200 LANGUAGE, VOLUME 76, NUMBER 1 (2000) 4, "The realist paradigm' (129-77), explores the role of consciousness-marking devices in the type of novel that came to be called 'realist'. Particularly interesting here is F's discussion of the transition from drama to fiction, with the dramatic scene serving as a precursor to the narrative episode in the novel. Ch. 5, 'Reflectoralization and figuralization' (178-221), develops a new vocabulary for describing phenomena noticed by previous narratologists in their work on the notions of 'voice' and 'perspective' in narrative. Reflectorized tellers have the knowledge of a more or less omniscient narrator but the perspective and speech style of a character in the story; figuralization comes about when a narrator implicitly evokes an observer figure not actually present in the story. Ch. 6, 'Virgin terntories' (222-68), shows how experimental texts use 'odd' personal pronouns to enrich narrative's resources for the presentation of consciousness. For instance, narratives using one and it toy with the category of person, reinforcing F's overall argument that readers must 're-cognize in terms of familiar natural parameters and frames what initially appears to contravene natural or commonsense expectations' (223). Meanwhile, second-person narratives (along with stones using one) create new 'identificationaF potentials for stones, prompting readers to project themselves into the narrated action. Present tense and future tense narratives, as well as stories told in nonindicative moods (e.g. the imperative), present additional challenges for recipients who must sometimes work against all odds to narrativize such odd texts. Ch. 7, 'Games with tellers, telling and told' (269-310), studies narratives purposely designed to be inconsistent. Certain kinds of inconsistencies can be made sense of—i.e. narrativized —by recourse to the hypothesis of an incompetent narrator. Other narrative texts incorporate conflicting discourses harder to interpret by means of basic storytelling frames and parameters. Ch. 8, 'Natural narratology' (311-75), reviews the results of the discussion, compares the author's approach with those developed by previous narrative theorists, and argues for the claim that experiencing constitutes the core parameter for narrativization, subsuming plot- or action-based parameters. [David Herman, North Carolina State University.] Negation and polarity: Syntax and semantics . Selected papers from the colloquium 'Negation: Syntax and Semantics ', Ottawa, 11-13 May 1995. Ed. by Danielle Forget, Paul HirschbUhler , France Martineau, and María-Luisa Rivero. Amsterdam: John Benjamins, 1997. Pp. viii, 367. The editors ofthis book have brought together seventeen papers on syntactic and semantic aspects of negation and polarity. No further attempt has been made to arrange the diverse perspectives that these papers provide under specific sections or to relate or unify the ideas by proffering an introductory chapter. Nonetheless, the topics broadly fall into two categories —negative concord (NC) and negative polarity (NP). NC in relation to «-words is addressed in papers by Joäo Peres and Jacob Hoeksema; to Neg-criterion by Viviane Déprez, Liliane Haegeman, and Daniel Valois; and to A'-binding by Paul Rowlett . NP in relation to long-distance licensing is discussed in a paper by Anastasia Giannakidou and Josep Quer; to scalar models by Michael Israel; to only by Laurence R. Horn; to NP items in Maori by Elizabeth Pearce; and to focus by Eugene Rohrbaugh. The rest of the papers are "The syntax of French negative adverbs' , Anne Abeille and Daniele Godard ; "The syntax of sentential negation in French and English', Denis Bouchard; 'Non-negative negation and Wh-exclamatives', M. Teresa Espinal; 'Negation as a reflex of clause structure', Aafke Hulk and Ans van Kemenade; 'La négation comme expression procédurale', Jacques Moeschler; 'De partitif et la négation', Claude Müller. There are examples from dozens of different languages, and many language families are involved in cross-linguistic studies of these notions. Turning to NC, a conventional wisdom suggests that languages can typically be divided into negative concord languages and double negation languages, according to an appropriate categorization on the behavior of their «-words. This view is extended in JoAO Peres's paper 'Extending the notion of negative concord', where certain parametric variation on the licensing of «-words is seen to give rise to three different dimensions for slicing up languages. No...