The first three chapters of this book focus on French indefinite pronouns, centrally plusieurs, certain(e)s, les uns… les autres, personne, rien, and quelques-un(e)s. They deal with the apparent (but non-existent) synonymy of certain of these pronouns and their functioning within whole texts (chapter 1), the notion that the domain of operation of referring pronouns is at the meso- rather than the micro-textual level (chapter 2), and the relationship between pronominal reference and frames (chapter 3). Together with a fourth chapter that focuses on the definite singular clitic pronoun il, they make up Part I. Part II (four chapters) is devoted to the question of thematic development within French journalistic texts, the issue of criteria for textuality, based on an atypical text (dialogues in a surrealistic play), text understanders and their judgements of discourse coherence, and the organization of sentences within a written text with regard to its contribution to the stylistic dimension of textual meaning. The textual data used as a basis for the analysis are taken from broadsheet journalism (Le Monde, Le Nouvel Observateur, and the Canadian Hansard) and also from plays and novels (sometimes translated from English, Russian or Hungarian).

As is to be expected of a work that deals with (indefinite and personal) pronouns and with textual cohesion, the notion of anaphora plays a key role in the author’s thinking. Nagy relies on an entirely text-based definition (‘Le terme anaphore est à comprendre ici comme l’ensemble des procédés de renvois à un élément antérieur ou à un élément postérieur du texte’, 39) that makes no allowance for the targeting of a referent or sense unit ‘outside’ the text, in the understander’s ongoing mental representation of his/her interpretation of it, in terms of a relevant context; nor does it seem to allow for the (not infrequent) situation where there is no appropriate textual antecedent, or where targeting a given referent is achieved indirectly. This essentially text-centered conception of how anaphora functions reflects the author’s all-embracing view of text, whereby text is reified and hypostacized, implicitly presented as a dimension capable of describing and accounting for certain factors that, according to others (including this reviewer), more properly fall under the heading of ‘discourse’ (the mental representation of the negotiated content of a communicative event).

One apparent exception to Nagy’s approach to anaphora is her presentation and analysis of the 3rd person pronouns in Anna Gavalda’s novel La Consolante in chapter 4. Ironically, the chapter’s subtitle refers to the personal pronoun il as an ‘élément de cohésion textuelle’. The uses of il in the extracts given are however highly unusual, and not really ‘cohésifs’ at all. Indeed, there is no immediate ‘textual antecedent’ for the pronoun to achieve cohesion with. The text starts off with a clear instance of the ‘in media res’ device characteristic of modern literature (though this is not noted as such by the author): ‘Il se tenait toujours à l’écart’. Yet this does not prevent Nagy from using a co-referential index i to mark this occurrence. The use of antecedentless il continues, adding to the uncertainty as to the identity of the character referred to, until, in sentence 38 (Nagy numbers each sentence, but does not inform the reader of the import of the intervening material not cited), the feminine pronoun elle is suddenly used to refer to the same individual. The referential gender shift is (apparently) motivated via a dialogue between two of the characters, one of whom is unsure whether the individual in question is ‘un monsieur ou une dame’ (sentence 32). Interestingly, Nagy places a (here, as elsewhere, highly idiosyncratic) coreferential index i on both ‘monsieur’ and ‘dame’, despite the fact that these nouns are heads of indefinite NPs used to classify the individual at issue, not to refer. The uncertainty is lifted in sentence 36 by the announcement that the person is ‘Un monsieur bien sûr. Mais qu’il appelait sa nounou’ (il refers to Alexis, a character introduced earlier in the narrative).

In chapter 7, in particular, Nagy seems to contradict her previous text-centred approaches, the chapter being subtitled ‘Sujets interprétants et jugements de cohérence’. The author rightly emphasizes that texts are there to be interpreted by human subjects and she uses the term mentalisation to make her point. The focus on users’ interpretations highlights in particular their need to be compatible with the extra-linguistic ‘facts’ of the matter being evoked (the ‘fragment[s] du monde’, 59), a factor characterized via the term constringence (Nagy’s translation of a term introduced by S. Janos Petöfi). But this would seem to hark back to truth-conditional theories of language, more familiar from formal-semantic approaches. In Nagy’s illustrations taken inter alia from Shakespeare’s Hamlet, the exchanges of dialogue at cross-purposes derive in fact from the different knowledge bases of each character (and of the audience), and the one each attributes to the other (see Charolles, Reference Charolles and Lane2005 for instances where text fragments appearing at first sight to be locally incoherent are subsumed under a macro-topic, thereby ensuring higher-level coherence).

The author’s text-centred approach is further apparent in her analysis of theme-rheme structures in chapter 5, based on Combettes (Reference Combettes1983). Here, the various theme-rheme relations structuring a text are directly annotated on individual sentences, rather than (as befits a discourse-level information-structure phenomenon) on discourse units. Thus, in an article on French politics from Le Monde, the subject of each sentence systematically receives theme status, presumably because the subject function is closely associated with topics. However, this is by no means always the case, since (among other information-structural values) the referent of a given subject may be presented as an integral part of the situation being evoked, the utterance in question being ‘thetic’ (conveying all-new information). This is the case in the complex initial sentence of the article, where the main clause l’inquiétude grandit chez les europhiles surely conveys the gradual development of an overall situation, rather than predicating of ‘anxiety’ (an abstract concept) the property of ‘increasing among Europhiles’, as is implied by the annotation ‘(T₁)’ on l’inquiétude.

A similar case arises in the author’s analysis of an amusing article, also from Le Monde, about former US President George W. Bush arriving on an aircraft carrier to compliment the forces having fought in the second Iraq war. The author records her difficulty in connecting the subject of the final sentence (‘La ruse n’a pas plu à tout le monde’), marked as Theme 4 ‘(T₄)’, with a preceding theme in the text. To remedy this, she posits a new type of textual progression, termed progression conceptuelle, whereby the theme unit ‘la ruse’ would summarize and encapsulate the ‘contenu de tout un bloc de texte antérieur’, thereby ensuring both its thematic continuity and potentially moving the text forward in terms of what it implicitly predicates.

Secondly, given that the subject NP la ruse clearly relates to the implied point of the text (its intended perlocutionary effect), this would fall under the reviewer’s conception of ‘discourse’ (see above), as would all the theme-rheme annotations that Nagy assigns elsewhere to segments of the texts at issue. For we are surely dealing here with readers’ interpretations of the text, rather than with the text qua text. In line with Cornish (Reference Cornish2018), the subject NP la ruse in the final sentence would be classified as an instance of discourse deixis, defined as ‘the act of “cognitively targeting” a discourse representation accessible in working memory, and creating out of it via inference a partially new discourse entity. This entity will thus have both new and given properties at the level of discourse’.

Nagy’s démarche in this slim volume is at turns sophisticated and over-simplified. Sophisticated in that the style of writing is lucid, nuanced and often insightful, the author’s mastery of the various methodologies adopted being clearly in evidence; over-simplified and somewhat problematic in that the different chapters deal with disparate areas of research, without any connections being made with the concerns of the others, resulting in a number of contradictory positions. The various chapters are in fact translations by the author of stand-alone papers previously published in Hungarian. Of the 58 bibliographical references, 22 are in Hungarian (many by S. Janos Petöfi); the remainder are predominantly pre-2000. Great swathes of post-2000 work on the issues raised are thus ignored -- most importantly, work on the ‘discourse’ dimension of text-oriented research.