of clinical practice. None the less, it is no good trying to refute Lacan's ideas, since one will seek in vain for the systematic observations on which his conclusions are based. For example, there is little reference to clinical material of his own, which means that his interpretations of his own cases cannot be assessed since insufficient material is given. For Tallis, then, Lacan is a prime example of how not to proceed in diagnostic medicine. He reinforces this view by a detailed account of Lacan's presentation of the so-called mirror stage. This is supposedly the period in infant development between six and eighteen months during which a unified sense of the self, of one's personal identity, is formed and established. Lacan's account of the self has had enormous importance for literary theory; in particular, it has been widely used to justify the notion that referential realism is impossible. The realities of self and world are thought to be composed of fictions that belong solely within the symbolic realm, closed off from any contact with what might lie beyond language. However, as Tallis shows, Lacan's theory assumes that which it sets out to explain. Lacan tries to establish that the child's sense of self is mediated through an identification with his or her own mirror image, and his argument is based on a series of dialectical shifts between self and Other. But, to pull this off, Lacan begins with the assumption he needs to establish, namely, that the child can recognise his or her own image in the mirror. Furthermore, Lacan never explains how, if the self is a fiction, the child is able to refer so many different images of its own body to the same self. These and other objections reveal the fundamental nonsense that Lacan's account succumbs to. It is a nonsense Tallis links to the muddles associated with some versions of genetic epistemology — the process of describing how the child comes to acquire the adult metaphysical world-view. This is to do descriptive psychology and ontology at the same time, and the result is confusion of the sort we find in Lacan.