BOOK REVIEW Meaning and Interpretation Umber to Eco with Richard Torty, Jonathan Culler, Christine Rose-Brooke, Interpretation and Over-Interpretation, Edited by Stefan Collini, Cambridge University Press, 1992. The present volume provides invaluable access to contemporary debates about meaning, interpretation and the possibility of over-interpretation. It comprises the 1990 Tanner Lectures delivered by Umberto Eco, as also the papers presented by three participants at the ensuing seminar. Eco's Reply completes the volume. Stefan Cellini's introduction is commendable as much for its well selected background information as for its restraint in not imposing a perspective on the reader. The texts are allowed to speak for themselves. The directness of presentation and engagement immediately confront one adding a refreshing tension to the act of reading itself. Going into the 'archaic roots* of the contemporary debate on meaning ('most post-modern thought will look very pre-antique') Eco argues that the rationalist 'modus* and that which 'escapes the norm' have, since the days of Graeco-Roman antiquity, been dialectically related in the history of human thought. Consequently, recognising the potentially unlimited range of interpretation does not imply firstly, that interpretation has no object and secondly, that every act of interpretation can have a 'happy end'. To interpret a text, states Eco, is 'to explain why words can do various things (and not others).* Developing a sort of Popperian principle, Eco claims that even if there are no rules that allow one to ascertain which interpretations are the 'best', and therefore, authoritative ones, there are criteria for determining which are 'bad' ones. 'Over-interpretations typically fail to reject apparent similarities (possible between anything and any other) as irrelevant even when this relation is recognisably minimal. Attributing maximal possibilities in such cases renders the 'wonder' motivating the interpreter closer to 'paranoia' than to a creative sundering of culturally established norms of significance. Eco defines the text as a 'device conceived in order to produce its model reader'. A' the same time the 'initiative' of the model reader Social Scientist Vol. 20, Nos. 5-6, May-June 1992