The semantics of fictional texts is an area of literary studies with a long and eminent tradition. In recent years, this tradition has been revitalized and greatly enriched by a group of scholars who have joined forces to construct a theory based on the philosophical concept of 'possible worlds.' Thomas Pavel, Marie-Laure Ryan, Ruth Ronen, among others, deserve a special mention for their distinguished contributions, but the one name that is to be given the title of leading figure is Lubomír Dolezel. A pioneer in the application of 'possible-worlds semantics' to the study of fictional texts, in Heterocosmica: Fiction and Possible Worlds, Dolezel brings together the results of his extensive and original research - spanning beyond two decades - in the field of literary meaning. Before analyzing the high theoretical import of the book, one should emphasize that the book also possesses a practical merit. In effect, it offers itself as a (long-awaited) unifying reworking of a considerable number of papers previously published. What adds to the interest of Heterocosmica is that the reworking of these writings is intended not as a mere assembling, but rather as a project of theoretical re-elaboration and refinement. As the author states in the preface, the result is a new work "both in overall design and in particular formulations" (xii).

Dolezel's work is divided in four main parts - the prologue, two central sections and the epilogue - supplemented by a rich apparatus of notes and references of primary and secondary sources. The inclusion of a glossary of the key-concepts used throughout the text is a most helpful choice, as it provides students and scholars of literary theory with a comprehensive recapitulation of the particular cognitive vocabulary advanced by the author. Another point to be stressed is that the compositional method adopted by the author is greatly beneficial to the reading and understanding of the dense conceptual content of this book. In articulating such content in three different kinds of chapters - 'theoretical' (T), 'theoretical with examples' (TA) and 'analytical' (A) - Dolezel attains a clear expositional framework inspired by the classic triad of epistemological modes of poetics ('theoretical,' 'analytical,' and 'exemplificatory').

The prologue - "From Nonexistent Entities to Fictional Worlds" - focuses on the problem of defining the concept of fictionality. As such, the opening section serves to lay the theoretical foundations of the entire study. After having affirmed the exigency of tackling the problem in an interdisciplinary fashion - by establishing a frame of exchange between conceptions and developments in philosophy, linguistics, semiotics, and literary studies as such - the author proposes his own definition. He contends that fictionality "is primarily a semantic phenomenon located on the axis 'representation (sign)-world'; its formal and pragmatic aspects are not denied but have an auxiliary theoretical role." (2) This narrow definition is a mere point of departure, as an essential formula to be expanded and clarified. Dolezel argues that in order to avoid a reductionist perspective on the semantics of fiction the first move is the adoption of a mild pluralist ontology. According to this position, there are as many universes of discourse or domains of reference as the number of conceivable possible worlds. Following Adams's 'actualism,' the actual world is accorded the status of an ontological standpoint - against the 'possibilism' of those who put it on the same ontological footing as possible worlds - but, simultaneously, it is also denied any exclusive monopoly in terms of reference. On the basis of this mild pluralist ontology, Dolezel shows that fictional discourse creates its own domain of reference as a set of "nonactualized possible states of affairs."

It should not go unnoticed that this theory of fictionality is conceived and directed against a specific target: those conceptions that deny the legitimacy of the notion of fictional reference. With clear and convincing arguments, Dolezel exposes the weaknesses of some of the most prominent theories. Russell's logical realism (fictional terms lack reference), Frege's two-language semantics (characterized by the distinction between a cognitive or referential language and a poetic or pure-sense one), Saussurean self-referentiality and, lastly, mimetic criticism with its universalist hermeneutics (exemplified by Auerbach and Watt) are all reiterated and criticized. Despite substantial differences, all these approaches share a significant common denominator. They are all predicated upon what Dolezel calls the 'one-world frame,' an ontological model which assumes the existence of one single universe of discourse, the actual world. The alternative is a 'multiple-world frame' - the pluralist model - which eludes the (ontological) reductio ad unum of the 'one-world frame' and affirms the validity of fictional reference by "positing possible worlds as the universe of fictional discourse..." (16). Within this universe, the possible worlds of fiction constitute a specific category. Because of the incompleteness of their structure - "only some conceivable statements about fictional entities are decidable, while some are not" (22) - fictional worlds must be distinguished from the complete worlds of logical semantics.

Moving from this general tenet - widely shared by philosophers and literary theorists - Dolezel goes on to argue that the most significant feature of fictional works is that they are products of "world-constructing texts." Such a differentia specifica is explained by attributing to literary texts a "special illocutionary force" in virtue of which possible worlds are called into existence in the form of semiotic objects. Unlike "imaging texts," which provide information about the actual world, "[C]onstructing texts are prior to worlds" (24), and as such their sentences are exempt from truth-valuation. Indeed, the textual typology elaborated by Dolezel - "world-imaging texts"/"world-constructing texts" - represents a valid alternative to the relativistic drift of Goodman's radical constructivism.

The prologue introduces the two central sections,"Narrative Worlds" and "Intensional Functions," that constitute the conceptual core of Heterocosmica. Dense and complex, on both the theoretical and the analytical plane, this vast region of the book exemplifies the unusual mélange of rigour and creativity that characterizes the author's body of work. "Narrative Worlds" and "Intensional Functions" form a pair of distinct but complementary sections, each containing an introductory chapter that defines the basic vocabulary. Borrowing Carnap's terminology - such as the extension/intension opposition - Dolezel pursues a bipolar fictional semantics: fictional works, that is, are conceived as both 'extensional entities' (world structures) and as 'intensional formations' ( forms of texture). That the extensive conceptual apparatus proposed - so extensive that we cannot discuss it here in detail -- results not in an abstract system but in a set of concrete and effective analytical tools is made manifest by the considerable number of textual applications (Hemingway's "Big Two-Hearted River," Dostoevsky's The Idiot, Kundera's Laughable Loves and The Book of Laughter and Forgetting, Kafka's The Trial and The Castle, Calvino's If on a Winter Night a Traveller, to mention only a few).

In "Narrative Worlds," Dolezel's initial contention is that fictional semantics must concentrate on "the macrostructural conditions of story generation..." (31). The key-notion here is not 'story' (as in classical narratology) but 'narrative world', namely the possible domain wherein a certain chain of events ('the story') is encompassed. This initial move and the subsequent appropriation of concepts and tools from action theory and cognitive psychology, serve to draw the lineaments of what may be called an 'anthropic' narratological model. Within such a model, narrativity is conceived as a property of "worlds-with-persons" rather than as a formal feature of the constellation of events. "World-without-person," contends Dolezel, "is below the threshold of narrativity. It is worlds with persons or, better, persons within worlds that generate stories" (33). The first development of this premise is the distinction between "one-person" and "multiperson" worlds. Two separate chapters are devoted to the analysis of their specific narrative horizons, a description undertaken with the help of a number of diverse textual examples. Notions like 'action,' 'event,' 'intentionality,' 'motivation' are first defined and then utilized as general categories to describe the stories of "worlds-with-persons." What is noteworthy is that the incursions in the territories of action theory, cognitive psychology and other related disciplines never smack of sterile eclecticism. On the contrary, they emerge as constituents of a coherent and perspicuous synthesis. Applied to individual texts, this synthesis provides compelling insights into the different narrative potentials of "one-person" and "multiperson" worlds, showing the superior malleability of the latter. The author's conclusion is that "[T]he semantics of narrative is, at its core, the semantics of interaction" (97).

Having individuated the general categories that shape narrative worlds, Dolezel shifts his attention on the other extensional aspect of their macro structures: the global restrictions which affect and 'order' the totality of the actions and events. To account for the presence and functioning of narrative constraints, the paradigm proposed is the logicosemantic concept of 'modality.' The assumption, here, is that narrative worlds generate stories which obey a logic of action and that the latter can be described on the basis of four systems of modalities. Drawing on the work of von Wright, Hintikka, Rescher, Dolezel posits an "alethic system," ( the 'classical' modalities of possibility, impossibility, and necessity); a "deontic system," (the modalities of permission, prohibition, and obligation); an "axiological system," ( goodness, badness, and indifference), and an "epistemic system" (knowledge, ignorance and belief). As global restrictions, modalities determine and control the configuration of the actions and events. The "alethic" triad delineates what is physically possible and impossible, whether what is (fictionally) the case corresponds to a "natural world" (where the laws of nature in the actual world are in force), to a "supernatural world,"where such laws are violated, or to a combination of the two, the "dyadic" structure of the "mythological world". The "deontic" and "axiological" modalities generate, respectively, the normative horizon (proscriptions and prescriptions) and the set of values/disvalues present in the narrative field of action. Lastly, the modalities of knowledge, ignorance and belief, in a manner reminiscent of Barthes' 'hermeneutic code,' define what every agent "knows, is ignorant of, and believes to be the case in the world" (126).

The study of narrative macro structures produces a descriptive method for representing the extensional meaning of works of fiction. But a project of fictional semantics can be deemed complete only if it is able to offer a theory of the aesthetic efficiency of literary language. In the second central section, Dolezel's chief contention is that such a theory requires an "intensional fictional semantics." As he puts it "extensional meaning is aesthetically neutral; only on the level of intension is aesthetically effective meaning achieved" (138). On the basis of this claim, the problematic and (nowadays) discredited distinction between the communicative systems of literature and science is reproposed in a convincing way. Literature and science, we are told, are not to be seen as distinct 'kinds' of languages but, rather, as different 'uses' of the constituents (poetic figures and devices, rhymes, connotata, and so on) of intensional meaning. In semantic terms, "literature (poetry) aims in the direction opposite to science: it is a communicative system for activating and putting to maximal use the resources of intentionality in language" (138).

Dolezel is aware that, being texture dependent, intensional meaning poses a serious heuristic problem: it is de facto non-paraphrasable. The solution he proposes is a "method of indirect analysis" based on the concept of function. In short, the intensional structuring of fictional worlds is seen as a product of the concomitant presence of the pair "authentication function"-"saturation function." The two components designate, respectively, the factor that determines what exists in the fictional world and the overall degree of explicitness, implicitness, and gaps ("zero texture") inscribed in the text. The "authentication function" is linked to the nature of performative speech act of the fictional text ( its being a special type of "constructing text") and defines the latter's (performative) capacity - "authentication force" - to transform 'possibles' into fictional facts. Thus, a fictional fact "is a possible entity authenticated by a felicitous literary speech act" (146). Moving from this conceptual basis, Dole el assigns the role of authenticating source to the persona of the narrator and, at the same time, propounds a gradatio in terms of narrative authority. Three main postures are individuated and described: the absolute authority of the Er-narrator whose anonymous discourse constructs a transparent domain of authentic facts; the quasi absolute authority of the subjectivized Er-form which "constructs fictional facts relativized to a certain person (or group of persons)" (153); and the relative authority of Ich-narrators who possess a limited scope of knowledge but also a privileged discursive status (that of fictional producers of the narrative). The chapter on the "authentication function" ends with the description of "self-voiding narratives" and "impossible worlds." These categories refer to two different strategies for obtaining the collapse ( the invalidation) of the above mentioned function: a pragmatic strategy and a semantic one. The pragmatic strategy of "self-voiding narratives" is exemplified by metafictional novels in which "the authenticating act is voided by being 'laid bare'" (162). In contrast, the semantic strategy of invalidation produces "impossible worlds" by introducing logical contradictions into the fictional domain of the narrative.

In the chapter devoted to the function of "saturation," Dolezel addresses the problem - a veritable vexata quaestio - of how to process and interpret gaps and textual implicitness. Contrary to Iser, he argues that gaps are not "stimuli or propellants for the reader's imagination" but "immutable projections of the world-constructing texture" (171). In other words, gaps are indices of the (necessary) logical incompleteness of fictional worlds and as such they authorize no filling-in. It is however the question of the recovery of implicit meaning that occupies the central position in Dolezel's reflection. 'Text theory' stipulates that implicit meaning is inscribed in the explicit texture in the form of markers - 'positive markers' (hints) and 'negative' ones (lacunae) - and that the latter permit the activation of procedures of inference (implication, presupposition). While adhering to this general assumption, Dolezel hastens to express a basic caveat. Inferential procedures, he argues, are "necessary but not sufficient for the processing of implicit meaning. Cognitive operations have to be activated in order to recover more than trivial or self-evident implied meanings" (176). The anti-mimetic contention advanced here is that fictional texts construct their own "fictional encyclopedias" - more or less deviant vis-à-vis the actual-world encyclopedia - thereby providing the reader with the relevant base(s) for his/her cognitive operations. With an elegant and appropriate image, the author asserts that the readers of fiction "must background the knowledge of their actual domicile and become cognitive residents of the fictional world they visit through the act of reading" (181).

This fine image of reading is taken up and expanded in the last section of Heterocosmica - "Fictional Worlds in Transduction" - where Dolezel assumes a decidedly critical stance toward the concept of 'intertextuality.' It is rightly pointed out that the intrinsic limit of such a concept is the reduction of literary space to a domain exclusively made up of relations between textures. To the radical textu(r)alism of "absolute intertextualists" (Barthes, Hutcheon), it is objected that the literary space is not only a system of interrelated intensional structures but also a store of fictional worlds (extensional entities). As objects of cultural memory, the fictional worlds of literature "enter into their own chain of succession, complementing and reinforcing or competing and undermining one another" (202). The double link (intensional and extensional link) characterizing the relation between literary works is defined "literary transduction."

More could be said about this rewarding and stimulating book that, no doubt, represents one of the most original texts of literary theory published in recent years. As a final note, we should observe that the lucid analytical rigor of the author as well as his terse prose are rare qualities in a field too often characterized by (confusing) conceptual fluidity and gratuitous obscuritas. But, more importantly, these two qualities are complemented by an even more rare spirit of theoretical creativity. This theoretical creativity, I think, is Dolezel's most unique achievement in Heterocosmica.



Luca Pocci

University of Toronto