Basil Lourié

POSSIBLE WORLDS OF DIFFERENT NARRATIVES*
FRANK R. ANKERSMIT, LUBOMIR DOLEZEL, AND NARRATOLOGY OF HISTORY.**

*Translated from the Russian by Michael Klebanov, edited by the author.

**The present work is a part of forthcoming monograph dedicated to the theory
of critical hagiography in particular and to the logical analysis of narrative in general. .

  As is well known, Aristotle actually built his "Poetics" around the idea of interpretation of any artwork as an "imitation" (???????). In the "Poetics" nothing is told about the poetics of historical narratives, yet it seems evident that they are even more of an "imitation". Aristotle's "imitation" is an imitation of reality: of characters, emotions etc.

In 20th century general poetics of "imitation", both for so-called literary fiction and historiography, was devised by Erich Auerbach (1892—1957), the German literary historian who emigrated to USA. His principal volume bears a characteristic name: "Mimesis: The Representation of Reality in Western Literature" (1946). Borrowing the language of slogans it might be said that in the domain of poetics Auerbach is Aristotle today.

However, today Aristotle doesn't meet all our needs in the field of logic and, consequently, in any other field as far as it's based on logic. Poetics is no exception, be it literary fiction or historiography (a propos, it's none other than Auerbach, though not only him, who explains why it would be better to enclose "fiction" in quotation marks: no one knows for sure where lies the difference between literary fiction and any other sort of literature; and if such difference does exist, it must be strongly affected by various fashion changes and is naturally very much relative).

The very concept of "imitation" works only when two conditions are simultaneously fulfilled; yet these conditions are never really fulfilled:

1) It is initially clear who can "imitate" whom, – which presumes the concept of "objective reality" or something of the kind;

2) The very mechanism of imitation is such that it presumes the possibility of unambiguous indication of certain, quite well-defined objects that are part of this very "objective reality".

If somebody allowed us on the spot to build up some kind of "Copenhagen interpretation" of literary text, we would immediately declare here that literature cannot be an "imitation". And really: what kind of reality is "imitated" by the results of measurement of coordinate and impulse linked by Heisenberg's uncertainty relations? Formally we can see here transgression of the condition (2), whereas Copenhagen interpretation tells us that it is caused by transgression of the condition (1). According to the Copenhagen interpretation, the true reality here is the process of measurement itself, but none of the magnitudes of values taken apart from the situation of their measurement. There is no reality to "imitate"; only the process itself is real, and nonetheless there are attempts to call it "imitation" without any ground.

Therefore the Copenhagen interpretation of "literary" or historic narrative might suggest that there is no "reality" that can be imitated, yet there is reality created within the narrative itself. As in the case of quantum physics, this is neither reality of subjective idealism nor "objective reality", a notion peculiar to the mainstream of Western philosophy.

No matter whether we speak of natural sciences, historiography or "literary fiction", the reality behind our narrative involves ourselves, which means the narrator, with all the necessity.

Therefore our tales contain no reality that is not "feigned", yet our fiction won't do without reality as well. This doesn't mean that the extent to which physical dimension, historic narrative and literary tale are "realistic" doesn't vary, but that the difference is precisely in the extent, that is, quantitative and not qualitative.

Both contemporary study of literature and contemporary science of historiography (I can't find a name for it) arrived at this conclusion even without initial assumptions as radical as "Copenhagen" notion of reality. One might feel sorry that it made the way to such conclusions longer, yet be glad as well that its course has been more elaborated and that now we have far better knowledge of its details.  

It appears that even if the condition (1) held, the condition (2) would be transgressed anyway. If "objective reality" existed (which is the belief of the majority of professional philosophers), its unambiguous "objective" description would have been nonetheless impossible. This was proved as a strict logical conclusion in 1977 by the American logician Hilary Putnam (Hilary Putnam, Models and Reality // Idem, Philosophical Papers. Vol. 3, Realism and Reason (Cambridge etc., 1983) 1—25, first published in 1980; see also other papers in this volume; see also: Hilary Putnam, Reason, Truth, and History. Cambridge: Cambridge University Press, 1981), though there were scientists who had similar, more or less vague intuitions before (we will yet return to this subject). Last 25 years saw many vain attempts to refute the "Putnam's theorem", while it is still a discutable matter. David Lewis, one of the first who failed to do it, died with hope that someone else will find an error there.

"Putnam's theorem" provided philosophical-logical ground for miscellaneous theories in various branches of science that evolved on their own without distinct knowledge of their premises. Cognitive linguistics, on its basis Elena Semino develops interesting approaches to the possible worlds of literary narratives (Elena Semino, Language and World Creation in Poetry and Other Texts (London—N. Y., 1997) (Textual Explorations, [1]) 117—233: Part III. Poetic Text Worlds as cognitive constructs), may be an example. Cognitive linguistics is mentioned here also because one of its founders dedicated to the "Putnam's theorem" remarkably extensive and non-expert friendly survey in his monograph: George Lakoff (1987), Women, Fire, and Dangerous Things. Chicago, IL: University of Chicago Press. Related approaches started to evolve in the 70's in the field of historiography as well. Actually, all the mentioned concepts started to evolve in the 70's simultaneously and regardless of each other, and since the 80's began, though not yet in a full measure, to interact.

What those concepts of "literary" and historic narratives that we are going to discuss have in common is a notion that any of them is busy with creating their own world, but not with "imitating" anything that is part of "objective reality".

In the literature studies this approach is associated mainly with the name of Lubomir Dolezel who started publishing his ideas in 1976. Now they are accessible in a concise monograph L. Dolezel, Heterocosmica. Fiction and Possible Worlds (Baltimore—London, 1998) (Parallax. Re-visions of culture and society). For a general exposition of the same approach see also Thomas G. Pavel, Fictional Worlds (Cambridge, Mass.,—London, 1986), where one can see an elaborated critics of Auerbach’s and structuralists’ approaches (both based on the notion of mimesis). As to the historiography, a similar approach was put forward by Frank Ankersmit whose conception was formulated in his monograph Narrative Logic. A Semantic Analysis of the Historian's Language (Den Haag: Nijhoff, 1983).

"Great narratology" or fictional history?

Facing Homer's poems we understand that we deal here with a mind that makes no particular difference between historical reality and reality of artistic fiction. Speaking of epos and the rest of folklore we take it for granted. However, can we be so sure that the subsequent literary genres underwent any radical changes concerning this matter?

This question is not idle, considering that we engage in hagiography where narrations that are quite epic dwell together with biographies preserving even pettiest everyday details of bygone ages… All these texts evidently functioned and were created for the same audience and for the same cult-serving goals. Thus, at least in the case of hagiography we can't speak of a radical change. It will be made still clearer when we consider more closely the logical order of historic narrative and "literary" text.

Let's begin with historic narrative: not for the reason that it's less complicate but reckoning that it has been, from the logical point of view, far less comprehensively studied, hence without further investigations we are likely to exceed the scope of existing material too soon.

The idea of historic and "literary" narratives returning back to their common roots has been formulated nowadays anew by the elder generation thinker, Paul Ricoeur, at the early 70's. Afterwards it was elaborated in his three-volume monograph Temps et recit (Time and Narrative. Volumes 1-3, trans. Kathleen Blamey and David Pellauer, Chicago and London: University of Chicago Press, 1984-1988): "historiography and literary criticism are summoned together to recreate the great narratology where equal rights of historic and fictional tale will be acknowledged" (emphasis is mine – B.L.).

The basic idea of Ricoeur's "great narratology" is common structure shared by historic and fictional tale. This idea is comprehensible intuitively and can be easily illustrated, for instance, by such genres as historic novel (fictional narrative involving historic characters and situations) or medieval chronicle (historic narrative replete with "legends", as historians are wont to call them). No wonder that historical novels are often meant to reconstruct the historical reality. However, consistent logical elaboration of this idea is fraught with certain difficulties, and its course has been only set by Ricoeur:

"…Common origin of historic and fictional narrative alone cannot explain their affinity, obvious even in their most elaborated forms: historiography and literature. Another reason of this steady compliance should be suggested: reunion of the narrative field is possible only in so far as configurative operations applied in both fields can be measured with the same measure. Construction of plot served us as a kind of such comprehensive measure. Hence it is small wonder that in fictional tale we found the configurative operation that we could compare historic explanation with. <…>. In this sense we just gave literature back what history borrowed from it".

It appears that the "construction of the plot" is inherent in any historic narrative as soon as it begins to deviate from fixation of facts lacking both the beginning and the ending (for instance, on the margins of some family Bible). Plot construction enlisted not only ancient and medieval historians but contemporary scientists as well, including those who boasted devotion to the "positive science". However, their labors were not researched by Ricoeur but by Hayden White who arrived at the same ideas though not without Ricoeur's influence (his main work is: Metahistory. The Historical Imagination in Nineteenth-Century Europe, Baltimore and London: Johns Hopkins University Press, 1973; Ricoeur scrutinizes White in the Time and Narrative, vols. 1&3).

Actually, this moment – issuing of Hayden White's "Metahistory" – is a starting point of contemporary revolutionary change of views on historiography.

Hayden White has analyzed works by only four, however most prominent 19th century historians (Michelet, Tocqueville, Ranke and Burkhardt), and has analyzed them as literary works. He found that they contained quite enough material for this sort of analysis.

White's work's weak strain was his notion of the theory of literature, in ultimate accordance with the structuralism of the day. Today he admits it partly on his own accord:

«“Metahistory” belongs to the certain, "structuralist" phase of evolvement of western humanities. Today I would have written it otherwise. Nonetheless I think that it contributed to the comprehensive theory of historiography because it treated with equal graveness both the status of historiography as a written discourse and its status as an academic discipline. Most of the historical interpretations of historiography proceed from the assumption that along with scientification of history in 19th century historical researches lost their millennial ties with rhetoric and literature. However account of history remains rhetorical and literary as long as it proceeds with using regular grammatical language and writing as preferred means to convey the results of the research of the past. While historians go on with using regular language and writing, their representations of the phenomena of the past, just as their thoughts, will remain "literary", "poetical" and "rhetorical", differing from everything that is considered specific "scientific" discourse.» (Author’s Preface to the Russian edition of H.White’s « Metahistory »: Х.Уайт, Метаистория: Историческое воображение в Европе XIX века/ Пер. с англ. под ред. Е. Г. Трубиной и В. В. Харитонова (Екатеринбург, 2002); retroversed from Russian).

White has called his approach "tropology", study of poetical tropoi used by historiographic narrative acquires its "literary" quality. Actually, today White regards the kind of "tropology" he proposed in 1973 as relatively inadequate; since we don’t consider it a chief value of his work, there's no reason to discuss it here.

"Metahistory" suggested, we believe, an unsuccessful (structuralist) theory, however, its significance is in the new, explicitly formulated researching program that has been just outlined by Ricoeur: the possibility to interpret any historiographic text, even the most "scientific", as a literary text; or similarly, to consider both types of narrative combined by Ricoeur into the single object of research, proceeding from their internal affinity.

There is another reason why White's works was perceived as far more revolutionary than the works of Ricoeur.

Ricoeur promoted his thoughts of the comprehensive "great narratology" following the concept of mimesis, whereas White chose the structuralist paradigm. This choice has been possible because it avoided the necessity of referring to completely different, logical foundations of theory (see Pavel about logical premises of structuralist literary criticism). However, structuralist approach defiantly emphasized the elimination of fundamental difference between historic narrative and literary fiction.

In the modern historiosophy, preserving the primacy of "objective reality", Ricoeur and White's approach stimulated development of new conceptions of "historicism" (J.Margolis, The Flux of History and the Flux of Science (Berkeley—Los Angeles—London, 1996) 188—193: criticizes Ricoeur and White for "losing the realism of historic (and physical) past" and concludes that exactly these mistakes create a stimulus for the new historicism). The situation reminds of the Vienna Circle's reaction on the Copenhagen interpretation (minding that Karl Popper, too, created his own version of "historicism").

Meanwhile, things moved on swiftly. In 1980, a work emerged that soon could be used as a bridge between structuralist literary criticism and logic, so as to abandon eventually the structuralism forever: monograph by G. Lacoff and a logician M. Johnson (Metaphors We Live By, Chicago: University of Chicago Press, 1980). This work formed so-called cognitive theory of metaphor that was further promoted not only in G. Lacoff's cognitive linguistics, but also in the narrative logic of Dutch philosopher Frank R. Ankersmit (therefore, in spite of all the difference of their subjects, cognitive linguistics and narrative logic are both genetically and structurally related disciplines).

Published in 1983, the conception of narrative logic exists hitherto without essential changes. We will have to give an account of it: not for the reason that we consider it appropriate for the purposes of critical hagiography (for this purpose we intend to propose somewhat different paradigm) but because it would allow us at the next stage to establish more sufficiently the logical status of single "great narratology" for critical hagiography.

Historiography and literature as functions

Literature: Frank R. Ankersmit, Narrative Logic. A Semantic Analysis of the Historian's Language, Den Haag: Nijhoff, 1983; Idem, History and Tropology. The Rise and Fall of Metaphor. Berkeley: University of California Press, 1994.

Trying to compare hagiographic narrative with some notions used by contemporary science we turned to descriptive methods of natural science (or, more precisely, to the notions of research programs), as well as to historiographic and literary narrative. If we want to advance, it will be important to mind all three of these approaches as a whole, side by side. Disciplinary dissociation of our days won't allow us just to advert to some research where ontological and epistemological foundations of all the three approaches are scrutinized concurrently. It's quite normal for contemporary science that a scientist who made a remarkable contribution to one or even two of these approaches wields but superficial or outdated knowledge of the rest. Regretfully we, too, can't carry out work exceeding even the tasks of Ricoeur's "great narratology": this would demand a separate and very extensive research; yet we have to outline at least some theses essential for the further practical work.

Our primary field of examination is an area where two tendencies, wittily designated by Ricoeur as «la fictionalisation de l’histoire» and «l’historisation de la fiction» (Temps et recit, III (Paris, 1984), 331—348) are explicit. We are starting with the first tendency that actually became an object of Ankersmit's narrative logic.

We will consider only some of the narrative logic's theses that are most approximate to the ideas we intend to promote (see, besides Narrative logic, Ankersmit's condensed account in Six Theses on Narrativist philosophy of History: Chapter One of History and Tropology). Let's make an immediate reservation that in truth Ankersmit's logic has much more in common with Dolezel's semantics of possible worlds to be discussed below; besides, both these concepts originate from 70's atmosphere that gave birth to Ricoeur and White's approaches; however, their detailed comparative analysis is out of scope here. If someone else turned to this task we would only welcome it because hitherto these, so closely related currents of thought never interacted in any way; yet we comfort ourselves with hope that our hagiographic observations would establish a fertile ground for this sort of work.

Furthermore we would like to point out that Ankersmit, who continually refers to natural sciences, tends nonetheless to perceive them in Kuhn's terms, seeming to overlook Lakatos completely (or at least never mentioning his name). This seems to be quite a regrettable misunderstanding because it's namely Ankersmit's principal notion, "narrative substance", is most similar to Lakatos' notion of "research program". Ankersmit himself admits it indirectly by noticing proximity between his "narrative substances" and Kuhn's scientific paradigms, adding also that such notions are inappropriate to Kuhn's "normal science" (Narrative logic). However, in our opinion, the very notion of "normal science" is inappropriate to science itself, as Lakatos shows it. Hence we will take liberty to expound Ankersmit on the background of Lakatos.

Narrative logic was built to answer the question: what is actually real in history? The fact is that events of the past remains in the past: and, in this sense, cease being real. Along with that it's impossible to deny them any status of being real since they are nonetheless somewhat different from the products of our imagination. Considering how hard is it for historian to distinguish between fiction and reality, we shouldn't expect that philosopher will find it easy to differentiate their logical-ontological statuses.

Here are Ankersmit's theses (taken from his Six Theses on Narrativist philosophy of History), showing clearly that they regard the notions similar to Lakatos' "research programs" He calls them "narrative interpretation" or, in other cases, "narrative substances":

"4.5.1. Our musings on the past are covered by the same thick coverlet that doesn't refer directly to the past but to its historical interpretation, and to discussions over competing historical interpretations. <…>

4.6.1. Facts of the past may be selected so as to help or to thwart narrative interpretation, yet they never can determine those interpretations <…>. It's only interpretations that can confirm (or not) other interpretation".

Examples of narrative interpretations are: "cold war", "renaissance", "mannerism", "industrial revolution". These examples even outwardly resemble notions of research programs: but in truth Ankersmit's notion of narrative substance is much broader and more complicate; below we will give a brief account of it intended for a reader who has at least basic knowledge of modern logic.

Ankersmit unconsciously reiterates Lakatos when he asserts the primary status of systems of interpretations (narrative substances / research programs) in relation to empiric facts (of history / natural science): "interpretations logically precede our models (notions) of things" (thesis 4.7.6). Systems of interpretations are subject to change and development, but these changes do not correspond mutually and unambiguously with experimental data. Each historian knows it all. There exist entire branches of science where impossibility of unambiguous interpretation of empiric facts causes continuous methodological problems: archeology, for example. Attributing this or another archeological "culture" to this or another people, whose existence is inferred from written sources, always constitutes a problem.

Therefore it seems advisable to recall the old name of natural science: natural history. This would allow us to redefine notions of historic narrative so that it might include natural-historic and socio-historic narratives all in one.
The proximity between historic narrative and descriptive method of natural science was ignored (or, more precisely, has been forgotten) by 20th century historians because descriptive method of natural science was compared with that of history on the level of scientific theories but not on the level of research programs that alone can evince the proximity of conceptual organization. The reason, probably, is Kuhn's popularity that obscured Lakatos in circles that were not concerned professionally with philosophy of natural science.

One more peculiar feature that can be found in Ankersmit's narrative logic is the proximity between historic and literary narrative: both of them are constructed by the means of tropoi (that are reduced by Ankersmit, following Lakoff and Johnson, to the notion of metaphor). Under such conditions this proximity becomes so evident that the opposite problem arises: how to distinguish between historic narration (research) and literary fiction?

Ankersmit answers this question so (once again his Theses are quoted):

"1.5.1. Historiography creates narrative interpretations of socio-historical reality; literature puts them to use".

This statement alone, even without detailed logical analysis (that we will do separately below) causes a feeling that there must be some problem about this differentiation. It's hard to imagine historian's labor that doesn't presume the use of interpretations it produces, as well as author's labor that doesn’t create such interpretations; and indeed, we've got into the habit of reading good historians' works as if it were good literature. In Russian cultural tradition Karamzin is a fine example; and if we appealed to the world culture in its entirety, it would be enough to recall historians of Ancient Greece and Rome.

In the same way, there is no reason to deny literary fiction creating narrative interpretations. Author's aspiration to create narrative interpretation of history is explicit in historical novels and especially in such genre-exceeding works as "War and Peace" and "Joseph and His Brothers" (it may be implicit in other cases, but it's always there). However, even an author trying to replace a historian (like L. Tolstoy) remains an author for all that.

This is all the more true for hagiographical monuments. They all claim they give an account of "history", yet they are "literature" all the same: not only Passions epiques but Passions historiques as well. Practice of contemporary science shows that the interest revealed by both historians of human society and historians of literature to the monuments of hagiography is more or less the same. It is grounded in the mere life itself but not explained by present historiography theories (and running a few steps forward we'll add: nor justified by present theories of literature as well).

Therefore let us introduce a certain amendment into Ankersmit, for which reason we actually turned to his concept of narrative logic.

Historiography (of both social and natural history) and literature are not definitions of different narrative modes but two different functions that are inherent in each one of the narratives. In other words, the difference between historic and literary narratives is quantitative and not qualitative: the only question is which of the function is more apparent. Maybe we in our "hagiographical corner", where both function are expressed more or less equally, could start seeing it more and more clearly. However, we will have yet to discuss the essence of literature as a function of narrative.

Historic narrative in the semantics of possible worlds

Now we are going to discuss the principal notion of Ankersmit's narrative logic, narrative essence: as well as the reasons why we don't like it and how can we remake it so that it would suit our own narratology that we are trying to work out here with hagiographical monuments in sight (and out of this reason we'd like it to be as universal as possible, just like hagiographic narrative itself that includes features of both types of historic narratives and of literary narrative alike).

Ankersmit considers (in Narrative logic) narrative essences at different levels: from abovementioned abstract notions to narrative images (essences) of ordinary things like table or chair. In the last case, when narrative essences of certain real things are discussed, Ankersmit doesn't encounter complicate logical problems: but such essences are not much wanted for the study of historic narratives (later logical objects of this sort will be thoroughly considered in the cognitive linguistics where G. Lakoff will introduce for their sake the notion if idealized cognitive model (ICM), see Lakoff, Women, Fire, and Dangerous Things).

Narrative essences that Ankersmit works with are being described as collections of statements that are nonetheless not just conjunctions of separate statements but endowed with specific content surpassing the sum of contents of separate statements. That which we have just now called "content" (for instance, content of the notion of "Renaissance" or "cold war") is ought to be called, turning to Frege's terminology, "sense" (Sinn). Then, the question is what constitutes the meaning (Bedeutung), or reference, of narrative substance.

If we choose to regard the statements that form narrative substance as sentences, then, according to Frege we'll have to infer that their reference should be the meaning of verity (which could be generalized as a notion of the meaning of verity of narrative substance). However, narrative substance simply cannot be false by definition, it cannot be falsified at all, being but constructed by our own mind (see above, Ankersmit's thesis 4.6.1; or thesis 4.7.1: "logic of narrative is strictly nominalistic"). Thus Ankersmit arrives at logical predicament: "if we shared Frege's position, all the acceptable Nss [narrative substances], no matter how different they may be, would designate invariably the same thing, that is, truth. I can’t imagine how to understand this solution. It looks rather like an unusual figure of speech than like a reasonable answer to a reasonable question". (Narrative logic). Hence Ankersmit suggests his own solution: narrative substances are devoid of reference at all:

"It might seem strange that such Nss as "Renaissance" or "state" have no referents in historic reality, but there is no other solution" (ibid.).

"Therefore we ought to overcome our intuitive resistance and abstain from postulating in historic reality things that refer to terms like "mannerism", "cold war", "state" etc" (ibid.).  Ankersmit observes logical predicament that originates, in its turn, from his resting upon Frege: "If Nss designate nothing, then we can reasonably ask: how in fact does reference to real things occur". By the way of answer Ankersmit "contents himself with most elementary cases of reference to things in reality, that is, those cases when the subject of statement designates a certain individual thing" (ibid; see for comparison the ICM conception by G. Lakoff). Ankersmit realizes that his proof isn't too much strict and least of all can be called exhaustive. The only answer he gives is: "I hope that more complicate modes of reference to reality can be ultimately reduced to these cases. If this hope will prove groundless, I don't think that the considered problem concerns only narrativist philosophy" (ibid.).

It's necessary to agree with the last sentence. We will face the same problem in literary criticism, and it certainly won't use up all possible examples.

Thus, at the very basis of Ankersmit's narrative logic we find a hypothesis that the author himself doesn't consider proven. We actually deem it quite false; however, we choose to seek the true solution on the road paved by Ankersmit himself.
If we recall that the very doctrine of narrative substances originates from tropology, and tropology, in its turn, has matured in the depths of structuralism, we shouldn't wonder why Ankersmit came to such solution.  It would suffice to recall the teaching of Jakobson about the poetical function of language that generates tropoi and its self-reference (see above) to make things clear.

Structuralism, as Thomas Pavel has clearly shown (Pavel, Fictional Worlds), failed to create an efficient conception of literary criticism precisely out of reason that it tended to interpret literary text too simplistically, by analogy with a word or a sentence: whereas in fact it dealt with an object far more complicate. Predicaments of Ankersmit's narrative logic are analogous: he tries to build the logic of historic narrative while resting on very narrow and, moreover, very approximately elaborated area of the philosophy of language.

It is no wonder that Ankersmit has found the solution of his narrative logic's cardinal problem within the very logical-philosophical tradition this logic emerged from. It is astonishing however, how close did Ankersmit get to the possibility of solution that wouldn't make us "overcome our intuitive resistance". Ankersmit himself mentions the possible compatibility of his ideas with such semantics and even introduces the notion of "narrativist worlds" that might be postulated in accordance to each of the possible worlds (ibid.), but in different context (referring to Leibniz whose logic is important to Ankersmit, and to whose ideas various kinds of semantics of possible worlds go back) and without revealing true interest to this topic.

Discussing the same problem of reference of narrative essences, Ankersmit writes in particular:

"…Most of the terms we use when discussing the past don't correspond to any identified objects. What is crucially important is whether these objects designate a certain Ns, or they are just specific concepts. Thus, since the term of the "fall of the Roman empire" designates Ns, we can agree with Munz when he writes: "There is no use to imagine that such thing as the fall of the Roman empire has really been and then investigate whether the explanation of either Gibbon, Rostovtsev or Seik (…). Gibbon's “Decline and Fall of the Roman Empire” and Rostovtsev's history are not just two attempts to describe the same event but two completely different historical narrations. There is not a single combination of events given two different causal explanations but two narrations about combinations of events" (ibid, emphasis is mine. The author quotes: P. Munz, The Skeleton and the Mollusk // New Zealand Journal of History 1 (1967) 107—123, esp. 122).

The emphasized words are actually definition, though implicit, of the notion of narrativist worlds as possible worlds. These are not Ankersmit's narrativist worlds supposed to correspond to other possible worlds, worlds of things, but possible worlds that are constructed independently in every historic narrative. We introduce this notion here in the same way as Dolezel introduces the notion of possible worlds for the analysis of literary fiction (about Dolezel see below).

"The fall of the Roman Empire" in the narrativist world of Gibbon and "the fall of the Roman Empire" in the narrativist world of Rostovtsev are not just two different events but two different events in two different worlds. Both of them are true, but each one is true only in its own narrativist world.

We will have yet to discuss how much relevant for the historic narrative is the notion that was grounded by Dolezel only for the literary narrative; yet for the present let us notice that while introducing the notion of possible world for historians with their mental reconstructions we are abstaining, for a while, from any speculations concerning the ontological status of possible worlds. This question should be discussed along with Dolezel's concept.

Returning in the meantime to the initial Ankersmit's question about reference of narrative substances, we'll have to answer it in the course of Frege's ideas interpreted in much the same way as Dolezel did it for the literary narrative:
The referent of narrative substance is a respective meaning of verity (always positive, that is "true") in a respective narrativist possible world.

Possible worlds of the three narratives.

Literature:L. Dolezel, Heterocosmica. Fiction and Possible Worlds (Baltimore - London, 1998: Parallax. Re-visions of culture and society) (all the following references to Lubomir Dolezel refer to this edition).

See also materials of the Nobel Symposium, 1986 (featuring in particular Thomas Kuhn and Lubomir Dolezel), where for the first time ever works on logic, philosophy of science, linguistics, literary criticism and art criticism were filed under the same cover: Possible Worlds in Humanities, Arts and Sciences. Proceedings of Nobel Symposium 65 / Ed. Sture Allen (Berlin—N. Y., 1989) (Research in Text Theory / Untersuchungen zur Texttheorie, 14).

Let's lay our cards upon the table: we are going to start expounding the concept whose significance for our suggested view on hagiography is similar to that of Delehaye's critical hagiography, although the author of this concept, Lubomir Dolezel, wrote nothing about hagiography as such.

The book "Heterocosmica" summarizes more than 20 years of development of this concept due not only to the efforts of Dolezel's efforts alone (who started publishing on the subject in 1976) but also those who experienced his influence. One of them is Thomas Pavel whom we've already mentioned above. In spite of its opposition to structuralism this concept goes back, through Jan Mukarovsky and the Prague Circle, to Russian formalism of 20's and even further back to A. Veselovsky (due also to the fact that Dolezel is an immigrant from Czechoslovakia and Pavel is from Romania; Dolezel has an excellent knowledge of Russian literature works of Russian formalists in their original); however, its principal structural element, which might be called its skeleton, is Anglo-Saxon (by its origin) analytical philosophy. Dolezel, who works in Canada, reasonably wonders at the addiction of literary critics in Anglo-Saxon countries to French and German schools to the detriment of their own tradition of philosophy: "Analytical philosophy has preserved the sober spirit of critical thinking at a time of bloated verbosity" (p. X). Dolezel's ideas, though yet in their early form (by the early 70's), were grasped and developed in Russia as well. Very important contribution, to our opinion, is Vadim Rudnev's book "Прочь от реальности. Исследования по философии текста" (М., 2000, ХХ век плюс. Междисциплинарные исследования).

Dolezel, it seems, provides the development of literary criticism with a link that would allow actualizing Veselovsky's old intuitions and adequate binding of historical literary criticism with critical hagiography.
However, prior to using Dolezel's concept we will point out its aspects that we deem unsatisfactory. Let's make a reservation: Dolezel dealt almost exclusively with the European literature, whereas our suggestion is that his ideas cannot be automatically extended to the cultural languages of other eras and peoples.

Once more about historiography and literature as functions

Dolezel introduces his notion of "fictional text" (p. 24—28, 235—236) on the basis of strict differentiation between what he calls I-texts (world-imaging texts; he identifies the term "imagination" with "representation") и C-texts (world-constructing texts), putting literary texts in the C-category. This differentiation seems obvious when the modern literature is concerned, and here Dolezel's position is strong enough. In particular, he can easily object to "mimetic" criticism that considers all texts as belonging to the I-category (after all, according to the concept of mimesis fictional works do not construct their own world but just "imitate" the only world that really exists).

However, this obviousness on which Dolezel's concept is founded doesn't last in historic perspective. For instance, let's turn to the medieval recensions of "The Romance of Alexander the Great" by Pseudo-Callisthenes, as well as to this late antiquity novel. This instance seems convenient out of reason that this novel was thoroughly examined by A. Veselovsky in his work on the shaping of novel and tale genres (see the Russian edition: А. Н. Веселовский, Из истории романа и повести. Материалы и исследования // Сборник Отделения русского языка и словесности Императорской Академии наук (1886) № 2, с. 1—511, 1—80), and also because we could see how easily this novel has transformed from a secular to a hagiographic one. Veselovsky didn't know yet about the ties between this novel and hagiography but he demonstrated quite clearly its direct connection to the latest European chivalry novels. Actually, although European chivalry novels are not hagiographical genre, they, too, are not devoid of hagiography, as far as they incorporated hagiographic legends (for instance, of the Holy Grail).

Thus, are various novels about Alexander C-texts or I-texts?

Regarding them in the light of their place in the history of literature (as did Veselovsky) we have to state that these are texts belonging to C-category, that is, fictional texts. However, their authors, editors and medieval readers scarcely chose to put them out of the I-category, which is the category of historic narratives that enlisted various biographies and chronicles.
Such reasoning might be relevant for nearly all the hagiographic monuments as well, which at once makes Dolezel's classification unacceptable as an instrument: we are just short of data in order to relate beyond doubt this or other hagiographic text to I-category or C-category. Putting "historic" hagiography in the I-category and "epic" hagiography in the C-category doesn't solve the problem. We had so far enough chances to see that the difference between these genres is rather quantitative and not qualitative. Thus, there exists an extensive "zone of transition" (to which panegyrics belong), and it often happens that the text seeming a "historical" hagiography turns out to be an hagiographical novel (for instance, the Life of Theodore of Edessa, the Life of Gregory of Agrigentum etc).

We arrive at the same point as in the case of historic narratives: in general, we can't speak of two types of text: I-type and C-type. Yet we can and even should speak of two functions showing in different texts to a variable extent: we will proceed calling these functions "literary" and "historical". Hagiography particularly urges us to insist on that since these functions reside there both distinctly and simultaneously.  

Dolezel, however, would very likely reject such conclusion since it doesn't agree with his idea of semantics of possible worlds. He would scarcely agree with our speculations on the interpretation of historic narrative. Hence we will have to display more comprehensive panorama of "multiworld" semantics and formulate our own thoughts concerning its usability for narratology, and at the same time formulate anew in Dolezel's terms our idea of both historic narrative and narrative of natural science.

Possible worlds: "a third way" between Kripke and Lewis

The notion of "possible worlds" goes back to Leibniz but no one ventures now to determine what he exactly meant by this. Leibniz expounded his thoughts about possible worlds rather fragmentarily, not so much in his tractates (here Theodicea must be mentioned) as in his scientific correspondence. Anyhow, Leibniz became a source of inspiration for many philosophers in most various areas of knowledge.

Leibniz's ideas have influenced, through his disciple Christian Wolff (incidentally, Lomonosov's teacher) on certain German literary critics in 18th century, especially on A.G. Baumgarten, Wolff's disciple considered a founder of aesthetics as a separate discipline), but this tradition fade away soon. Dolezel quite consciously resolved to restore it in the context of modern science.

In general resurrection of logics of possible worlds in philosophy is associated with two names: Alexius Meinong and, already in our days, Saul A. Kripke. Since Kripke's article "Semantical Considerations on Modal Logic" was published in 1963 (Acta Philosophica Fennica 16 (1963) 83—94), various versions of the semantics of possible worlds begin to evolve in the modern analytical philosophy. As the title shows, in his article Kripke turned to the concept of possible worlds for the sake of modal logic (we are leaving now the topic of modal logic as we'll have to deal much with it further on).

Among the various semantics of possible worlds two concepts may be noted as extreme: Kripke's own and that of David Lewis.
Kripke considers "our" only real world as one of the elements of multitude of logically possible worlds. Thus, Kripke's possible worlds are abstract notion intended for interpretation of statistic nature of regularities of the one and only real world.

Such, abstract and not real, possible worlds provide decent "stuff" for building up the logic of possible worlds of literary fiction: therefore Dolezel with his general ideas of semantics of possible worlds remains more or less within the limits of Kripke's approach, which is, we should say, most popular in the contemporary philosophy.

However, the notion of a single "actual" world is by no means as obvious as it might seem. Difficulties that literary criticism using multiworld semantics encounters with this notion were pointed out neatly by Elena Samino, who also paid her attention to the difficulties resulting from attempts (characteristic for Dolezel, see above) to differentiate precisely between fictional and non-fictional narratives (Samino, Language and World Creation… 83—84: "What is taken as “actual” is not an absolute notion, but is dependent on historical, cultural and ideological facts"; compare with ibid., 83—85, 108—109).

David Lewis in the early 70's declared an opposite position that he called "modal realism". The most mature and thorough account of it is given in his monograph The Plurality of Worlds (Oxford, 1986).

"Modal realism" is defined there as "The thesis that the world we are part of is but one of a plurality of worlds, and that we who inhabit this world are only a few out of all the inhabitants of all the worlds» (p. vii); and further: "Absolutely every way that a world could possibly be is a way that some world is…" (p. 86). In the early 70's Lewis was nearly the only one of the professional philosophers who adhered to such views, but already in a decade they became extremely wanted (the monograph we have just quoted was formed on the basis of public lectures given in 1984).

There must have been a sort of correlation with the change of generations among the physicists that eventually led in 90's to the drop in popularity of the Copenhagen interpretation and the rise of the renovated Hugh Everett's "multiworld interpretation" as its successful rival.  It remained almost unheeded as it emerged in 1957 but was favored with intensive development since the late 60's, though already without Everett's involvement. The polls that were carried out in 90's among the physicists give different results concerning the relative popularity of both Copenhagen and "multiworld" interpretations, but even as late as in 70's it was impossible to imagine that the latter could appear as much as competitive at all;  it's remarkable that among the supporters of the "multiworld" interpretation are reckoned such prominent physicists of our time as Richard Feynman and Steven Hawking.

"Multiworld interpretation" strives to be rather a scientific than a philosophic conception, yet it undoubtedly involves philosophic thesis claiming that wave function presents quite real states belonging, however, to different worlds that exist simultaneously but can't get in physical contact between them. Thus the notion of "objective reality" is being saved but there is a dear price to pay for it: "objective realities" become innumerable.

Lewis avoided referring to "multiworld interpretation" of quantum physics in his philosophic works but in his last lecture, that actually became his philosophic testament, this taboo was lifted (D. Lewis, How Many Lives Has Schroedinger's Cat? // Australasian Journal of Philosophy, 82 (2004) 3—22; all this issue (Nr 1) is dedicated to Lewis and his uncommon views). This lecture, that was given three months before his death in 2001 and published only in 2004, suggests an explanation of the principle of "modal realism" built entirely around the interpretation of quantum mechanics. There Lewis thoroughly criticizes the multiworld interpretation", but only for logical inconsistency and inner "discrepancies". His own position is even more radical, and he draws from (his own version) of Everett's concept such conclusions as have never been drawn before.

The whole article, starting from the title and epigraph:

I really pity
Schrodinger’s kitty,

is dedicated to the hard lot of the cat from mental experiment by which means Schroedinger wanted to prove the absurdity of Copenhagen interpretation. This cat happened to share hermetic camera with the sealed vessel containing prussic acid. Nothing threatens the cat as long as the vessel remains sealed off. However, the same camera is equipped with device that can break the vessel if it detects a particle generated in consequence of quantum event – radioactive decay (we deliberately simplify the description of Schroedinger's construction). Then camera will be filled with vapors of prussic acid and the cat will die. The essence of experiment is in the fact that the event in macro-world, probable death of cat, appears to depend directly on the event in micro-world, act of radioactive decay.

Event of macro-world is determined by the laws of quantum mechanics characterized by probability. Interpreting this probability in accordance with Copenhagen principles we arrive, in absence of observer, at the superposition of all the possible states: that is, we should assert that camera contains a kind of mixture of two cats, the dead one and the live one, being described by Schroedinger' wave function. Having opened the camera we would commit an act of observation (the so-called collapse of the wave function will take place) and then see the cat dead or alive, but in this act we would be a necessary part of experiment's reality ourselves. Niels Bohr solved this paradox asserting that in absence of observer it's impossible to speak of reality in strict sense of the word. Out of this reason he denied Schroedinger's wave function immediate ontological status.

In the "multiworld" interpretation everything happens otherwise: all the possible states are recognized as real and their mutual incompatibility is explained by their reference to different worlds. The event of radioactive decay that secures the simultaneous probability of both cats, dead and alive, is one of the acts dividing the world into two "branches": two cats eventually emerge, one of them dead, another alive. There can be no collapse of wave function because instead realization of different states takes place by the means of multiplying possible, and becoming real, worlds.

David Lewis carried this concept through to an end, - rather a cheerless end, we have to say. Schroedinger's cat, it appears, deserves pity not because of its probable murder, and even not because it "gets smeared" between the states of life and death, but exactly out of reason that it's liable to stay invariably alive in some of the "branches" of possible, and real, worlds. Any situation fraught with possible death of living creature appears to have an alternative, and this alternative, in accordance with the principle of "modal realism", is also always real. Therefore a peculiar kind of immortality emerges where a human continuously decays but does not die. This immortality reminds of "the immortals" (Struldbruggs) in the third of Swift's "Gulliver's travels" (Ch. 10), yet this one is much more excruciating. Gulliver's interlocutors were losing only hair, teeth and gustatory senses, but in fact nothing prevents the decay of living body from getting much farther. "Under such conditions eternal life turns into eternal torture" (it's hard not to recall that these are words of a man who was dying of diabetes for a year and had an implanted kidney). If wave function never collapses, life becomes eternal, but an hour may come when it won't be worth living. This, according to Lewis, should be a true reason to pity Schroedinger's cat, not the threat of death:

"Everett’s idea is elegant, but heaven forefend it should be true! Sad to say, a reason to wish it false is not a reason to believe it false"(p. 21).

Thus, we reviewed two extremities of the semantics of possible worlds. Both of them appear to "save" the notion of reality of "objective reality", that is, reality that is quite independent from the "observer".

Both extremities lead to paradoxes. In Kripke's semantics it's impossible to explain the laws of quantum physics, whereas in Lewis' semantics everything is possible, but at a cost that makes even Copenhagen interpretation look quite innocent…

However, our goal is to find such semantics of possible worlds that would be adequate to Judeo-Christian cosmologic tradition and to hagiography in particular. It's obvious that neither Kripke's nor Lewis' semantics are appropriate here.

If our conclusions are true, such semantics should be sought in accordance with Copenhagen interpretation of quantum theory. That means that we have to regard possible worlds the way Niels Bohr regarded Schroedinger's wave function: to renounce an unambiguous definition of their ontological status.

It should be obvious enough what this conclusion means with reference to the narrative of natural sciences, since nothing we say here exceeds the concept of the Copenhagen interpretation's originators: possible worlds are real in the way they "bear relation to" reality, that is, their measure of reality is proportionate to their probability values. This measure is always less than 1, since a full reality ("measure of reality" = 1) is revealed only in the situation of interaction, when it's possible to speak of the "collapse of possible worlds" (in analogy with the "collapse of wave function").

In contemporary formal logic similar ideas also take shape: see, in particular, A. Anisov’s article “The Logic of Indeterminatedness” in А. М. Анисов, Логика неопределенности и неопределенности во времени // Logical Studies. Online Journal http://www.logic.ru/Russian/LogStud/08/LS_8_r_Anisov.zip

 This is the way author formulates the ontological essence of his ideas:

"Our approach suggests that possible worlds don't exist along with a real world [as in Kripke's case – B.L.], but constitute it in their totality [here likeness to Lewis shows – B.L.]. The real world disintegrates into possible worlds because uncertainty is objectively inherent in it [whereas Lewis not just admits uncertainty, but his very idea of "modal realism" was needed to help avoiding the admittance of uncertainty as an objective feature of the real world – B.L.]. More exactly, possible worlds in our terms coincide in a certain part of the real world and differ only in regard to its uncertain part. This part is actually uncertain due to the fact that in reality it's impossible to reduce it to something particular".
We'd like to add to this reasoning just one remark: the certainty of the real world is possible only as an approximation that we could accept. The world, if we want to describe its "objective" (that is, existing without our interference) features appears to be, to say so, uncertain by definition. Uncertainty is just a price we pay for objectivity, and that's exactly what Copenhagen interpretation of quantum theory tells us at such a great length.

Usually semantics of possible worlds has been applied to the interpretation of quantum theory in order to find an alternative to Copenhagen interpretation, whereas we have used it so as to formulate in different terms the same Copenhagen interpretation. That's why we arrived at the notion of "degree (measure) of reality” that assumes values in the interval between 0 and 1 (this notion will be explained more accurately below). Let's say again, however: our speculation holds nothing that would seem new to contemporary philosophy. As a matter of fact it's trivial and just as reliable as Copenhagen interpretation itself.

However, we claim that this semantics of possible worlds can be applied to interpretation of historic and fictional narratives as well. And this declaration is neither trivial nor obvious: moreover, this text has been actually written in order to base it. There we'll have to consider possible worlds of both historic and fictional narratives (mainly on the base of hagiographic material) and then reach the opportunity to compare the layout of these worlds to the possible worlds of natural science.

Let's notice, running a few steps forward, that it's exactly the homogeneity of the internal layout of possible worlds of all three narrative types that allows them to get along peacefully together within the limits of Judeo-Christian narrative tradition, both apocalyptic and hagiographic. As for the ontology of possible worlds described here, we will have yet to consider it on another level below.

The principle of ontological homogeneity and rigid designators

There are two possible approaches to the semantics of possible worlds. What we propose is so-called "possibilism": none of the possible worlds are designated a priori as real but all of them are considered as "possible" (however, the term of "possibilism" isn't very accurate, since we speak of the uncertainty of possible worlds: see Anisov's article mentioned above; whereas uncertainty may be caused not by an element of accidence but by so-called illegibility inherent in phenomena: see below). Kripke and Lewis' approaches are "actualism": a certain world (as in Kripke's case) or all the worlds (as in Lewis' case) are considered real and not just "possible". Dolezel joins "actualist" ontology proclaiming completely fictional nature of his possible worlds of fictional narratives: they are not real and not even "possible" but appear to be exclusively mental constructions (р. 13).

We have just now voted for "possibilist" interpretation of possible worlds of all three narratives. If for the part of the narrative of natural science we can be backed at least by an authority, though not indisputable, of the Copenhagen interpretation, and for the part of the narrative of social history a glimpse of something intuitive, if nothing else, can be caught (especially after what we've said above; from the point of view of formal logic see also important notes in Anisov's article), but regarding the narrative of literary fiction our thesis wants explanation even on the preliminary stage.

Thus, why are we not ready to join Dolezel so as to consider possible worlds of fictional narratives completely fictional?

This text is actually intended to give a comprehensive answer to this question. We will examine only one problem that remains unsolved in Dolezel's narratology but in our "possibilist" narratology it doesn't emerge at all. This question is closely related to the needs of critical hagiography as well.

Dolezel formulates it with the means of Kripke's notion of "rigid designator" that Kripke himself relates to proper names (p. 18—19, 233—234; see: S. Kripke, Naming and Necessity, Cambridge, Mass., 1980). "Rigid" (contrary to "non-rigid" of "accidental") designators, according to Kripke's definition are those "designating the same object in each of the possible worlds".

Thus, for example, rigid designator "Napoleon" designates at the same time historic Napoleon, Napoleon of "War and Peace" and, say, Napoleon from the play by German expressionist Georg Kayser where Napoleon manages to escape from the island of St. Helena and proceed with his career in New Orleans.

In the traditional ("one-world") literary criticism scrutinizing literary works featuring characters from the real history is fraught with essential difficulties. Assuming that fictional characters act along with historic ones such literary criticism cannot keep any notion of homogeneity of literary work. This may mean, for instance, that different sorts of semantics are requested for Napoleon and Pierre Bezukhov. Dolezel's position in this sense is very strong, since for him Napoleon of the "War and Peace" and Pierre Bezukhov are just characters referring to one of the possible worlds of fictional narrative.  

Dolezel therefore formulates the principle of ontological homogeneity of the possible world of fictional narrative. He considers this principle fundamental and it appears still more fundamental for us.

However, Dolezel's approach meets its own difficulties, none easier than those arising before the "one-world" literary criticism.

Dolezel has to assert that "Tolstoy's Napoleon is no les fictional than his Pierre Bezukhov". Dolezel couldn't say otherwise since he decided to proceed from the absolutely fictional nature of the possible worlds of fictional works. Since the principle of ontological homogeneity takes effect, all the characters of the fictional work should be equally fictional.

But here inner logic of Dolezel's reasoning makes him contradict the facts both well-known and obvious: Tolstoy wrote his novel as a kind of historiosophical research or, more precisely, a certain mental experiment in the field of history. The author of "War and Peace" himself pursued the same goals as does historian at his professional work: although instead of academic methods he chose what we have just called a mental experiment. In particular, Tolstoy tried to restore, though with the specific means of literary fiction, exactly an image of historic Napoleon. Actually, we have no slightest reason to reckon (at least, a priori) that Tolstoy managed to cope with this goal less successfully than, for instance, the author of historical monograph about Napoleon, E. Tarle: whom nobody, however, would blame for fiction, - at least, for literary fiction.

If Dolezel thinks that restoration of the historic image of Napoleon in "War and Peace" can by no means be as reliable as the analogous (by goals) work by Tarle, he should be expected to prove it. In particular, it would be necessary to argue why author's subjective intentions (his claim to build up historiosophical research in the form of a novel) could bring to nothing but creation of the next literary fiction.

However, Dolezel not only gives up such explanation but sinks into still deepening contradiction to his own principles. He has to carry the restriction of principle of ontological homogeneity in regard to fictional works further on.

Running a bit forward, we'll say again that for us, contrary to Dolezel, the principle of ontological homogeneity will remain intact. And in particular, as it regards rigid designators, we will insist on the solution that attributes the certain, non-one and non-zero at a time, "measure of reality" (see below) to Tolstoy's Napoleon, Tarle's Napoleon and Pierre Bezukhov at once. As for Tarle's Napoleon, it should be clear, after all that was said above about historic narrative, that, for all the respect to historian's labor this Napoleon (narrative substance "Napoleon") should not be mixed with Napoleon of historic reality. However, in regard to Tolstoy's Napoleon, it should be noted that, for all the possible disagreement with author's historiosophical views this Napoleon cannot be ultimately denied historic reliability. At last, in regard to Pierre Bezukhov, the question concerning his ontological status is most interesting, and we will deal with it later on.  

"Theoretical digressions" and their truth conditions

The problem of historic characters in the fictional work is like a hatch, not very large by the sight of it but hiding an abyss behind it: an abyss of quite real and even physical reality that no fictional world could contain.

At the first glance, there exists a ready solution of this problem which Dolezel actually tries to use (p. 27). It has been developed yet within the frame of classical, "one-world" literary criticism.

The unquestionable fact that fictional work can contain explicit statements about the real world is reduced by Dolezel to what Boris M. Eichenbaum whom he quotes called "theoretical digressions" (in the monograph "Young Tolstoy", 1922). According to Dolezel, these "digressions" don’t refer to the fictional world of literary fiction but to our real world, whereas, while analyzing the possible worlds of fictional worlds they should not be taken into account.

Already at this point one question arises: why can authors of fictional works who construct, according to Dolezel, absolutely fictional worlds of their own fiction manage these "theoretical digressions" referring to our real world? How is such procedure logically possible?

Representatives of the "old" literary criticism didn't have to face this problem or rather it hasn't been too acute. After all, in the "one-world" model everything refers to our "actual" world in this of other way, fictional images and non-fictional reasoning alike. John Searl, whose article Dolezel quotes (p. 236—237, n. 45) as an example of considering this problem with regard to the contemporary philosophy of language and narratology, is no exception.

Following the "one-world" notion of fictional literature and unconcerned with the principle of ontological homogeneity of the world of fictional work (since he admits no particular world of fictional work), Searl can easily afford differentiating between fictional and non-fictional fragments inside the fictional work. Even so he encounters no logical contradictions:

"Sometimes an author of fictional narration puts into the narration statements that aren't based on fiction and aren't part of the narration. Let's consider the most well-known example: Tolstoy begins "Anna Karenina" with a sentence: "Happy families are all alike; every unhappy family is unhappy in its own way". This, I think, is a statement made in earnest and not based on fiction. This is a true assertion. It is an integral part of the novel but it takes no part in the fictional narration. When Nabokov opens "Ada" with an intentionally wrong quote from Tolstoy, saying: "All happy families are more or less dissimilar, all unhappy ones more or less alike", he contradicts Tolstoy by implication (and chaffs him as well).

Both statements are true assertions, even if Nabokov's statement was made in the way of ironically quoting Tolstoy. Such examples compel us to make the last differentiation: between fictional work and fictional discourse. Fictional work doesn't necessarily consists entirely (and usually doesn't consist entirely) of the fictional (based on fiction) discourse".
Searl also couldn't escape the difficulty of explaining why proceeding from or, at least, in connection with the discourse that is "based on fiction" it's possible to make "true statements". Yet this is a common problem of the "one-world" semantics of fictional narrative. In Dolezel's case this problem is not only being solved but also worsened: he still needs different semantics for different parts of fictional work in order to analyze it, and yet no logical formulas are suggested to ensure their interrelation. Still worse, Dolezel's poetics is being threatened by the more problematic question of the "firmness" of fictional work, that is, a question of firm basis for integrity of a building made of materials so heterogeneous and fastened in a fashion so obscure (if fastened at all).

However, detailed scrutiny shows that the very Dolezel's observation, regarding ontological heterogeneity of the world of literary fiction and the world of "theoretical digressions", isn't in line with the facts.

Traditional literary criticism from Eichenbaum to Searl did not consider question of distinct separation between fictional and non-fictional discourse within the fictional work crucial (let's recall that both discourses were considered related, though in different ways, to our single "actual" reality). However, for Dolezel it's a question of defining boundaries of the fictional world of a fictional work within this very fictional work.

The absence of a distinct separation between "true assertions" and fiction may prove fatal for the philosophy that places nothing between "actual" and "fantastic".

Let's consider it in detail, remembering what we've already said concerning the fallacy of differentiation between I-texts and C-texts suggested by Dolezel.

Dolezel, in accordance with his notion of "I-texts", introduces for Eichenbaum's "theoretical digressions" the notion of "I-digressions" (p. 27). In the cases that he considers the boundaries of such digressions are quite evident, though at times these digressions take too much place and even start living their own life. Thus, Dolezel examines aesthetical essay written on behalf of the main character of Huysmans' "A Rebours" to which Remy de Gourmont refers as a work by this fictional character and not by Huysmans himself (as a closest analogy within the margins of Russian literature, the treatise on Chernyshevsky written on behalf of the main character of Nabokov's "The Gift" may be referred to). Eventually Dolezel asserts that "I-digressions" are but "fiction’s means of judging the affairs of the actual world», but their very autonomous being as "digressions" proves that the fictional text as such doesn't interfere in the actual world's affairs (p. 54).

Whereas the first half of Dolezel's thesis doesn't raise doubts (of course, these "digressions" are but means to judge affairs of namely the actual world), its second half where "digressions" are opposed by this feature to the fictional text as such is rather vulnerable to criticism.

We won't discuss in detail the question how can reasoning about the world so alien (according to Dolezel) to fictional work as our actual world find room within the limits of the fictional work without finally damaging and destroying it. Let's consider other difficulties that are not few.

Is a boundary between "digressions" and fictional texts as such always as evident as in the case of Guysmans and Nabokov?

Here is an example a bit more complicate (for Dolezel's concept): Andrey Bitov's "Pushkin House" (1971; editio princeps 1978 in the USA; my thanks to Nune Barsegian for this reference). In the text of this novel main character's articles on Russian literature are included to their full extent; these articles were even published separately (see the Russian Edition in А. Битов, Статьи из романа, М., 1986) before censorship in USSR would allow publishing the novel itself in full. There is no doubt that these articles include for the most part the thoughts on Russian literature that are dear to Bitov himself. However, the context of the novel shows that the main character's views on Russian literature, and of course on this life at large, aren't identical to those of the author at all; moreover, the author maintains critical view on his character as well. It's obvious that were these articles written not on behalf of "Levushka" but of Andrej Bitov, they would be quite different.

Where is the boundary between I and C here? It turns out that we have to speak of fictive I-digressions, so that we arrive at the contradictio in adjecto.

Fictive nature of I-digressions in the modern literature can be even more evident, as shown by the instance of Milorad Pavic's Dictionary of the Khazars which is constructed as essentially faithful, though supplied with modern additions, restoration of the scientific work that never existed in reality, Dictionary of the Khazars (Lexicon Cosri), allegedly published in 1691 in Latin by the learned man Daubmannus and destroyed in 1692. Judging by its content, this restoration indeed contains scientific data and appears, to the extent of author's scientific erudition, to be rather adequate introduction to the study of Khazars. The possibility of reading Pavich as educational literature is, maybe, not the most optimal solution, yet it scarcely raises any doubts even within the scientific community.

In the case of Pavic’s famous Dictionary of the Khazars it's particularly hard to distinguish the "digressions" in the text. If these are "digressions", what do they digress from? After all, if we leave out everything but the correspondence of fictional characters, then all the plot ties in the text will be lost. It's worth to mention that in the novels of Huysmans and Nabokov, and even of Bitov scientific works of their characters are set apart from the development of the plot.  Whereas Pavic, who retains all specific features of scientific account, just erases the boundary between the "digression" and the "fictional text": his "theoretical digression" is fictional and "fictional text" is theoretical… Dolezel's theory dismisses such possibility, and this is a serious reason to put the theory in doubt.
The most important thing here is that we get an occasion to raise the question in regard to "theoreticity" of text in general. Isn't the reason fictional texts can so easily incorporate "theoretic digressions" referring to the actual world's affairs, that these "digressions" may be defined as "digressions" only in terms of style but not of philosophy?

On the philosophy level Dolezel formulates the difference between "digressions" and literary text as such through the notion of truth conditions: «Because they obey truth-conditions different from the fictional text proper, I-digressions cannot be used as a source for a semantic of fictionality» (р. 27).

However, this is just more accurate formula of what was just said. The fact is that truth conditions for "digression" and for fiction differ, according to Dolezel, inasmuch as in the first case they refer to the "actual" world and in the last case to the fictional possible world. If that is so, then indeed they cannot be similar.

However, this Dolezel's thesis misses an evident fact: all the truth conditions that in the "actual" world could be applied to "theoretical digressions" in literary fiction are in no way different from the truth conditions in the respective fictional possible world. Thus, in the "fictive" (according to Dolezel) world of "Anna Karenina" "every unhappy family is unhappy in its own way" and it raises no doubt, whereas it would be quite pertinent to doubt the verity of this thesis for the "actual" world, and that's precisely what Nabokov did: so that in the possible world of his "Ada" this thesis, having been reversed, may also be considered an undisputable truth.

These two examples, especially after we, following Searl, put them together, show something that Searl missed and Dolezel following him consistently rejects: "theoretical digressions" and their truth conditions belong to possible worlds of respective fictional works, whereas their plausibility in the face of the "actual" world of readers is but an artistic move ("semiotic channel", using Dolezel's expression) involving reader into fictional possible world. Of course, this is by no means the only method of revealing possible worlds of fictional works but it may be considered one of the most important.

We have gradually come to the conclusion that Dolezel's principle of ontological homogeneity of fictional possible world works much more effectively than its discoverer guesses. "Theoretical digressions" are not digressions in terms of philosophy since the very differentiation of I-texts and C-texts is not philosophically grounded. Let's repeat that the case in point is not the difference of text types but of functions inside the same text.

However, we still can't appreciate the scale of the phenomenon we presently scrutinize.

 Literary fiction as a virtual construct and mental experiment

The observations we made above might face the following objection: Bitov, and especially Pavic, are "postmodernism", whereas proper literature is quite different… If such an objection were just it would presume not only essential restrictions in regard to the fictive nature of possible worlds of a fictional work but also essential conceptual difficulties with carrying these restrictions through, that is, differentiating between "right" and "wrong" literature. 
Fortunately, this objection is wrong, too. Let's turn therefore again to the "proper" literature, that is, such as cannot be suspected as postmodernist.

Let's turn, however, to the "War and Peace".

We have already noted that a fictive nature of Tolstoy's Napoleon isn't a simple matter. Although this Napoleon differs from the historical one, he is nonetheless a "narrative substance" that a priori is none the worse than the other "narrative substance", Tarle's Napoleon. And what's about Pierre Bezukhov? Is he, at least, fictive?

The answer depends on the point of view.

Pierre Bezukhov is a logical object to which certain qualities are attributed, and which displays these qualities while interacting with other objects of the possible world of Tolstoy's novel. In these interactions only some of the qualities of the logical object "Pierre Bezukhov" are primordially known, whereas others are being revealed in the process. They are being revealed since we have a prior knowledge of the laws inherent in the possible world of Tolstoy's novel. However, not all of these laws are primordially known to us. The novelist's plot was actually intended to reveal them by observing the interactions within this world of logical objects whose qualities are for some part primordially known and for some part are to be revealed. Experimental material found in the novel is being processed in a similar way to solving the system of many equations with many unknown quantities. As a result, this way allows to draw conclusions in regard to qualities of the possible world of the novel, including qualities of its possible logical objects, among them Pierre Bezukhov.

In the previous paragraph we have just given a more extensive statement of our thesis that the novel "War and Peace" is a mental experiment. Speaking yet more strictly, while just slightly paraphrasing what Tolstoy himself says in his "theoretical digressions", the novel includes detailed description of a mental experiment, its analysis and an account of conclusions, both in a stringent and a popular manner. An attempt to reconstruct Napoleon's personality is only a part of this mental experiment that is basically dedicated to the study of historical laws of social development.

What can be said then of Pierre Bezukhov's reality?

If we chose to understand "reality" as his possible correspondence to a certain historic person, then Pierre Bezukhov is fictive. We can't regard the proper name "Pierre Bezukhov" as a rigid designator: even if a certain historic person with such name were found, it would be incorrect to say that the designator "Pierre Bezukhov" refers to the same object in all possible worlds. Two different persons named Pierre Bezukhov, one of them from Tolstoy's novel and the other from historic reality, would appear to be two different objects but not two "incarnations" of the same object in two different worlds, real and fictive. Dolezel's notion of transworld identity (p. 17) can't be applied to Pierre Bezukhov as an integral person, yet it can be applied to Napoleon (or, more precisely, to all Napoleons: whether Tarle's, Tolstoy's or Stendhal's, or Napoleons of various legends and stories including that about New Orleans). That's why the position allowing considering Pierre Bezukhov a purely fictional character is quite legitimate. That's exactly how Dolezel considers him.

However, the position allowing considering Pierre Bezukhov in a different manner is no less legitimate. From such point of view the most important thing is that Pierre Bezukhov takes part in a mental experiment. Recalling the mental experiments carried out by physicists we would suggest that Pierre Bezukhov bears in this regard the closest resemblance to Schroedinger's cat. Schroedinger's experiment, just as in "War and Peace" does, presumes that nothing that occurs would raise doubts concerning its theoretical and practical possibility in terrestrial conditions. At the same time, in spite of the presence of technical possibility to carry out the experiment in fact, and not only in theory, no cat suffered in various versions of Schroedinger's experiments performed by various scientists: all these cats invariably remained fictitious.  

Should we say that Schroedinger's cat is as unreal as Pierre Bezukhov? We can say so but physicists will scarcely show interest in this literary aspect of Schroedinger's cat's "non-reality". "Literary" argumentation wouldn't help us to approach problems inherent in physicist's mental experiment. Therefore physicists have to treat Schroedinger's cat as a real one.

In a like manner, if we wish to understand (even in order to challenge) Tolstoy's historiosophical concept, we will have to treat fictive Pierre Bezukhov as real. However, we will have yet to discuss reality and fictivity status of characters of literary fiction, after introducing more necessary notions.

Inner affinity of mental experiments in physics and literary fiction has been noticed long ago and is strenuously exploited by science fiction authors who derive their plots, or at least starting points of plots, from these experiments. However, literature theoreticians somehow overlook this field where discourse of natural science and fictional discourse so obviously cross.

Nevertheless, we have to draw for literary criticism a conclusion in regard to undisputable (up to date) ability of mental experiments to provide us with real knowledge of things.

Since mental experiments of natural scientists don't appear to be a pure fiction but describe certain aspects of reality, "mental experiments" made by authors of fictional texts also describe reality, though obviously not all of them do it in equal measure (still, mental experiments of natural scientists also tend to be different, and sometimes even fallacious).

Thus, we arrive at the different formula of the thesis we suggested above regarding ontological status of possible worlds of fictional narrative, thesis stating that these worlds are not purely fictive but endowed with a certain "measure of reality" that cannot be determined a priori (that is, in particular, cannot be determined a priori as zero, as Dolezel does).  

Ontology of possible worlds: systematization and conclusions

At a certain stage we will have, following Dolezel, to consider his scheme describing fictional work in  "the way it works" (Russian Formalists wrote about "the way it is made", whereas Dolezel writes rather about "the way it works" than just "the way it's made). However, our main subject is not a literary fiction but hagiography that involves not only fictional but also historic narrative and even a narrative of natural sciences. Therefore, we can't try and apply Dolezel's theory to hagiography without clear notion of mutual relations of all three narratives. Hence let's present all that was said hitherto in this text in a more systematic form.

Thus, we have compared literary fiction with mental experiment but only on the basis of so-called "realistic" literature and of such experiment that could be carried out in practice. Nonetheless, mental experiments don't necessarily have to be practicable, whereas literary fiction doesn't necessarily have to be "realistic" (no matter what we choose to understand as "realistic").

Very often mental experiments present impossible to realize though theoretically possible situations. For instance, so-called twins' paradox that is often used to illustrate the relativity theory presumes that a human should move with near-light velocity which practically impossible though theoretically acceptable. In this experiment twins are figures yet more fictive than Schroedinger's cat (that can at least practically exist) but it doesn't prevent this experiment from discovering important features of the real universe.

The closest analogies to such mental experiments that may be found in literary fiction are characters and situations theoretically possible yet not found in daily life. Such characters and situations are typical, for instance, for Dostoevsky (blamed for "anti-realism" by many critics who asserted that his characters are a fruit of morbid imagination). However, the contemporary literary criticism holds almost in a single voice that by the means of his unusual characters Dostoevsky managed to tell about reality something that no one could tell in other way. It looks like a complete analogy with mental experiments like twins' paradox, theoretically possible but practically unfeasible.

Finally, there are mental experiments that not only practically unfeasible but also theoretically impossible. As an instance, a journey through time in Goedel's universe may be named (technical impossibility is unquestionable yet it's still under question whether it's possible theoretically), as well as measuring distances in the universe by means of "telescope" sending signals in the shape of loops engirdling the torus-shaped universe (it may be theoretically possible only if at least the main these of the superstrings theory is correct). All these experiments deal with a possible world where a certain assumption, whose applicability to our world is under question, is considered true together with a number of other assumptions that are unquestionably incorrect for our world (these are assumptions causing things that in our world are practically unfeasible to become practicable in the possible world of mental experiment). Nevertheless, such experiments contribute to the study of reality, too: at least by the means of clarifying certain interrelations of phenomena inherent in reality.

In literary fiction such experiments may be associated with the possible worlds featuring creatures whose existence is virtually impossible in our world. Very good examples are Lewis Carroll's "Alice in Wonderland" and "Through the looking-glass". Their value in this sense is still higher because they actually appear to be not just fictional narratives but scientific treatises in disguise of fictional narratives, consisting entirely of mental experiment in strictest sense of the word. Other examples are Swift's "Gulliver's Travels" and all the literary anti-utopias of 20th century: these are true mental experiments in politology and sociology, though characters of these novels may not be humans (for instance, in the cases of Swift and Orwell). However, utopias also draw up a long row of mental experiments, more or less fictional: from Plato's "Republic" and Campanella's "The City of the Sun" to some soviet science fiction authors who wrote about the Communist future (David Lewis wrote on the verity of literary fiction by analogy with philosophical treatises in his "Postscript C: Fiction in the service of truth" (1983, written on reissue of his article "Truth in fiction").

This kind of works where scientific goals (world-imaging, according to Dolezel) are pursued along with literary aims (world-constructing) can refute any theory of literary fiction where the notion of fictionality is identified with the function of world-constructing.

On the other hand, such theories of literary fiction are refuted by the mere presence of the world-constructing function in scientific (including natural science) texts. That's why we have been so far giving only examples of mental experiments; yet we intend to show in this text that the matter cannot be reduced to mental experiments, and, moreover, that no science – neither history nor natural science - can do without constructing its own scientific worlds (however, we spoke of history in this connection above).

The presence of the world-constructing function in scientific treatises contributes to their ability to assume the form of a fictional work. Once it was considered quite acceptable for a scientific work, and only in 19th century this fashion has been abandoned. Plato's dialogues are classical example of this form, presenting a genre that survived until the New Age. This genre was deemed legitimate even for natural science, to say nothing of philosophy and logic.

As late as in 18th century equal rights of fictional and scientific discourse in regard to scientific discussion haven't been entirely forgotten. Thus, Voltaire replied to Leibniz (on his "Theodicea", 1710) by the means of a "philosophic tale" "Candide or Optimism" (1759) where Leibniz himself features as a ludicruous character.

Theory of literature where fictional discourse is opposed to scientific on the basis of a certain qualitative distinction is but a self-justification of a certain literary practice exercised in European culture of last centuries; actually, this is a sort of literary manifest, not scientific theory of literature. That which corresponds to the self-awareness of certain literary circles (that is, first of all, authors and critics) is not enough yet to be considered theory of literature.

Postmodern, having transgressed the convention of 19th century literary fashion, partially restored natural equilibrium between literary and scientific discourse. We mentioned it above, but much more striking examples could be found. It's hard, however, to find an example more striking than Jorge Luis Borges, unquestionable classic of Postmodernist tradition (significance of his work for the philosophic analysis of literary fiction was particularly noted by Vadim Rudnev; our attention, moreover, was drawn to it by A. Muraviev).

The mere, in Borges' case, obligatory intertextuality doesn't agree with the theories of "fictionality" that expect it to be supplied with a chemically pure world-constructing function: after all, any quotation claims interpretation, that is world-imaging. Even if this world-imaging chooses fictional worlds of literature as its subject (which in Borges' case is not necessarily so: his "intertextuality" extends to philosophy and even natural science as well). However, intertextuality is a theoretical issue too important to be treated in haste. We will deal with it separately.

Borges allows us to trace the destiny of a concept born quite conventionally by 20th century science but then refused by scientific community and adopted, due to Borges, by literary fiction. We speak of the Dunn's multidimensional time concept.

At the moment of publication (1927) Dunn's concept appeared no more and no less "bizarre" than many others hypotheses of the time, but already in 30's it became clear that this hypothesis shouldn't expect any particular demand. Theoretically it says nothing of the further destiny of the hypothesis. Nevertheless, Dunn's hypothesis caught the sight of Borges who dedicated to it a theoretical essay in his "Other Inquisitions" (1952) and, moreover, built on its basis plots of some of his fictional works.

In Borges' case literature chooses to act in a sphere that in a certain way "adds" to the scientific sphere of action. And this is typical, let's emphasize, not only for Borges, not only for Postmodern, not even only for the contemporary literary trends like science fiction, but for literature in general. Not for nothing we suggested the examples of Voltaire's "Philosophic Tales", and Swift; yet we could refer to the more ancient instances like "historical novels" about Alexander (not hagiographic ones) and even "The Golden Ass" by Apuleus who expounds his very definite theological, philosophical and ascetic ideas under disguise of an amusing story (that ends up very naturally with a liturgical description of an initiation ritual in the mysteries of Isis).

We will have yet an opportunity to discuss in detail the other side of the considered phenomenon: of the function of human fantasy in natural science; it cannot be reduced to construction of mental experiments that has been scrutinized here only because an analogy with literary fiction is too obvious in their case. In the meantime we just want to point out the wide-known fact that somehow remains unheeded only by literary critics: any scientific theory can be applied only to ideal situations that can be found nowhere but within a human mind. 

Real situations always appear to be more complicate, whereas a good theory can be considered good only as a reasonable approximation used for their description. Even if a scientific theory doesn’t operate with such abstractions as, for instance, "ideal gas", it anyway presumes excluding certain factors that cannot be excluded in reality. Therefore absolutely every theory of natural science applies to a possible world where those and only those factors are relevant that this theory intends to illustrate. It often happens that a main body of theoretician's work amounts to building an appropriate possible world in a proper way.

Until we started to examine how possible worlds of various narratives are organized, we can't say much in this regard: yet something, and maybe most important, can be said right now:

Any human cognitive activity leads to constructing of possible worlds.

Literary fiction is one of the sorts of cognitive activity. This is not its single function, yet necessarily inherent in it.

It would be reasonable to expect that a structure of possible worlds emerging as a result of cognitive activity appeared similar regardless of the sort of cognitive activity that determined their construction. So far we have found that this thesis can be indirectly confirmed by the fact of impossibility to differentiate strictly between the possible worlds of three narratives: fictional, historic and that of natural science. In order to prove this observation strictly we'll have to examine an inner structure of respective possible worlds.

Ontological status of such possible worlds is indefinite. This conclusion directly follows from the fact that speaking of possible worlds we adopt a language of description that doesn't presume a procedure of observation (by the analogy of probabilistic description in quantum physics). Therefore we should attribute to each of these worlds a "measure of reality" that will be less than 1 and more than 0. Here we have to notice that the notion of "measure of reality" has been so far introduced as intuitive, so that it will have yet to be formulated in a logically correct mode (we'll deal with this task later on).

Literature as well-conspired science

Before we finally set out on our journey through possible worlds, let's make yet another stop. It is, however, unnecessary for those who, being interested in literary criticism, have a correct notion of research programs and their way of action in the history of natural science. However, there are reasons to believe that those aren't many. Therefore we'll try to review the principal theses by using different examples.

Let's regard an example of "War and Peace", a novel quite obviously constructed in a likeness of historiosophical reasoning involving a remarkable amount of very extensive illustrations, as not so typical. In literature more acceptable practice is entrusting characters, and not an author, with all the state-of-mind considerations: to the extent of impossibility to identify an author's position with a position of a certain character. The last phenomenon is typical for Dostoevsky's novel (though not for all of them). Mikhail Bakhtin in his ground-breaking "Problems of Dostoevsky's poetics" (1928) defined this method as "polyphonic novel": what constitutes an author's position is not what he explicitly states (he may actually state nothing) but the very attempt to compare different states of mind of his characters, each of them wins his sympathy though not to the same extent. Author's position in polyphonic novel appears to be closest to the position of experimentalist in natural sciences who tries, and not always successfully, to carry out an experiment that would help him to make a choice in favor of one of the competing hypotheses.  

Following the course of such experiment is most effective when laboratory journal, which means drafts, in this case, is close at hand. In regard to Dostoevsky's novel this conclusion is often can be met. As far as his drafts and private correspondence allow to judge, the most dramatic struggle between Dostoevsky as empirical scientist and Dostoevsky as psychology theoretician took place over the novel "The Idiot"; this struggle alone delayed the completion of the novel for half a year (for detailed analysis see our article on the “dogmatics of love in the late works by Dostoevsky”): В. М. Лурье, Догматика «религии любви». Догматические представления позднего Достоевского // Христианство и русская литература. Сб. 2 / Отв. ред. В. А. Котельников (СПб., 1996) 290—309, особ. 299—301). Eventually the empiric won over the theoretician, leaving him, moreover, without any theory for a fair couple of years.

Dostoevsky formulated his task in the language of religion, not of science: to picture the "Prince-Christ"; yet quite scientific aspect regarding the human psyche was present there on the background. At least, if today we consider possible to extend the notion of science to psychoanalysis and many other psychological theories offering ground to various therapeutic methods, then we should seek no reason to exclude from this notion psychological theories of Christian ascetics that directly concern the problems posed by Dostoevsky for his own ends. However, while working on "The Idiot" Dostoevsky not only lacked any knowledge of Christian ascetics is essence but also never suspected that Christian ascetics is granted with any theory (only in several years, in 1870's will he find out something about this theory). By then Dostoevsky sincerely believed that humanistic ideals of 19th century "masters of thoughts" indeed represent the Christian doctrine of man. That's why he was so persistent in his efforts to create, through the use of inappropriate materials and methods, a personality as close as possible to Christ. As his principal design gradually proved unsuccessful Dostoevsky several times "lowered the plank" but nonetheless couldn't put up to his satisfaction with any result save the most hopeless: complete and final idiocy of his character.

It might seem that the Prince Myshkin's personality has initially been so artificial that it could match any possible ending. Yet it's no less obvious that if Dostoevsky retained his right to act so willfully, he would never become Dostoevsky.

That's exactly what reminds us of a situation of mental experiment in natural sciences: certain participants in an experiment may possess qualities impossible in our world, but the results of properly conducted experiment can never be arbitrary.

Having arrived at the negative result of a mental experiment Dostoevsky the theoretician resolved to revise his theory. His initial (1867) project was intended to show, by the means of an incarnate Christian ideal, a real image of Christ that had been distorted by the West. Such indeed was his thought that resulted from the shock he experienced before Holbein's "Dead Christ" (it may be reckoned that at this very moment he started working on the novel). That was theological project but it demanded a scientific ingredient: a certain concept of a human. As a result of his failure and following two more unsuccessful attempts to create a new novel Dostoevsky concludes that the concept of a human also should be sought in the East and not in the West. Thus was conceived the "Life of a great sinner" from which all his great novels of 1870's emerged.

By the analogy with science it should have been said that Dostoevsky adopted a new research program set to compete with the European humanistic one that remained unknown to him as long as he worked on "The Idiot".

This situation is quite typical for competition of research programs and for development of science in general.

Researchers face an experiment whose results contradict the current theories. Thus, theory should be modified, for which purpose a choice should be made between two fundamentally different ways: either to modify within the limits of the same research program or to turn to another one. However, experiment itself never allows concluding which way is more promising. Such conclusion may be drawn only as a result of competition of entire scientific conceptions.

In Dostoevsky's case after the failure of "The Idiot" two ways remained: either to try modifying somehow the humanistic concept of a human in order to bring it somehow near the Christian ideal, or to turn to some entirely new sources in order to form fundamental ideas as well as (quasi-) scientific psychological theory. Dostoevsky had been trying for two years to go the first way but eventually turned to the second way.

All this reminds of science to such extent that a question arises, and not for nothing: what then is the difference between the practices of an author and a scientist?

This question isn’t idle at least because a popular answer to it is wrong. Usually it is said that an author is led by intuition and a scientist by logic. However: first, no scientific discovery can be made without intuition, whereas choosing a research program for the purpose of building up theories adequate to empiric material is a procedure that cannot by deduced from any logic; besides, if logic really works, it submits to Goedel's incompleteness theorem (1931) asserting that every system of logical affirmations should contain no less than on affirmation that cannot be proved within the limits of the said system, that is, by the same "logic". Therefore, no matter how relations between science and logic are defined, there's one thing that cannot be claimed: that science is founded on logic. Science uses logic.

And second, one must not say that an author doesn’t use logic and is stranger to scientific theories. Even if we put aside special cases of authors who consciously used in their works data from certain sciences, it always appears that some inner logic is inherent in the possible world created by various authors, and that an author himself translates beside his own will into his work such extent of awareness of scientific ideas as is peculiar to his social layer and to himself in person. 

That's why the distinctions between literature and science are rather quantitative than qualitative. They come, for the most part, from the method of presenting the material and from the degree of logical stringency applied to a procedure of inference. Thus, in literature conclusion may not be explicated, whereas procedure of inference may be camouflaged.

Nevertheless, since the distinctions are quantitative, there can be no strict differentiation between literature and science, therefore frontier works like "Alice in wonderland", Plato's dialogues and "Philosophic Tales" will ever spring up – not to mention popular scientific literature like Abbott's "Flatland".

We need this comparison of logical structures of scientific and literary text because our hagiography involves both of these discourses at a time. If both discourses have a homogeneous internal organization, then the possibility of such combination won't raise any unnecessary questions. But if we followed the views dominating the contemporary narratology, then the task of building up common logical theory of hagiography would appear theoretically insoluble. That's why we have to care so much for a correct theory.

It's necessary to say, for justice's sake, that the point of view on fictional creative work as a variety of science (or otherwise, on science as a variety of creative work) that we stated here isn't completely alien to contemporary aesthetics, though undoubtedly marginal. Once it has been rather sharply articulated by Otto Baensch: "…the function of art is not at all causing a subject of perception any pleasure, no matter how noble it may be, but imparting to him something he never knew before" (O. Baensch, Kunst und Gefuhl // Logos. Internationale Zeitschrift fur Philosophie der Kultur 12 (1923—1924) 1—28) – though, proceeding from positivist notions of science as "objective" knowledge. In pendant to objectivist concept of scientific knowledge Baensch introduced the concept of "objective feeling", which allowed him to speak of objective sensual substance of the world that art exists to express. However, whatever could be said of Baensch objectivist conceptions, the "hard core" of his aesthetic theory was similar to what we here speak about.

 Hagiography as a narrative of secondary sacralization

Literature: C. Levi-Strauss, La Pensee sauvage (1962, The Savage Mind, 1966); Totemisme aujourd’hui (1962, Totemism, trans. Rodney Needham, 1963); Late Eleazar Meletinsky’s classic “The Poetics of Myth” and “The Origin of Epics” : Е. М. Мелетинский, Поэтика мифа. 3-е изд., репринтное (М., 2000) (Исследования по фольклору и мифологии Востока) [1-е изд. — 1976]; он же, Происхождение героического эпоса. Ранние формы и архаические памятники. Изд. 2-е, исправленное (М., 2004) (Исследования по фольклору и мифологии Востока) [1-е изд. — 1963].

Having mentioned the affinity between literature and science we have trodden the ground where we have, by necessity, to take account of their common origin, gradual evolvement from the myth.  It's precisely in this aspect of origin that hagiography reveals its striking difference from both literature and science, for all their outward likeness. The principal difference is that while science and literature started evolving from the myth only following its desacralization, hagiography, being a narration of sacred objects and saints, became a new sacralization of desacralized, though different by its nature from the sacrality of myth.   

Hagiographic narrative may not vary structurally from historic or fictional one (and scientific as well, taking into account within the margins of hagiography also cosmologic apocalyptics, not to mention the possibility of impregnation of scientific narrative of more particular nature). However, the difference will anyway remain on metastructural level.

This difference is analogous to the difference between myth and fairy tale. Both of them may look as completely similar stories. Yet it happens that a certain tribe tells this story as a sacral one, that is, as myth, whereas other tribe tells the same story as an ordinary fairy tale, bereft of a sacral content.

In the process of desacralization of the myth it gave birth to two folklore genres: fairy tale and epos. Epos kept the etiologic and historic components of myth but they consequently started to refer to some formative period, actual for a given environment, instead of referring to the beginning of times, the world's creation. For instance, to the times of ‘Troyan War’ or ‘Vladimir the Bright Sun’. On the background of such a present-day perception of epos, classification of hagiographic genres according to Delehaye looks only more persuading: Passions epiques are exactly epic martyrdoms and not anything else, and not for nothing they are related to the "formative period" of martyrdom, to the times of Decius or Diocletian (it's remarkable that the first year of Diocletian's reign, 284, heralded the beginning of the "Era of Martyrs", the starting point of Alexandrian Christian chronology).

Fairy tale as a genre, of course, also isn't alien to hagiography, but it usually follows "in the wake" of epos and submits to its goals. This genre becomes especially productive when it concerns the origins of relics, particularly insignia of royal power (let's recall that fairy tales often tell of hero winning a kingdom). For instance, ancient Russian Narration of the Babylonian kingdom is a hagiographic legend telling about origins of insignia of the royal power, and also about geopolitics (ties between Russia, Byzantium, Georgia and, as we suppose, Varangians), however, it's a typical fairy tale by its form: royal insignia appear to be bestowed in a place that is hard to reach and is guarded by typical fairy tale dragons (that is, we deal here with a fairy tale about dragon guarding a treasure). However, hagiographic legend, for all its formal likeness to a fairy tale, isn't a fairy tale anymore, just as it doesn’t become a myth (in archaic sense of the word). It isn't a fairy tale because it becomes a sacral narration (whose sacrality, besides, is endowed with quite evident political and ideological meaning). However, this sacralization differs from the sacralization of myth. Maybe, it would be appropriate to call such sacralization relative: contrary to the absolute sacrality of myth in archaic religions or to Christian dogmatics.

Relative sacrality, as well as veneration of sacred objects and saints in general, matters in Christianity only in a context of cult as a whole. Thus, sacral meaning appears here to be borrowed but not authentic as in the case of the symbols of myth.

All we have just said of fairy tale can be applied to a later literary genre also used in hagiography, that is, a novel. For example, "Life of St. Eustathius Placida" is written strictly in accordance with the standards of antique novel. Engaging plot of the novel assumes a new sacral meaning as it gets into hagiography. In the case of St. Eustathius Placida this meaning implies service to the Cross as a new state ideology (namely ideology of state, not of general Eustathius alone). It was necessary to accomplish the vision of the Cross to the emperor Constantine with an explication for broadest masses.  

It's also well known that secondary sacralization may easily become a subject to secondary desacralization in the context of the Renaissance and the New Age culture. Renaissance painters demonstrated this possibility somewhat earlier and more persuasively than authors, but authors eventually came up with them: secular scenes and hagiographical plays became quite common practice almost everywhere in 18th century; Pushkin read with enthusiasm Slavonic Menologia as a collection of fairy tales, Flaubert borrowed plots for his French novels from ancient hagiography (and wasn't the only one who did so at all).

Here is, however, what we'd like to emphasize as a principal conclusion from this discussion of the affinity between hagiography and genres of literary fiction as well as archaic myth: all of them exploit the specific mode of thinking that was never gone since humanity denounced archaic forms of religion, and still retains its ruling position in almost all the spheres of human activity, being only partially supplemented with a scientific analytic thinking.

Regretfully, experts in archaic mentality, whose peculiar laws were first paid due attention by L. Levy-Bruhl, and subsequently were extensively described by E. Cassirer and C. Levy-Strauss, never showed particular interest to history and methodology of science. Therefore there are no somewhat extensive researches in this area that we could rest upon, so we'll have to content ourselves with general observations.  

Raising a question in regard to literature as a special form of scientific knowledge is necessary now owing to a change of the common idea of scientific knowledge as such.

After Levy-Strauss it's easy to say that much in literary fiction appears to be founded on the specific logic that for a first time in history revealed itself in archaic mentality. We have learned now that:

"The mistake of the naturalist school was that they supposed that natural phenomena are what myths try to explain [in Russia the statements of this school were shared by the founders of folkloristics and even of literary criticism in general: A. N. Afanasiev (1826-1871) and F. I. Buslaev (1818-1897), Veselovsky's teacher], whereas natural phenomena are rather means by which myths try to explain realia that don't referring not to natural but to logical order" (C. Levy-Strauss, The Savage Mind).

However, after Lakatos and Quine (whom we'll discuss below), it's almost obvious that a theoretic science plays the same role. This had been taken into account neither by Levy-Strauss nor by his successors. It's obvious, actually, that research programs are realities of a logical order, yet only building up research programs lets our theories, invented in order to explain natural phenomena, be what we perceive as cognition: cognition both of physical sense and of sense in general. For all the difference between scientific and mythological mode of thinking it's by no way the difference of grounds. If Levy-Strauss is right in calling mythological mode of thinking entirely "natural-science" (at least in quotation marks), then we have to add that the scientific mode of thinking – at meta-scientific level – appears to be quite mythological as well.

This inner affinity of mythological and "natural-science" modes of thinking revealed itself in a most distinct way in Judean cosmologic apocalyptics from which hagiography descended. The difference between Judean cosmology and Babylonian myth is particularly perceptible when 1 Enoch and MUL.APIN are compared: for all the cosmological similarity, what makes these texts differ strikingly is that deification of celestial bodies, as well as polytheism in general, disappears in the Jewish apocalypse. Deviation from Babylonian mythology occurs both in natural-science and religious sense, whereas the actual natural-science content, retaining its religious significance, ceases to be actually religious, because celestial bodies are just celestial bodies now, and no gods.

On the whole, the narrative of Judean apocalyptics is an epical narrative, and rev. Delehaye appeared to be right once more when he chose to consider hagiographic apocalyptics (already of Christian age) within the margins of Passions epiques. In pre-Christian Judean apocalyptics typical features of epos strike the eye, and first of all, reference to the Mediator, symbolic figure of one of the "formative periods" but not of the mythological beginning of the universe.

Of course, epic narrative was sacralized already in Judean apocalyptics. How much appropriate will it be to call its sacralization secondary? From the historic point of view, it's possible to assert the secondarity of sacralization of epic narrative only for certain cases (first of all, for the Enoch cycle, whose prototype was legendary Shumerian high priest Enmeduranki, and for Daniel, whose prototype, Danel, is found in Ugaritic epos). It's quite probable that, for instance, in the case of Ezra there was no epic prototype at all: only historic prototype. In this case, it's possible to speak of secondarity of sacralization of epical narrative of the Apocalypse of Ezra only in a logical, but not in a chronological sense. As most of the heroes of Christian epic hagiography, Ezra of the Apocalypse of Ezra became an epic figure only at the same time as he became a sacral figure.

At any rate, hagiography emerged as a result of secondary sacralization of epos, and it happened yet before hagiography got estranged from the Judean apocalyptics.

Intertextuality, texture and hagiographical narrative substrate

It's easy "to miss" the combined presence of world-constructing and world-imaging functions in a contemporary literary text also because in the literature of the New Age a function of writing fictional text was quite ultimately opposed to a function of reading. When both modernism and postmodernism dared to infringe upon the border between writing and reading, it appeared that almost a new, special theory of literature is needed in order to to explain, in particular, posmodernist concept of text as a full-fledged part of reality and of reality as text.   

Dolezel's theory is no exception; as a result of his analysis of the technique of "postmodernist rewriting" he arrives at the conclusion that the postmodernism brings literature as such to an end: characters and other elements of the possible world of a certain fictional work are easily removed to another, no matter how different world: as easy as child plays with "Lego" parts. However, "The game is no longer exciting, and it is time to invent a new one", р. 226). With this phrase and in this tune Dolezel accomplishes his theory of literature.

Maybe, in the contemporary Western literature things indeed go so ill, but intertextuality as such is not to be blamed here. Intertextuality is a symptom necessary but not sufficient for pronouncing a diagnosis of "postmodernism" (intertextuality is a term introduced in 1967 by Julia Kristeva to designate one of the principal features of the poetics of postmodernism: strict and hidden quotation originating from the peculiar attitude to the text; see Julia Kristeva, Le Texte du roman: approches semiologique d'une structure discursive transformationnelle (The Hague: Mouton de Gruyter, 1970; repr. 1976) 209  p. (Coll. "Approaches to semiotics," 6), published on the basis of her 1967 doctoral dissertation).

However, intertextuality with its conceptual implication of the attitude to a text as part of reality doesn't take root in postmodernism and even not in modernism. It inevitably arises in any place where literary activity presents itself as, most, of all, a process of self-expression of the Tradition (and not of individual author or epoch etc.).

That's why the "postmodernist rewriting" structure that seems to contemporary theory of literature something extreme and marginal would look in traditional (from the word "Tradition" with a capital letter) literary systems a particular case of a principal method of text constructing. For instance, "postmodernist rewriting" (as literary critics of our time could call it) of the Old Testament is one of the most productive genres of "intertestamentary" period, and of the library of the Qumran community in particular.

For all the genres of hagiography intertextuality is an absolutely normative. "Intertextuality" is, in particular, nothing else than literary critic's designator of what critical hagiography designates by the term of "hagiographic substrate" (we'll discuss another function of intertextuality in hagiography below). Therefore Dolezel's theory of "postmodernist rewriting" should help us to explain how hagiographic substrate functions, and besides, prompt us once more to emphasize the difference between Dolezel's and our interpretation of the ontology of possible worlds (it's remarkable that Dolezel dedicated to this theory the Epilogue (p. 199—226), and not one of the chapters in the main body of his book. Almost all those who reviewed the book noticed that this is no epilogue but an important part of the theory; however it's possible that Dolezel intended that the "postmodernist writing" is an epilogue to the development of European literature). 

Theory of intertextuality in semantics of possible worlds is founded on the theory of reading. This isn't surprising, and in the context of Judeo-Christian tradition even obvious, as within its limits any text, openly or at least implicitly, is a commentary on a number of preceding texts, that is, every text is for some part a targum (translation to other culture's language) and for some part a midrash (commentary by means of various illustrating tales). Every new text is a commentary on the older texts, and a commentary is actually slow reading.

The outdated theory of reading presumes author's activity and reader's passivity: author creates a text and by its means affects a reader. Today there's no need to prove that the process of reading cannot be reduced to reader's passive experience of an effect exerted by an author through the text. The question is just how exactly to take account of reader's activity. Semantics of possible worlds allows doing it in a very subtle way.

An author creates by the means of his text the possible world of his work. A reader reconstructs this world by the means of the same text (Dolezel, in this case, digresses from the problem of adequacy of text understanding). However, a reader can become an author of the next text and construct the new possible world using "building materials" from the preceding one. Dolezel called this mechanism "transduction" of possible worlds; he chose a term that combines an idea of both conduction and transformation.

A possible world W2 created as a result of transduction of another possible world W1, presumes for the sake of its adequate understanding a reader's knowledge of the world W1. Dolezel himself draws a parallel with a perception of historic novel demanding that a reader, in order to understand it, should possess a certain knowledge borrowed from the real world: say, about Napoleon (р. 222—223). Thus Dolezel spontaneously approaches our ontology of possible worlds: he just has to admit that the information about the real Napoleon is available to any reader only from historiography that all by itself creates possible worlds none worse than an historical novel. Not for nothing it's so important when analyzing historical novel to find out whom of the historians an author had read. Very seldom this list appears to be too long: an instance of "The Captain's Daughter", backed by Pushkin's own historical researches that developed eventually into the "The History of Pugachev Rebellion", is an exception (though even more striking exceptions may be found: fir instance, Bryusov's "The Fiery Angel"); whereas an instance of "The Master and Margarita", where M. Bulgakov's historical erudition is backed by almost a single, though very extensive book by the Professor N. K. Makkaveysky, is quite typical. Therefore there's no reason to perceive any principal distinctions between the ontological statuses of the possible worlds W1 and W2.

What is non-trivial in Dolezel's theory: that is, what's its difference from the intertextuality concept accepted by the 70's structuralism (whose insufficiency was noticed, however, by Julia Kristeva himself whom Dolezel quotes)? The difference is that the intertextuality concept here becomes referential and ceases to be purely semantic (in the language that we, following Dolezel, are going to adopt below, it would be proper to say so: intertextuality is a phenomenon both intenstional and extensional at once, and not merely intensional). Intertextuality becomes apparent not only on the level of literary language (in the choice of words, figures of speech, metaphors etc) – on which level it has been usually studied – but also on the level of the possible worlds that were built in each one of the works bound by transduction (see especially р. 201—202).

Now we can define a notion of hagiographic substrate much more accurately.

When we introduced it, we spoke of the substrate as a peculiar "language of hagiographic narrative". We didn't pose a problem of specifying it from the point of view of poetics. Now we can pose this problem and also introduce another useful term borrowed from Dolezel: texture (in English tongue, one of the firm meanings of this word as a term used by literary critics is "peculiarity of author's language"; however, Dolezel endows it with special meaning).

Texture is precise verbal expression of narrative, that is, exactly the way it's formulated literally. It should be immediately made clear that we, contrary to Dolezel, are going to perceive the term of "texture" more broadly, extending it even to the texts of folk legends that are being fixated with an absolute accuracy (for Dolezel who analyses only works by known authors of literary fiction of the New Age this reservation isn't relevant). Therefore, Therefore by analogy with natural science, the word "precise" from our definition of texture should by understood as relative: "with precision of" (we'll get an opportunity to formulate our definition of texture more strictly later on, after having introduced a certain special logical apparatus).

Intertextuality, according to the above said, becomes apparent not only on a structure level but also on a level of reference which is a level of respective possible worlds. That is so for hagiographic substrate as well.

Hagiographic substrate exists on a level of reference (that is, of the components of respective possible world), and on a level of texture (of literary style and language).

If it existed only on a structure level, then ordinary methods of literary analysis would suffice for its analysis, and we would content ourselves with conversation about "literary dependence", not of hagiographic substrate as such. However, dependence on a level of reference that cannot be tracked down by traditional methods of literary criticism is much more important for understanding of any literary work (and particularly hagiographic). That's why hagiography didn't wait for Dolezel and developed, quite apart from literary criticism, its own methods of analysis. However, nowadays literary criticism keeps up with a level of critical hagiography they know what to say to each other and can cooperate on equal terms. It's necessary to admit that in regard to the structure of possible worlds of hagiography Dolezel, quite unconsciously, systematized many separate observations of bollandists and created a stimulus to notice something else that they never described.

Intentionality of creating and understanding possible worlds

Before taking a look at the internal organization of possible worlds we have yet to consider one more notion, a notion of intentionality (which in no way should be mixed with other notion, a notion of intensionality, to be considered below!). On the whole we will follow Dolezel's way of understanding intentionality (p. 57—59).

Intentional (from the Latin intentio) is any action of rational (and only rational) creature pursuing a certain goal.

Considering complicate strategy of human behavior it's possible to observe that his or her goals may change yet it doesn't affect the mere fact of intentionality of his actions.

Non-intentional are actions undertaken when a human acts not as a rational creature but, for instance, obeying the laws of nature. In certain acts of this kind an aspect of intentionality may be present as well: for instance, satisfying hunger isn't intentional in general but may be quite intentional in a specific situation. In terms of St. Maximus the Confessor, customary for patrologists, a notion of intentionality could matched with a notion of ?????????? (but not a notion of ?????: a mere wish, without action, cannot be intentional by definition).

Non-intentional are also casual acts of a rational creature.

Now we are getting to the actual reason of introducing the notion of intentionality: intentionality should not be mixed up with consciousness.  

Thus, (here Dolezel cites an example from A. Donagan, Choice: The Essential Element in Human Action (London, 1987) 110), car driving is an intentional action. However, it's performed in a different way by beginners and by experienced drivers: what makes beginners take conscious decisions is performed automatically by experienced drivers. Nevertheless it would be absurd to conclude that an experienced driver acts less intentionally.

It's no less obvious that the activity of both creator and reader of narrative is intentional; however, it's difficult to differentiate between conscious and unconscious within its limits. A notion of intentionality allows us to avoid it and thus to escape sinking into the depths of the psychology of creativity and the psychology of perception. In order to understand the logical structure of possible worlds it would be enough to know that it was built intentionally, whereas the question of how consciously its makers tried to use fiction, and how much they failed to notice that they followed their own fantasy, won't be essential to us.

That would remove initially absurd that is, non-relevant questions of how "truthful" and "pious" people could "lie". An answer to such questions is clear beforehand: if they indeed were subjectively pious, they could not lie, yet it doesn't mean that their intentional activity was entirely controlled by their consciousness (where such control is possible, any work granted with artistic qualities can scarcely be created: whereas artistic qualities are considered very desirable for hagiographic documents as well).  

Speaking of narratives' readers we should mention also the possibility of non-intentional (and at the same time, non-casual) component of reading. It refers to the sphere of manipulating consciousness, and in regard to hagiography it refers to the hagiography's functioning as a principal ecclesiastical mass-media. Within the margins of critical hagiography we'll have to limit ourselves with just raising this question. Detailed answer to it is possible only by the means of interdisciplinary research that would combine methods of critical hagiography with methods of history and sociology.

Interior of possible worlds

We are getting, at last, to the internal organization of narrativist possible worlds. We will principally follow the theory of Dolezel who developed it through the use of very restricted material of European literary fiction of the New Age, and doesn't reckon that it can be applied to narratology in general: especially if a notion of naratology is to embrace the narrative of natural science. However, here we'll take liberty to disagree with Dolezel.   

All the contemporary narratology emerged from the modern philosophy of language that, in its turn, emerged from one principal source: logics of Frege who resolved to turn to internal organization of a natural language so as to understand better the language of mathematics and the language of science. With its further development this, initial for analytical philosophy, intuition of the unity of human language – be it natural language, language of science or poetical language – so distinct when it concerns the founding fathers of analytical philosophy, Frege and Wittgenstein, has been gradually lost. Eventually this intuition became so forgotten that since the second half of 20th century it became a habit to repudiate it. Usually this repudiation hid the interdisciplinary disunity: literary critics had little notion what's going on in philosophy of science, whereas philosophers took no interest in literary criticism and linguistics. Therefore even most striking parallels remained unnoticed (we have already mentioned it by the example of Dolezel and Ankersmit who obviously could enrich narratology of each other, but it would be even better if they took account of Quine and Putnam's philosophical research to its full extent...).

Quine's aphorism: "Philosophy of science is philosophy enough" reflects our position as well as Frege's initial intuition. However, and not in the least due to the same Quine, we can recall in due time that science is in some way present in historiography, and even in literature. After all, and we will yet return to it, it's namely Quine who said that, in spite of all his subjective faith in modern physics he has to admit, from epistemological point of view, that physical objects and Homer's gods differ "only in degree and not in kind"…

Therefore we'll try to keep up with Dolezel, without forgetting, however, about the sources of analytical philosophy and those of its ideas that we deem principal, that is, those that give the necessary ground for judging the applicability of our conclusions to the narratives of various kinds. Dolezel's book, as we have already cited, is a call to turn to the tradition of analytical philosophy; that's why even when we disagree with its author it can be appreciated as an attempt to go farther following his call.

In its most important conclusions our analysis will arrive at the concurrence with empiric laws inferred by Delehaye. It seems to us that mathematician's soul, as Father Delehaye has been, regardless of how imperfect our further logical findings may appear, would heed with benevolence at least the mere fact of applying to his science the philosophy that was born, just as a critical hagiography itself, out of the spirit of mathematics.

Two principal terms: extensionality and intensionality

Literature: Gottlob Frege, “Logics”:Г. Фреге, Логика и логическая семантика. Сборник трудов / Пер. с нем. Б. В. Бирюкова. Введение и послесловие Б. В. Бирюкова (М., 2000); Rudolf Carnap, Meaning and Necessity: a Study in Semantics and Modal Logic, Chicago: University of Chicago Press, 1947.

The thought concerning the necessity of finding some way do differentiate between the meaning of what is said and the object of what is said, though not quite obvious, occurred nonetheless on a regular basis in many linguistic doctrines. For instance, respective terms of "connotation" for "meaning" and "denotation" for designated object were proposed yet by John Stewart Mill in his "System of Logic" ('A System of Logic, Ratiocinative and Inductive, Being a Connected View of the Principles of Evidence and the Methods of Scientific Investigation, 2 volumes', London, Parker, 1843; gratis to B. Biryukov's Afterword to the Russian edition of Frege's 'Logic and Logic Semantics') that, in his turn, appealed to the tradition of late scholastics (nominalism). It's there that a word "intension" first appeared as a synonym of "connotation". This word originates from the Latin intensio (not to be mixed up with intentio!).

However, the most important is to say a right word in a right time and in a right place: that is, in the context of development of what has been called analytical philosophy. And that's what did Gottlob Frege having published in 1892 an article ‘On Sense and Reference’ (Ueber Sinn und Bedeutung). Notions of "Sinn" (usually translated to English as "sense") and "Bedeutung" (usually translated as "reference" and not as "meaning", because "meaning" retains its simultaneous relation to both "sense" and "reference") were introduced in more or less the same sense as J. St. Mille's "denotation" and "connotation", but they appeared to be inscribed into more complicate conceptual construction that after Frege was being depicted schematically as "Frege's triangle":

Sign (Name)

signifies                          represents

Object (Meaning)                  refers to                   Sense (Sense of Sign)

This construction displays Frege's principal idea: a sign though it's sense points at its object.

"Sense" is defined by Frege as "a mode of representing the denotation (meaning) in the sign". For instance, in the sign "house", the sense will consist in the notion of "house" or otherwise, in the whole class of objects we name as "houses" (in this case we digress from possible ambiguity: thus, "house" in a meaning of "dwelling", and "house" in a meaning of astrologic term are just two different signs), whereas a meaning of this sign is a certain house that we may wish to point at with its help. It's obvious that wishing to point at the same specific house we may picture in our mind quite different houses so that the presumed class of objects named "houses" may appear in rather different forms: the same signs usually include many modes of representing the same denotation.

A certain order may be introduced into this diversity if we start considering the dependence of sense from denotation, that is, of qualities (predicates) of designated individual itself. That's what Frege's disciple, Rudolph Carnap, actually did. Carnap (along with Wittgenstein), created a link between Frege's tradition and Anglo-Saxon analytical philosophy, especially due to his emigration in 1936 from Germany to USA (carried out with Quine's help), where he started publishing in English. 

Instead of the terms of "sense" and "meaning" Carnap gave preference to "intension" and "extension" respectively. It's referring to these terms that our tradition of literary criticism (Dolezel and all the like-minded) speaks of "intensional" and "extensional" content of text.

However, the principal difference of Carnap's concept from Frege's initial concept, of course, didn't consist in terms but in his definition of intensional as a function of extensional (we intentionally slightly modernize original terminology of Carnap who used the word "function" in a sense more narrow than the habit of 70's made it). If we wished to draw "Frege's triangle according to Carnap", we would have to reverse the arrow's direction in the base of Frege's triangle and to replace the word "assigns" with the words "is function of". Carnap didn't deny what Frege said but drew attention to what he never said.

Carnap approached the problem of reference otherwise than Frege: in regard to qualities of denotation. Coming back to our example of the sign of "house", it could be explained so: the "sense" – but henceforth we'll say: the "intensional" – of the name "house" are those qualities of house that allow us to determine that a given object belongs the class of houses; the extensional of a name, as in Frege's case, is this very house that we speak of.  

The cited example can give a general notion of Carnap's respective ideas (for their full survey see: R. Carnap, Introduction to Symbolic Logic and Its Applications / Transl. by W. H. Meyer and J. Wilkinson (N. Y., 1958) 40—42), and such notion is enough for us: regretfully, namely out of reason that Carnap, as well as Frege, failed to describe precisely enough what intensional is, so that logicians were left with a fair amount of work to do after him. However, it is Carnap who gives us "universally accepted" notion of intensional meaning that we have to begin with before proceeding to any specifications.
It's possible to perceive quite well the difference between intensional and extensional content at least by the examples given within the limits of this definition.

Thus, extencional "Dostoevsky" (a great author, also a certain historic person) corresponds with rather various expressions: "Dostoevsky", "F. M. Dostoevsky" (and not his younger brother M. M. Dostoevsky, "Fedya" (in the diaries of his wife, Anna Grigorievna), an author of "Crime and Punishment", "great humanist", "reactionary" and so on. Being applied to the same person (extensional) these expressions vary by their content. The content of their difference will just be intensional. Thus, just "Dostoevsky is a name of a great author, "Dostoevsky" with initials is a name of the man differing from those sharing his family name, "Fedya" is the same man's name in the family circle, "author of a certain novel" is clear enough, "humanist" and "reactionary" are just certain qualities of the same person in certain persons' eyes. If, meeting these expressions, we can perceive by some signs the intended person then we can easily interpret them. Retelling pointing at a certain person of Dostoevsky can become an interpretation. In the retelling all the differences between those expressions will be lost yet we will understand precisely who they point at.

Thus we arrived at practical explanation of a notion "intensional", that is, of what it means "to be a function of an extensional" or, just the same, "extensional function". Formally we speak of a function (or, more precisely, of functions, since we speak of intensionals of several different statements referring to the same Dostoevsky) that imparts to Dostoevsky his various qualities expressed in the statements referring to Dostoevsky. This is not very stringent definition yet its lack of stringency is almost entirely compensated by the practical law that we have just found out:
Intensional content of a text is a content that gets lost (changed) in retelling. Extensional content is a content that is stable in regard to any retelling of a text.

It's a traditional opinion that literature differs from science exactly out of reason that science strives for unambiguity of reference and consequently to elimination of intensional content, whereas literature, on the contrary, is built exactly on it. This idea was arranged by Frege who suggested distinguishing the "poetical" language that lacks reference from both ordinary and scientific "cognitive" language. Even nowadays Dolezel doesn't entirely distances from this tendency (p. 136), although his book shows that literature is also built on extensional content no less than on intensional, whereas he finds much to criticize about Frege's idea of "poetical" language and other related ideas associated with "auto-reference" (р. 3—6, 228—229; also see above).

However, an idea of total extensionality is utopian even for science. Carnap dreamed of creating for the needs of science special "extensional language" where reference to all denotations would be unambiguous, that is, bereft of intensional content. Such was the essence of Carnap's "Thesis of Extensionality": "[A] universal language of science may be extensional". Carnap attached great importance to this thesis, while clearly reserving that it's unproved and acceptable "only as supposition"; nonetheless, he considered his thesis "fairly plausible" since it corresponded to all the known (to him) cases of scientific description (R. Carnap, The Logical Syntax of Language (London, 1967) (International Library of Psychology, Philosophy and Scientific Method) 245—247; first published in 1937). However, already in the last years of his life Carnap's belief in potential unambiguity (extensionality) of the language of science appeared to be anachronistic. By then, only mere formulas of quantum mechanics (not interpretations but namely mathematic formulas) amounted to three, and it was clear that their physical meaning is by no means equal, no matter how to interpret them. Language of mathematic formulas appeared to be as intensional as any other. However, another instance might be taken from 19th century physics where Bolzmann inferred statistical formulas for physics' basic laws, having created thus new discipline: statistical physics.

However, no facts, of course, could shake Carnap's philosophical theory that was, at its basis, not Carnap's theory at all but the most important intuition of science of the New Age. As we have more then once said, philosophical theory may be refuted only by the means of another philosophical theory, and precisely such theory emerged in 50's. We'll turn to it as soon as it will become relevant to our further elaborations.

And so far let's notice that now we can draw with an easy heart our own, narratological Frege's triangle: texture (see above) – extensional content – intensional content. However, at the present we'll content ourselves by drawing such triangle only in our mind, since so far we have said almost nothing concerning how to draw arrows in such a triangle./p>

We'll summarize this chapter by the means of very simple formula: Intensionality is extensional function to texture, that is, such function I for which extensional content of the text E is its domain of definition, and texture of the text T, that is, its literal verbal expression is its set of values:
Т = I (E)

Ontological relativity and intensionality of being

Literature: Willard Van Orman Quine: Two Dogmas of Empiricism // http://www.ditext.com/quine/quine.html (this web edition is preferable to the printed ones because it's synoptic edition of two different versions of the article, from 1951 and 1961); Ontological Relativity (Columbia, 1977); Hilary Putnam: Reason, Truth, and History. Cambridge: Cambridge University Press, 1981, see especially the articles "Problem of Reference" and "Two Philosophical Perspectives.

Soon we'll have to turn more in detail to the notion of texture: the very verbal material of which text is created.
Generally speaking, it consists not just of separate statements but of finished phrases. Frege's most important innovation consisted in regarding namely phrases and not separate words as a principal channel to convey the meaning. That is, for instance, not a separate word "house" but a phrase "this is a house". In spite of all its importance, this innovation hasn't been radical enough therefore we'll have to consider texture as a whole; but so far we'll stop at what phrase offers us.

Phrases point at certain facts and not at properties of object as such. What are these facts? That's what most interesting.

During Frege's life around this question another interesting triangle has been formed: of three philosophers, Frege, Meinong and Bertrand Russell. Meinong published in 1904 a paper named "On the theory of objects" (Ueber Gegenstandstheorie), and Russell immediately reacted with review and, above all, with his famous paper "On Denoting” (1905). Since that, extreme positions in the further discussions on denoting were determined: any possible point of view had to fall somewhere between Meinong and Russell.

Thus, in regard to phrases Frege reckoned that their sense is what's told in them, and their denotation, since phrase cannot designate any object, is just a truth value, that is, whether what's told in phrase is true or false. In regard to non-existing objects of "poetical" language Frege suggested to reckon that they lack meaning (reference) and therefore, in particular, phrases of this language can't possess any truth value: "that's not why we love them", as they would say today, precisely expressing Frege's attitude to to artistic fiction. 

Meinong suggested considering all possible objects, regardless to whether they exist or not, as possible objects of reference: that is, he was the first who introduced the notion of reference to object regardless to whether object exists or not and even regardless to whether it's logically possible (according to Meinong, not only absolutely non-existing thing like Pegasus the winged horse but also something like round square can become object of reference). Accordingly, for every phrase such possible world may be imagined where its truth value will be "true" and not "false".

"In contrary to <…> the preference of real that hitherto in fact is not principally overcome in any science, today, however, there exists undoubted need for the science that worked on its objects without being restricted by namely separate case of their being, so that it could in this sense be designated as being-free [daseinfrei]. The science about object as such, or pure object, I have called the theory of objects" (from Meinong's auto-summary that became his scientific testament (Selbstdarstellung - Die Deutsche Philosophie der Gegenwart in Selbstdarstellungen - vol. I - Leipzig 1921).

Meinong kept basic structures of Frege's idea of language, but their most important aspect, initial reference to reality, was withdrawn from them. Frege started dealing with linguistics namely in order to understand this reality, whereas Meinong demonstrated that Frege remained within specific reality of language that is not too much connected to the world external to our mind. That was a challenge, and it was answered not so much by Frege as by far more radical Russell.

Frege's thesis about lack of reference for imagined objects was brought by Russel to its logical limit: attaching the truth value of "false" to all the phrases describing imagined situations.

Russel's famous example in his article "On Denoting" is a phrase "the present King of France is bald". In 1905 there already has been republic in France, on which base Russel infers that this phrase is false. Meinong in his stead should have inferred that this phrase may be false or true depending on what possible world it refers to, and Frege would have said that this phrase can have no truth value at all.

According to Russell, not only phrases featuring “present Kings of France”, “Pegasus” and round squares turned out to be false, but also scientific theories featuring essences that were afterwards rejected be science (for example, flogiston, special matter of burning according chemical theories accepted until approximately 1830's). At a first glance it looked convincing, but… Subsequently revealed "buts" appeared to be very serious. They were well summarized by Charles Crittenden: following Russell we should have admitted that "The Odyssey" is a "historic narration" making wrong statements in regard to history (Cyclops, Sirens etc) (Ch. Crottenden,  Unreality: The Metaphysics of Fictional Objects (Ithaca, 1991) 25; cited via to Dolezel, р. 227].

Semantics of possible worlds became an opposition to Russell within the margins of Anglo-Saxon analytical philosophy, and it seems likely that its most adequate ontology was suggested by Quine (though he he didn't put special interest in multiworld semantics but just used it as a formal logical move, referring to Leibniz) that from his views Putnam currently draws subsequent logical conclusions. Putnam has been closely tied with Carnap and even strove to follow him when it was possible, yet it emphasized only too distinctly his deviation from Carnap's philosophical tradition. Quine's article "Two Dogmas of Empiricism" (1951) may be considered a philosophical manifest of his new trend; this paper, in our opinion, contains philosophical ground for theory of literature, too, though hitherto nobody considered it this way.

It's a common opinion that Quine never managed to refute formally Carnap's program of creating extensional language and quite "objective" description of reality.  Anyhow, namely his philosophy became a death sentence to the scientific objectivism. This sentence had to be taken seriously when Putnam executed it: Putnam's theorem, that we've already mentioned, sounded as a burst of fire at the theory of truth as conformity, that is, as a certain "objective" description of reality analogous to mimesis in poetics. Philosophers hitherto wander around the corpse of this theory, hoping to find a mistake about verifying its death; however, as is well known, in such cases chances to find a mistake sweepingly lessen with time: no one succeeded in doing it for a quarter a century that has passed since Putnam's theory, that became a logical proceeding and formalization of Quine's principal philosophic intuitions, was published. As for philosophers' hopes that the decomposing corpse of scientific objectivism is "rather alive than dead", thus it should be until the new generation replaces the old one (as Thomas Kuhn wrote in "The Structure of Scientific Revolutions"; in this case we deal namely with a quite conscious (for its authors) revolution in philosophy as a whole and not only in its Anglo-Saxon school).  

Thus, let's try to summarize shortly what actually happened.

Before Frege there dominated a naive view on language (in terms of the philosophy of language found in, say, J. St. Mill), presuming that language operates with notions that all of them correspond to something in the empirical reality: as if "houses", "trees", "atoms" and other things really exist regardless of our consciousness.
Frege introduced the differentiating of sense and denotation having thus shown that, although the classification of objects that due to it we can determine classes of "houses", "fir-trees", "atoms" - that is, what he defined as "sense" - isn't arbitrary but is formed in our mind while enjoying no objective existence apart from our mind. How arbitrary the classification of objects in language may be, we started to realize indeed only due to the achievements of linguistics of the last decades: thus, the list of words chosen by Lakoff as a name for his book: "women, fire and dangerous things", refers to objects belonging to the same grammatical class in one of the languages of aboriginal Australians.

However, Quine resolved to take even more radical step that found its final logical substantiation in Putnam's theorem and, accordingly, emerged into more consistent and modern version as Putnam's philosophy of internalism. And namely, Quine resolved to acclaim that the objects that we can see in reality are also for some part the function of our consciousness: that is, not only their classification ("sense", according to Frege) but also denotations.

This conclusion should be understood not in the sense that these objects don't exist at all, and it's only our sensation that exists (then such point of view could boast nothing new, it would mean just repeating Berkeley), but in the sense that a human principally has no way to know whether the objects exposed to his mind correspond to what is located outside his mind in the outer world, and that constitutes both natural and logically necessary manifestation of human limited nature. 

This principle in general is obvious enough. Let's consider a certain external object and its representation within our mind. In order to appreciate the mode of its representation we are to get an independent access to it: but it's just like lifting ourselves by our hair. We try to evade this limitation by comparing many different representations of supposedly (and it's impossible to define it precisely) the same reality. On this account our knowledge of the world may be different and, in particular, more or less precise. That's why impossible to say, for instance, that all scientific theories or all points of view on a certain mundane question are equally good. On the contrary, practice shows (and here empiricist approach interferes) that things are not equal at all: but that is all. There's nothing else that practice shows: there can be neither absolutely precise (unambiguous) representation (reference), nor even precise method of appreciating the degree of an "error".

This may be expressed in a more simple way, in the language of pure logic. We discuss the problem of determining the conformity between two areas. It's possible to do it only enjoying an independent access to both of these areas. If a human isn't able to take the position of God, then in regard to the question of human perception of external reality it's impossible to do it.
Let's emphasize that in analytical philosophy such conclusions are drawn without any assumptions regarding the ontological status of external reality as such.

More precisely, formally Putnam abandons this question by using the method of "putting out of brackets" suggested in 1913 by Husserl; this method allows speaking of mental representations of reality regardless of this very reality (see E. Husserl, Ideen zu einer reinen Phanomenologie und phanomenologischen Philosophie. Erstes Buch: Allgemeine Einfuhrung in die reine Phanomenologie. Edited by Walter Biemel. The Hague, Netherlands: Martinus Nijhoff Publishers, 1950). Ideen zu einer reinen Phanomenologie und phanomenologischen Philosophie. Erstes Buch: Allgemeine Einfuhrung in die reine Phanomenologie. Edited by Walter Biemel. The Hague, Netherlands: Martinus Nijhoff Publishers, 1950. In his time, Husserl got the stimulus to develop his phenomenology namely from Frege who reviewed one of his early pre-phenomenological works; however afterwards the ways of phenomenology and analytical philosophy, for all their inner proximity, didn't cross for a long time, only last decades saw the reverse tendency gathering strength).

In fact Putnam really believes in the existence of "objective reality" and shows no sympathy to Niels Bohr as a philosopher (contrary to Quine who reveals such sympathy). However, the difference between Putnam's internalism and Copenhagen interpretation exceeds the limits of internalism as such that limits itself with statements regarding the cognitivity of external world, without special concern towards its ontological status.  

In regard to the question of ontological status of external reality internalism doesn't exceed the limits of what Quine formulated in 1968 in the article bearing the same name as a principle of ontological relativism:
"What makes ontological questions meaningless if they are considered absolutely (and not relatively) is not their universality but their quality of being a logical circle. The question in the form of "What is F?" may be answered only by turning to the following term: "F is G". An answer has only relative sense: a sense relative to non-critical acceptance of G".

Here is told, though in other words, just what we, retelling Putnam, have written a bit above: about the impossibility of independent access to the reference between our consciousness and external object.

Putnam's principle of internalism is formulated as a development of Quine's ontological relativism:

"Internalism doesn't deny that in regard to knowledge experimental initial data play their role, knowledge is not a tale that knows no limitations but internal conformity; however it does deny the existence of such initial data that wouldn't be formed to a certain extent by our notions, by the dictionary that we use in order to fix and describe them <…>. The initial data upon which our knowledge rests are conceptually infected, yet it's better to have infected initial data rather than to have no data at all"(H. Putnam, Reason, Truth and History, p. 77, italics are by the author).
Thus, in our perception of external reality an element of "randomness" or, more precisely, an element of being conditioned by our internal mental states (this circumstance is actually emphasized by the term "internalism"). It regards not only intensional (as Frege thought) but also to extensional content of our notions.

As for purely logical aspect of the matter, that is, scrutinizing the problem regardless to our mental states, Quine's philosophy lets know, and Putnam's theory strictly proves that, regardless to truth value of phrases, both intensional and extensional content of the words constituting the phrase may vary without limit. Thus, the same state of things in the external worlds (or, more precisely, in each one of possible worlds) may be correctly described (in the sense that it might be described by phrases with truth values of "true") by a random number of mutually incompatible interpretations (that is, these statements will be incompatible in regard to their intensional content).

Thus, semantics of possible worlds becomes an instrument allowing proving the absence of strict external predeterminacy of intensional content of phrases referring to objects. As for objects themselves, the above said means that the external predeterminacy is equally not characteristic for both intensional and extensional content of the corresponding terms. Two dependences at once take place, and both of them aren't strict: for some part, our classification of objects depends on the objects themselves, and for some part, variety of objects we're able to see depends on classification prepared in our mind beforehand. In the next pages we'll turn to examining these dependences more in detail.

One more notice should be made for readers used to philosophic categories of patristics or scholastics. None of the representatives of Anglo-Saxon analytical philosophy revealed professional knowledge of patristics, but some, and Quine in particular, showed noticeable interest in scholastics (not to mention that certain knowledge of scholastics is a part of the standard of European philosophical education). Quine reckoned that the question of the classes of objects corresponds to the medieval question of universalia and therefore ranked himself among the nominalists. Indeed, in his philosophy, where mental effort was needed to explain the reality of physical objects, any stimulus to substantiate also the physical reality of the classes of objects would fade away. Just as our ability to make up individual objects along with inability for "uninfected" perceiving the existing objects can't prove unreality of individual objects on their own, our ability to make up classes of objects along with inability for "uninfected" perceiving of these classes can't prove their non-existence.

In the old European tradition the argument between nominalists and realists was perceived in the way presuming that both schools agree in regard to reality of individual objects so that this question is obvious and out of discussion. Ontological relativism (and still more, Bohr's principle of complementarity) raised the question of individual objects almost as sharply as a question of universalia had been raised in older philosophy. Our own approach suggests that all the affirmations and limitations that the principle of ontological relativism applies to individual objects, are applicable to universalia as well.

Having described, at last, what's the essence of the principle ontological relativism, let's try and see what it turns out to be for narratology.

What matters for narratology is a relativity of the question "What is F?" The answer is a phrase of the sort: "F is this or another thing" where an explanation of "what F is" belongs to the field of intensional content (extensional content of the phrase is its truth value). In particular, any characteristic of the reality of being of F fall into the field of intensional content. That is the fact fundamental for narratology: being of any narrative substance (that is, any object mentioned in narrative) is an intensional notion). That's why a being of narrative substance may have different gradations that we have conditionally called "measures of reality".

We will proceed to specifying a notion of gradations in the coming pages, but meanwhile we have to realize as clearly as possible what intensionality of an object's reality means. As we have just made sure, it automatically and formally follows from the principle of ontological relativism: yet any formal conclusion is good to accompany with a simple explanation.

Dolezel (p. 147) formulates an analogous conclusion regarding intensionality (and therefore possibility of being gradated) of being, but for the objects of artistic fiction, and considers this quality of being a fundamental feature of both his theory of literature and fictive objects as such. If the first is undoubted, the second is wrong. Intensional (and therefore possible to gradate) is a being of any narrative object regardless to whether this object is a result of conscious (artistic, for instance), fiction. It's clear enough, recalling what we told above about the similar nature of fictional and historic narrative. In the case of the narrative of natural science the situation is also up to no essential distinctions. Such distinctions appear only if one, like Dolezel, shares illusion that natural science enjoys a direct access to the objects of external world whereas ontological relativity doesn't exist at all. However, ontological relativity exists hence any quality of possessing a real being is just one of the elements of intensional content of a term corresponding to the object.

Intensionality of being of a certain object reveals only what place it takes within a given narrative: how seriously does given narrative insist on its existence. Gradation of being of an object is determined by a mode of narration about an object: hence, depending on the nature of narration we perceive more or less seriously the statement made by narrator in regard to the being of an object. Let's recall that above we formulated the practical law stating that the content is intensional if it refers directly to the form of narration, that is, the content that may disappear during retelling. Any declarations in regard to ontological status of objects in three types of narratives (not only fictional, but also those of history and natural science) refer to the form of narration, that is, to its intensional content. However, we know it all quite well even by daily intercourse, where we often try to judge by the form of our interlocutor's narration the range of reality of what he tells us about.

Running a deal forward, let's notice that the principal idea of all Delehaye's critical hagiography, which is an idea of dependence of hagiographic narrative's historic content on its literary form, is precisely a special case of this general rule: intensional nature of being of narrative substances.

Texture and intensional functions

The time has come to turn from phrase to texture. Above (quoting Pavel) we mentioned that literary criticism found itself in rather a straitened situation because of its linguistic orientation. Linguistics (in its most developed spheres that actually set an example for the structuralist literary criticism) usually doesn’t rise above the level of phrase, which is by no means enough when the goals of literary analysis are concerned. On the whole it refers to the general problem of narratology (including natural science) that was generally posed by Quine. It directly concerns the impossibility of unambiguity of reference discussed above.

The essence of this problem is that any concept assumes its true sense (and therefore should be appreciated) only in the context of comprehensive state-of-mind or, in terms of multiworld semantics, in regard to its place in the arrangement of respective possible world.

Let's consider an example that was given by Quine in his "Word and Object" and subsequently modified by Putnam (in his "Problem about Reference") in connection with the proof of his theorem.

Let the phrase "This is a rabbit" be given and true. What does it tell about? Because of the ambiguity of reference the term "rabbit" may signify, as Quine points out, an animal, a piece of its meat and a reference to the respective species all at once. This example, however, may yet suggest that the reference of the term "rabbit" must necessarily presume some relation to "rabbitness". This illusion is carefully dispersed by the example illustrating the proof of Putnam's theory: there the term "rabbit" may signify the object "cherry".

Instead of reproducing the illustration of Putnam's theory we'll give another example proving that it's not so distanced from our daily life. It's easy to imagine the situation when we can with no difficulty detect the reference of the word "pych" used in regard to a girl, especially if it's pronounced with a characteristic "Eastern" accent. In the traditional one-world semantics we would speak of using a poetical trope, whereas in the semantics of possible worlds we must speak the way Putnam's theory was proved, of such language of a certain possible world where the word "peach" (at least, in the form of "pych" – status emphaticus of suppletive formation; let linguists correct us in the case we mixed up anything) may have a meaning "girl". There's a doubt that this meaning is fixed in the actual dictionaries of Russian language: dictionaries reflect the ambiguity of reference only in so far as it becomes a part of language standards.

In all the abovementioned cases we manage to overcome the ambiguity of reference because we deal with whole texts and not with separate sentences. That's why, as Quine pointed out, Frege's idea of regarding phrases of language as principal bearers of sense appeared insufficient. And that's why Dolezel insists on considering the texture of text as a whole and not as a chain of separate phrases.

It's within the texture that the mechanisms aiding to overcome an ambiguity of reference snap into action. How does that happen?

Traditional semantics (including "mimetic" poetics) always proceeds from an object to its name, and therefore appears principally unable to see necessary mechanisms.

Semantics that takes into account principle of ontological relativity (including Dolezel's poetics, though he restricts this principle to fictive objects), goes in both directions at once: not only from an object to its name but also from name to object. It takes account of our consciousness' ability to modify and even, alas, construct objects of the external world.

It's this ability of our consciousness that provides us with referential unambiguity even where it's absent, that is, generally speaking, everywhere (because initially there's no unambiguity anywhere). Referential ambiguity isn’t "revealed" (there's nothing to reveal here because nature lacks it) but appointed. From the point of view of pure logic such procedure may look arbitrary but, having been examined closely, it also can't be denied a sort of peculiar logic.

Despite the fact that general prolegomena to the respective logical theory were suggested half a century ago by Quine in "Two Dogmas of Empiricism" relevant theoretical researches are still very few and, all of them are oriented more or less ad hoc. Therefore Dolezel's researches appear to be still more important, though they are limited with a sphere of poetics. Practical aspect of what we are going to discuss here is likely to be developed best, to the extent of inferring a large amount of empiric laws, in the modern theories of mass media and information war; whereas its formal and logical apparatus is to be found in Putnam's theory of reference. Regretfully, Dolezel's terminology that was developed specially for our goals and therefore will be used as a basis for our own work, doesn't agree with Putnam's terminology that, though not intended specially for describing processes of imparting unambiguity to reference, boasts more detailed notions that our case demands. Therefore we'll have to readjust both terminologies ourselves.

Thus, recalling that we defined above the intensionality as an extensional function, that is, a function of extensional. This is a function for which extensional content is its domain of definition. However, now we have to find a function for which extensional content will be its set of values. By its sense it's a function that imparts unambiguity to initially ambiguous references of texture: that is, it's a function affecting the texture and namely texture is its domain of definition. It's quite obvious and has been partially illustrated in the Chapter 5.9.2 (by the example of quality of reality of being) that such function is intensionality.

Thus, we arrive at the formula:
E = I (T)

That describes extensional content as a function of texture.
It's not so easy to choose a name for such function. Dolezel calls it intensional function, but following the principle according to which we called intensionality itself extensional function, such function should be named textural. However, Putnam calls it both extensional function and intensional in a special, differing from traditional, meaning of the word. Willy-nilly we'll have to work this all out, yet we don't dare to drag all readers with us into this maze. Meanwhile, we proceed straight to conclusions.
By the way of conclusions we'll limit ourselves with building in general such Frege's triangle that would conform to the general theory of narratology. A philosophical summary of all that has been said above in regard to the theory of reference will be formulated below.

Thus, principal logical foundations of our narratological Frege's triangle will be statements formulated as formulas:

 T = I (E) - formula of intensional as extensional function to texture, and
E = I (T) - formula of intensional as "textural" function to extensional ("intensional function", according to Dolezel).

Comparing two formulas brings to light an evident, generally speaking, idea, from which, however, rather non-evident conclusions may be drawn: reference from sign to object and from object to sign (from texture to exntensional content of the respective possible world) is always mediated by sense (intensional content). What is non-evident here will be discussed in detail below. In the meantime, here is Frege's triangle for generalized narratology:

Drawn by A. Romanovsky.

Designations T, I, E correspond to formulas.

The edge ET is marked by dotted line because it corresponds to straight reference and extensional language, that is, to what is principally impossible. However, it may be considered as an unattainable limit of "extensionalization" of language that may be represented graphically as reduction to zero (in the unattainable limit) of the base EI, that is, until intensional and extensional become congruent.

Direction of arrows concurring with the edges of triangle corresponds to the order of narration about already existing reality: there are objects; their certain concept is formed in the mind and subsequently expressed in the text.  

Direction of arrows along the segment of the circle TIE is shown along the circle namely because it regards something that principally exceeds the limits of Frege's "triangle", that is, all the system of reference that Frege considered. This is schematic representation of what Frege principally couldn't see.

First, this is a certain function overcoming an ambiguity of reference in T at the expense of context or, more precisely, on the expense of clarifying the intensional content of context, which allows to arrive at the necessary degree of determination (not necessarily unambiguity) of extensional content. 

Second... this is even not a "certain" but the very same function E = I (T), that leads to creating the objects themselves.

This is not a reservation: this function really does create objects of respective possible worlds whose degree of reality is seldom initially known and never reaches 100 percent. It can't be helped: we haven't got, and aren't likely to get any other method of cognizing reality besides creating possible worlds (within the limits of our natural abilities, that is, without special divine revelation). That's what constitutes the principle of ontological relativity of whose special relation to the notion of texture we will speak below.

And in the meantime we'll proceed to specifying various technical details regarding intensional function E = I (T).

Intensional in semantics of possible worlds

Our task now is considering an intensional function in the semantics of possible worlds, for which purpose we'll begin our reasoning with the following example, referring so far to the one-world semantics, and not to intensional function but to intensionality.

Let's consider the word "I" (this example is borrowed from Putnam but modified in accordance to our needs and terminology).

Various people in certain situations may call themselves "I". Their common feature will be that all of them in this case have to figure as speakers. It's evident, thus, that "being I" is an intensional: because we can easily reproduce the direct speech by indirect, replacing first-person pronouns with third-person, while a referent (extensional) doesn't change. The same person may be called "I", "he", and also "you" and even "we", but an object of reference, that is a person itself, will remain the same.

It appears that the word "I" (element of texture, T) is corresponded with intensional content (intensional) that is extensional function, but not of a single referent but of many referents. Formula T = I (E) keeps its value but at the same time the value of I is a function selecting from all the objects of a given world only those possessing a quality of being animated, that is, can figure as speakers.

And who, generally speaking, possesses a quality of "being a speaker"? An answer will depend on the possible world at which this question is aimed.

Thus, in the possible world of Bertrand Russell's philosophy there can be no doubt: a quality of "being speakers" is possessed by humans and that's all. In the possible world of Frege's philosophy this quality would be possessed by humans and, maybe, the Lord as well. In the possible worlds of more traditional Christian views, humans and the Lord would be joined by angels and demons. In Homer's possible world animated objects would essentially multiply on the experience of the multitude of gods. In the possible world of CS Lewis' "The Chronicles of Narnia" humans would be joined by various dwarves and animals: however, not all animals but special ones (that are called in accordance: "talking animals"). And in authentic epos all animals would become "talking". At last, such possible world is possible where there is nobody but only "what", that is, there's nobody to speak at all (for instance, from such a world our present world evolved, if we give credence to Russell's philosophy and the rest of atheism).

In all these worlds the domain of definition of intensional (extensional function) of "I" will be rather different. In the last of our examples it will constitute an empty set.

We chose an example where extensional changes depending both on the context within the limits of each separate possible world and on the choice of possible world at a time.  The first kind of change isn't peculiar to all the terms, whereas the second is peculiar to all of them. Since the change of extensional occurs in both cases in a similar manner, it provided us with a convenient illustration of the meaning of change of extensional depending on the kind of a possible world.

Thus, in the multiworld semantics "…intensional specifies the way extensional depends on a possible world" (Putnam, Reason, Truth and History). This doesn't exhaust the notion of intensional, but this is its very important component that Putnam emphasized. Traditional idea of intensional prevented noticing it.

Speaking in general terms, Putnam's intensional IP (as we will call it; Putnam himself gives him no name but only reserves its difference from intensional in a regular sense) is a function of possible worlds indicating the meaning of a term in each one of the possible worlds:

T = IP (M),

where Мis a total of logical objects of each one of the possible worlds, that is, full extensional content of the given world (wishing to simplify formulas we are giving up indexes customary in formal logic, therefore our sign M lacks indexes corresponding to the difference of worlds, as well as indexes corresponding to the difference of objects within each separate world

Outside the limits of Putnam's intensional such intensional content is found that doesn't change from one possible world to another, that is, such as we discussed above. For instance, in the same possible world Dostoevsky can be both "a great author" and "an author of 'Crime and Punishment'", whereas "a great author", as well as "I", refer in this world not to Dostoevsky alone, though also not to any "speaker", but only Dostoevsky can be "an author of 'Crime and Punishment'". It's quite possible to imagine some other possible world, if not Dostoevsky's, then Turgenev's, and if not Turgenev's, then an author Karmazinov's from "Demons" – where even the term "a great author" could have only one and quite certain onject of reference, because if not Turgenev, then its caricature, Karmazinov, was aware of only one great author…

Let’s call this intensional content an intentional in the sense going back to Frege, or shortly, Frege's intensional - IF .

Then an intensional I in all our formulas is a conjunction of functions IP ^ IF : that is, both these functions should be taken into account simultaneously ("and", but not "or").  
Both functions, Frege's and Putnam's intensionals, are functions of extensional to texture:
IP  ^ IF : E > T

Now, in the interests of building a more general narratology, let's proceed to the logical apparatus allowing direct comparison of the problems of reference in the fictional and historic narratives on one hand and in the narrative of natural science on the other. 

Principle of uncertainty for reference

As our narratological Frege's triangle shows, our task now is examining of the problem of reference at the unusual visual angle: or, more correctly, from the unusual visual angle, the angle of Frege's triangle. Usually philosophy, without excluding phenomenology, built the theory of reference proceeding from objects, that is, from the apex E. Frege, analytical philosophy T. However, we'll have now to proceed from intensional, that is from the apex I.

Therefore it would be desirable to use the logical apparatus traditional neither for the philosophy as a whole nor for the philosophy of language in particular, but nonetheless not too recently invented. It is the formalism of so called fuzzy logic whose creation was declared by Lotfi A. Zadeh in his 1965 article (in regard to its common principles see: L. A. Zadeh, 'The concept of a linguistic variable and its applications to approximate reasoning, Information Sciences Vol. 8-9, 1975). It should be taken into account that fuzzy logic simply offers certain formalism convenient for discussing our problems, but it doesn't claim any special concept of philosophy of language or theoretical linguistics.

The convenience of fuzzy logic for us consists in the fact that it appears to be an apparatus developed specially for describing intensional meanings and their representation in signs.

Objects that fuzzy logic is usually applied to are much simpler than real narratives but they may be their model, though rough but principally more precise than models built of the objects traditionally studied by linguistics (of words and phrases). 
Basic notion of fuzzy logic is so called linguistic variable. This is nothing else but what we have been hitherto calling an intensional meaning: that is, a certain meaning that may be formulated in different ways.

In order to define linguistic variable it's necessary to define five parameters: its name (that is, common name of respective notion), universal set U where it can assume values, so called term set of terms corresponding to this variable, and also two functions: G and М.

Function G is so called syntactic rule determining the correspondence between the linguistic variable and its terms that is, each element of its term set taken separately. A notion of term of linguistic variable, for all its seeming proximity to Frege's notion of sign, is a notion specific for the apparatus of fuzzy logic and, moreover, such that allows to take an account of a word in a context: in the context of given narrative, and not in the given language as a whole.  

The value of each term is defined by the function М, so called semantic rule determining limitations for the set U: limitations corresponding to the scope of values of each term of a term-set.

Functions G and Мwithin the limits of the apparatus of fuzzy logic are not calculated in any way but simply defined proceeding from the defined sense. Thus, it's namely sense, that is, intensionality that appears to be the initial content formalized with the means of a notion of linguistic variable. 

Let's consider an example as primitive as possible: linguistic variable of "age".

It may be defined in the universal set of the amount of years, from zero to infinity, whereas for practical goals it may be defined in a shorter interval: for instance, from zero to a hundred years. The respective term set is constructed in accordance with the function G that also depends on the practical goal for whose sake we started talking about age: and not on the structure of natural language we use. For instance, we may speak of age only in the terms of "young"/"old", completely ignoring intermediate gradations, though any natural language could offer us rather nuanced scale; but we may also speak of age in the terms of age groups, as fractional as it may be, and it doesn’t necessarily presume an artificial situation: for instance, the entire tribal order of some peoples in the Eastern Africa is based upon the stratification of community with the means of several age groups with about eight-year intervals between them; in the languages of these peoples a binary opposition of "young" versus "old" may make no sense at all.

Thus, term set is a certain "division" of the value of linguistic variable. A specific mode of this division is defined by the semantic rule M, also proceeding not from qualities of natural language but from expedience in each separate case. Semantic rule defines the age interval that should be attributed to each term of the linguistic variable "age". In the examples given by Lotfi Zadeh, values bigger that 50 years are considered for applying to the term "old", but for the mentioned peoples of Eastern Africa defining values of human age for the groups older than 50 would be even more useless than expanding values of universal set of values beyond a hundred years in Zadeh's examples

It's very important to realize that Мand G aren't functions of each other but are defined together when a given linguistic variable is being defined. Thus, only expert can estimate the compatibility of notions of terms "great author" and casual "Fedya" with the person of Dostoevsky from the specific narrative (historic or, maybe, fictional). Say, in the biblical narratives the degree of compatibility of terms referring to animals with a quality of being "I" (be a speaker) is very low: yet a bit bigger than zero, therefore the case of Balaam's ass is described as a miracle. It's evident that in folk fairy tales, particularly archaic ones, it wouldn't look a miracle and hence would give no ground for building up a plot. The same quality of being a speaker in the biblical narrative is still less compatible with terms referring to unanimated world, and that's what the power of the expression "the very stones will cry out" (Luke, 19, 40) rests upon. What's all the listed limitations have in common is that they're not inferred from anywhere but are just defined when linguistic variable is determined. We won't consider now mathematical formalism allowing defining such restrictions, as in the given case these details aren't important to us.

Thus, let's formulate the difference between the term of linguistic variable and Frege's notion of sign: term is neither a word nor any other object of natural language. It's a semantic unit belonging to a given text (narrative).

Only within a given narrative can functions G and M be defined: syntactic and semantic rules that actually define the content of each term.

A term of linguistic variable concerns the sphere of text linguistics (as its very hard to determine what's included in this sphere). Although those aspects of text that may be considered on the level of words and phrases (phonology, morphology and syntax but not text linguistics as such) are used, in particular, for expressing intensional content, they nonetheless don't concern the notion of term as such.

At the same time, such key notions of text linguistics as explicit and implicit context are taken into account by the functions G and M.

It's also important to notice that terms are in no way attached to their names that may by replaced with any synonyms or even numbers. The same age category (that is, the same term) may be defined with one word (for instance, with a name of relevant age group in Oromo language) or in a way of description, as an author of fictional work does wishing to specify the character's age.

Term define the internal organization of texture of narrative on the level that it may be studied within the limits of regular linguistics (phonology, morphology, syntax and many aspects of text stylistics).

Term is a notion that doesn't correspond to any possible word of natural language but only to a scale of meanings defined within the limits of a given narrative. Therefore it's an elementary unit of narratology and not of language.

Now we have to apply the apparatus of fuzzy logic to describing the problem of precision of reference. Here we'll take account of the special research on determining the degree of fuzziness of fuzzy sets and functions, found in the scholarly monograph by A.Ryzhov:  А. П. Рыжов, Элементы теории нечетких множеств и ее приложений. Изд. 2-е (web edition: http://intsys.msu.ru/staff/ryzhov/FuzzySetsTheoryApplications.htm  (М., 2003); paper edition (1st): М., 1998 (under the name: «Элементы теории нечетких множеств и измерения нечеткости»).

Semantic rule M defines the values of terms as intervals (subsets) in the set U. These intervals may cross, and that's what usually meant when "fuzziness" of fuzzy logic is mentioned. In fact, this is not fuzziness itself but its rather rough modeling, since in ordinary fuzzy sets a function of membership ? of each term to each of the intervals. A function ? changes in the range between 0 and 1 inclusive. Its sense is in indicating the degree of conformity between a given term and a certain interval of values. For instance, reckoning that the term "old" undoubtedly refers to all persons older than 70 and undoubtedly cannot extend to persons younger than 50, this will mean that the function of membership of the term "old" is zero for interval from 0 to 50 and has a value of 1 for the interval from 70 and more. In the intermediate interval, that is, interval from 50 to 70 this function will assume various fractional values that may be defined as either constant values for smaller intervals between 50 and 70 or as some continuous function in the same interval.

Now let’s consider how the apparatus of linguistic variables may be applied to describing the problem of precision of reference.

For this purpose let's consider three term sets corresponding to the linguistic variable "blue". In the first of them there's only one term (correspondingly, only definition of presence of absence of a blue color is presumed); in the second, there are several synonyms differing by nuances of meanings: "raven blue, deep blue, turquoise etc"; in the third, there is a set of several dozens of narrow intervals of values of light wave lengths corresponding to a blue color.  

With this set of notions (and also with an expert's eye and a spectrometer) we proceed to classification of objects on the basis of presence of a blue color. The narrower will be the class to which we refer a specific object, the more precise will be an indication of this object. The class is designated with a respective term. The class, in its turn, will be more or less narrow depending on the number of elements contained in the term set that includes the term designating this class. In plainer words, the more precise is a scale our instrument is equipped with, the more precise will be our classification of objects, that is, our reference. However, a scale more precise is a scale featuring more possible values. 

Say, determining the difference between "sea blue" and "turquoise" isn't a simple task, and experts' opinion here will be not just subjective but even uncertain: whereas spectrometer allows determining a hue for certain. 

In the apparatus of fuzzy logic that will mean that the precision of determining the function of membership ? for each separate term will depend on the "scale calibration", that is, on the number of terms in the respective term set.

In most of the contemporary applications of fuzzy logic the membership function is taken for ordinary, that is, clear function. Yet this is only a particular case. In a more general case we should speak of sets that are sometimes called hyperfuzzy, which means that in such sets the membership function itself is fuzzy (graphically it's designated with a trace instead of a line). This is also approximation, but such that allows a rough account of uncertainty appearing when an attempt is being made to bring certain objects into correlation with semantic units of the respective narrative (terms).

On the qualitative level this rule is clear, though we haven't got means for inferring exact formulas. The case is that the fuzziness of the membership function ? will be the lesser the more terms n will be included in the term set of the respective linguistic variable.

If ? ? designates a change in the fuzziness of membership function during the transition from one term set to another, while ? n designates the respective change of terms in term sets, then we can infer a kind of Heisenberg's uncertainty ratio:

?? ?n? const

This expression means that the degree of uncertainty of reference in a given narrative depends on qualities of its terms: the more elements their term sets contain, the lesser is fuzziness (uncertainty) of each separate term, whereas measure of this mutual dependence (an analogue to Planck's constant) is a fundamental semantic characteristic of a given narrative.

Let's emphasize that we speak namely of narrative and not of language, hence the discussed rule relates to narratology and not to linguistics.

With a certain approximation it might be said that the more precise a reference we need, the more words will be needed. That's what constitutes the uncertainty principle for narratology.  

If we want to formulate this thought more correctly, we shouldn't speak of larger amount of words but of larger number of contextual ties of all sorts, not only explicit but implicit too. Only in the case we decide to explicate the implicit contextual ties, we'll find it possible to get more words when the precision of reference is increased. However, for instance, the art of poetry in many respects consists precisely in making an explication minimal.

Since we chose the way of interpretation of fuzziness inherent in narrative by analogy with the uncertainty in physics, we have no choice but to compare an act of reference with a process of measurement or observation. That's what we actually proceed to, after having a little break concerned with necessity of specifying some of the already introduced terms.

Dolezel, meaning for the most part European literary fiction of the New Age, defined texture as a precise verbal expression of text. Such definition only approximately applies to the most of works distributed before and even after the book-printing was invented: everyone who has at least basic knowledge of textology knows that the text of any work is given us in the form of manuscripts and printed editions that almost inevitably (and absolutely inevitably in the case of manuscripts) differ between them. A textological task of reconstruction, by the means of critical edition, of the "initial" ("author's") text is, by its essence, striving for impossible that may be praised only if this impossibility is being aware of.

Therefore, speaking of "precise" verbal form of a work, we speak of something initially approximate.

However, some of the narratives aren't too much attached to a specific verbal form. This refers, for instance, to folk legends not having a poetical form, and also to hagiographic legends while they function as tales, either spoken or occasionally registered on paper.

The content of such legends is easily translated to foreign languages, which shows that it's not determined with a precise verbal form.

The respective layer of text organization is easily detected also in the narratives to which the notion of precise verbal form is applicable. We do it each time when retelling their plot. Detailed retelling of the plot of any work allows detecting the level of its verbal organization that may be autonomous from the language this work is written in.

Such retelling, though inevitably wrapped in the words of some language, will not, on its own, consist of the words of natural language, but only of terms that we mentioned above. The same term may be designated with any synonym in any language, and also descriptively or with help of poetical tropoi.

Let's emphasize that, speaking of terms we speak namely of texture and not of the content of narrative.

The plot is a certain unity of both extensional and intensional content of narrative (we'll discuss it in detail below), abstracted from the specific verbal form of this narrative. However, for the purpose of retelling the plot we need some namely verbal form anyway. This must be a verbal form that doesn't consist of words. It consists of terms.

Term differs quite obviously both from an object of reference (since it's not an object but only designation of object) and from intensional (since it's not a notion of class of objects but only its name: but as a name it's "semantic", that is, possible to translate and not existing within the limits of one language only).

It will be yet more precise to say that term is a point in the instrument's scale. Instrument's scale is, in this case, a full term set to which a given term belongs.

It's clear that points on different scales may look similar, but their value is entirely determined by the calibration of scale. In the same way, terms that may be designated by the same words, may mean completely different things. The point on instrument's scale can't exist apart from this scale, and in the same way term can't exist apart from the term set it presumes. Yet term set is a characteristic feature of a given narrative.

Term is a basic element of narrative texture that is necessarily inherent in the narrative. The narrative of natural science, and also prosaic legends and spoken tales can altogether do without any other elements of texture organization, whereas term sets of their terms are most explicit. An ideal of fully explicated term sets is an ideal of Carnap's extensional language (see above). The more implicit presumed term sets of a narrative become, the stronger is an emphasis on the "external" elements of texture: that is, the more poetic and impossible to translate a narrative becomes.

It's possible that in the last years somewhere on the verge between narratology and text linguistics have already appeared researches that would allow serious investigations into what we have called terms: elementary "bricks" of narrative texture. However, they are unknown to us, therefore we, without claiming a serious research in this area, had to introduce at least elementary specification of logical notions that hitherto used to drop out of sight of all schools of narratology and linguistics that we're aware of. Most likely, it comes from the points of view traditional for these disciplines: after all, in order to see terms we had to move the position of the "observer" to the intensional apex of Frege's triangle (see above, in regard to the notion of linguistic variable

However, one formula of approach similar to what we suggested earlier. It's limited to a single type of narrative (mythological) and uses no special logical apparatus but enlists observations valuable for us as well. The author is Claude Levy-Strauss, and here's what he says in his pioneer "Structural Anthropology" (1958):

«…the place occupied by myth in the line of other kinds of lingual statements is quite contrary to that of poetry, however close they may be. Poetry is exceptionally difficult to translate to other language, and any translation is fraught with numerous distortions. On the contrary, value of myth as such can't be eliminated even by most unsuccessful translation. <…> The reason is that the essence of myth doesn't consist in style, form of narration or syntax, but in the story it tells. Myth is a language, but this language works on the highest level on which sense manages, if it's possible to say so, to detach itself from the lingual basis on which it was formed (C. Levy-Strauss, Structural Anthropology).

In our opinion, it would be more precise to speak of myth not as one of the kinds of "lingual statements" but as one of the kinds of narratives (and of poetry as another kind of narrative). Language of myth is a language of bare terms, that is, myth is built as a narrative containing nothing but the bare basis, whereas poetry, on the contrary, is a narrative where there is a maximum of possible formations upon this basis. From this point of view (organization of texture as a single bare basis made of terms) myth doesn't differ from various legends, hagiographic in particular, or, for example, spoken anecdotes. Probably, lack of development of this Levy-Strauss' approach in the subsequent narratology, where structuralism dominated, has something to do with Levy-Strauss' terminological fuzziness, principal for him as a structuralist: calling myth "a language" he implicitly denies it extensional content (reference differing from autoreference), and then separates this non-existing language from the "lingual basis" (natural language) that, generally speaking, happens to be an abstract notion that doesn't really exists out of narratives. It would be closer to reality to say that myth uses natural language as its own language not in its full, but only in its basic form, that is both necessary and enough for creating any narrative. 

Reference as observation and experiment

There can be theoretically two sources of uncertainty: stochastic (probabilistic) character of processes, and fuzziness of certain (or all) parameters of these processes.

The typical example of probabilistic uncertainty of phenomena of the physical world are the laws of classical physics in statistical form, as Bolzmann started to record them. Similarly to Galilei's physics, Bolzmann's statistic physics works with ideal (and non-existing in real nature) objects, but a model is one step closer to reality by at the expense of recognizing the impossibility of full determinism, even on the micro-level.

Probabilistic interpretation of quantum mechanics proposed by Schroedinger has a different meaning, along with Bolzmann's meaning: due to the presence of Heisenberg's uncertainty ratio in quantum mechanics. Interpretation of this sense lies in the sphere of research programs and exceeds the limits of physics as such.

According to the Copenhagen interpretation, Heisenberg's uncertainty ratio determines the minimal level of uncertainty that has no probabilistic meaning (though may be represented through a wave function having mathematical form of probability function) but reflects the uncertainty of physical reality itself, that is, its fuzziness.

Presence of such kind of fuzziness in reality is otherwise expressed in the impossibility of speaking about reality apart from the situation of observation (measurement). The result of each separate observation (measurement) is always distinct, but an unavoidable scattering of results of repeated observations witnesses that fuzziness is a quality of reality itself, as it submits to Heisenberg's uncertainty ratio.

All this exactly corresponds to our problems of reference.

Above we tried to speak of intensionality and its expression (terms) digressing from reality as such, that is, from acts of reference. Therefore we naturally arrived at the expression analogous to Heisenberg's principle that also speaks of reality preceding the situation of observation.

However, none of our mental structures responsible for our ties with reality exist without this very reality, and we need them only for performing a reference.

Performing a reference we arrive at the same ambiguity, the same scattering of data as those peculiar to observation (measurement) in quantum physics.

Often procedure of reference provides us with a certain unambiguous understanding of each narrative yet it would never coincide precisely neither with understanding of another person, even his readiness to perceive the given narrative were none lesser than our own, nor even with our own understanding after reading this narrative over some time later. It particularly concerns fictional works and poetry, for which reason we read them over and over again.

Quine's principle of ontological relativity displays at the qualitative level, while Putnam's theorem offers a strict logical prove, that reference can never be logically defined. However, we perceive the results of the acts of reference namely as certain (to this extent or another). As for the uncertainty of reference, it becomes apparent, just like Heisenberg's uncertainty, only in scattering of the results of the acts of reference.

Thus, human perception of reality appears to be similar to perception of physical measuring instrument, whereas our narrations about reality (narratives) appears to be similar to theories of natural science, which actually forces us to speak of general narratology including the narratives of natural science. There's nothing here to wonder about, because it would be more precise to say that measuring instruments constructed by human are nothing else but extension of his natural sense organs: if it were otherwise, then we couldn't interpret the results of measurements, yet it couldn't be otherwise anyway, since it's not likely that human is able to invent anything absolutely detached from his consciousness and experience.

The notion of natural sciences as sciences working with precise magnitudes and invariantly repeating measurements results has been a dream of the founders of European science of the New Age, yet in our era it's but a relic superstition, prevalent mostly among humanitarians.

The similar superstition is a thought that the narrative of natural science deals exclusively with existing external objects, whereas fictional narrative may involve both really existing (for instance, St. Petersburg) and fictive (for instance, Raskolnikov) objects. This thought appears most unsuccessful when it accompanies the attempts of building up referential poetics, as in the case of Dolezel who, regretfully, shares positivist view on natural sciences. This error prevents the person from being fully aware of the logical basis of even his own theory of poetics.  

Everyone remembers more or less clearly that science once operated with certain notions it later denounced as false: for instance, flogiston in chemistry or various unthinkable monsters in zoology. However, positivism, following Russell, reckons that those were just faults of science that are not necessary inherent in it.

In fact, it may be said, together with Russell, of any separate "mistake" that it's not necessary for science. However, presence of such "mistakes" as such is a necessary quality of scientific method.

For instance, we don't regard as mistake the presence of the notion of "electron" in the theories of electricity prior to the year 1913. However, the notion of electron proceeding from Bohr's quantum postulate, suggested in 1913 and subsequently accepted by science, led to differences far more serious than, for instance, the difference between dragons and pterodactyls: Bohr's 1913 model electron still remained a particle flying around an atomic core but it acquired a quality impossible for a normal particle: leaping from one orbit to another by passing the space between them. None of the physical objects of both classic physics and our daily experience possessed such a quality. However, this change of the notion of "electron" hasn't been final as well: after 1927 an electron finally ceased being a particle in a usual sense of a word and lost any possibility to be visually represented in our mind; instead, it retained a representation corresponding to the particle-wave dualism and the principle of complementarity. By the same time (in 1926) electron acquired yet another quality, spin, that may be presented to our understanding yet more conditionally (as spinning of particle around its own axis; however, its impossible to say that an electron, in a strict sense, is a particle, hence it can't possess an "axis" to spin around). With further development of the doctrine of elementary particles even such degree of visibility had to be renounced. Thus, when we speak nowadays of "flavors" and "colors" of quarks, it doesn't refer to the notions of flavor and color from our world any more than in Christian triadology "birth" refers to birth and "emanation" (literally, "evaporation") to evaporation

All three mentioned electrons, whether that prior to 1913, that of the quantum postulate period, or that of the quantum mechanics era, kept the same name but appeared to be completely different objects. The electron of classic physics was outwardly even less successful approximation of the electron of quantum physics than pterodactyl could be of dragon.

Addressing to Russell, we might wonder why in regard to chemical theories of 18th century flogiston should be considered a mistake and not an "approximation" to the notion of oxygen offering the best explanation of experimental data of the day. If the last is possible, then Russel made a mistake having put all imagined objects in the category of false and non-existing. Human cognition never possesses an absolute knowledge of the true and the existing, but uses various approximations to both of them.

In the modern philosophy this thought was most distinctly articulated by Quine, as early as in "Two Dogmas of Empiricism" (see above):

"The totality of our so-called knowledge or beliefs, from the most casual matters of geography and history to the profoundest laws of atomic physics or even of pure mathematics and logic, is a man-made fabric which impinges on experience only along the edges. Or, to change the figure, total science is like a field of force whose boundary conditions are experience. A conflict with experience at the periphery occasions readjustments in the interior of the field. Truth values have to be redistributed over some of our statements. Re-evaluation of some statements entails re-evaluation of others, because of their logical interconnections -- the logical laws being in turn simply certain further statements of the system, certain further elements of the field. Having re-evaluated one statement we must re-evaluate some others, whether they be statements logically connected with the first or whether they be the statements of logical connections themselves. But the total field is so undetermined by its boundary conditions, experience, that there is much latitude of choice as to what statements to re-evaluate in the light of any single contrary experience. No particular experiences are linked with any particular statements in the interior of the field, except indirectly through considerations of equilibrium affecting the field as a whole. <…> Total science, mathematical and natural and human <…> [is] underdetermined by experience. The edge of the system must be kept squared with experience; the rest, with all its elaborate myths or fictions, has as its objective the simplicity of laws".

Those who may find it hard to follow Quine's reasoning about the correlation between empirical and theoretical knowledge, the "edge of system" and "all the rest", are invited to run over it's summary in the short rhyme by Oleg Grigoriev (1943—1992):

Я ударился об угол —
Значит, мир не очень кругл.
[I hit the corner and so found
That this world is not so round]

Without referring specially to Quine, Thomas Kuhn presented analogous considerations for the more particular case of methodology and history of science. However, Quine speaks not only of science but of any form of human knowledge. And even more than that, he had almost formulated what we would like to explicate finally: that all the above said refers to artistic form of cognition as well. We have already told above that an artistic activity is namely a form of cognition.   

The problem of unambiguity of reference is in fact a problem of interpretation of experience that Quine speaks about. From the logic point of view, an unlimited multitude of interpretations is possible, yet we choose that which won't tear apart the tissues of our present knowledge. And if this tearing of tissue is quite inevitable, than we choose the way allowing to sew it over as easily as possible (however, if we liken the totality of our knowledge to tissue, it must be said that this is a special tissue that can't be darned yet can regenerate, analogously to cellular tissue of live organism).

The result of any act of reference shouldn't necessarily appear false, just as the result of any experiment isn't necessarily false. The matter is just that the notion of "verity" of result in both cases, be act of reference or experiment (and act of observation as well), is neither absolute nor "objective". Like all that's human, it's relative and in part subjective.  

In order to explain this thought more strictly we'll have to compare intensional functions with scientific programs. We will develop that point below. And in the meantime, here is a simple illustration to what we speak of here:

This picture represents a single object of reality with two different tissues of theory stretched over it like a fabric across the frame of a loom. A collapse of narrative uncertainty corresponding to the principle of uncertainty for narrative formulated above and analogous to the collapse of the wave function after measurement, will come when we read the drawing either top to bottom or left to right.

The drawing shows two things simultaneously. (1) In order to become comprehended, that is, interpreted an object should be represented by the term from a certain term set corresponding to a certain (initially defined) intensional content. In absence of any context, the depicted figure located at the intersection of vertical and horizontal rows could be interpreted in many different ways (for instance, as an architectural plan). Each of the two rows appearing on the picture, both vertical and horizontal, is a representation of a term set. However, the picture shows also that (2) term sets can be chosen in a very different way for the same object: it depends only on the initially defined intensional.

This illustrates the fact that intensional meaning of narrative is able to construct an object of reference, thus achieving an unambiguity of reference.

Gradations of the levels of beings and the question of reality of giants

"At this point they came in sight of thirty forty windmills that there are on plain, and as soon as Don Quixote saw them he said to his squire, "Fortune is arranging matters for us better than we could have shaped our desires ourselves, for look there, friend Sancho Panza, where thirty or more monstrous giants present themselves, all of whom I mean to engage in battle and slay, and with whose spoils we shall begin to make our fortunes; for this is righteous warfare, and it is God's good service to sweep so evil a breed from off the face of the earth."

"What giants?" said Sancho Panza.

"Those thou seest there," answered his master, "with the long arms, and some have them nearly two leagues long."

"Look, your worship," said Sancho; "what we see there are not giants but windmills, and what seem to be their arms are the sails that turned by the wind make the millstone go."

"It is easy to see," replied Don Quixote, "that thou art not used to this business of adventures; those are giants; and if thou art afraid, away with thee out of this and betake thyself to prayer while I engage them in fierce and unequal combat."

We don’t' think that many recognized this quote therefore we find it necessary to inform that this is Dolezel. (р. 148). He uses this example so as to introduce the notion of different levels of being": what we have called above gradations of being. Thus, according to Dolezel, one level of being is the fictive possible world of the novel "Don Quixote" created by Servantes and adequately perceived "from inside" by Sancho Panza, whereas another level is the special, fictive in regard to fictive, world of Don Quixote himself.

However, to tell the truth, Dolezel tendentiously excluded from his quote part of the last Don Quixote's sentence, so that we can't throw off the suspicion that he did it because of the persuasiveness of its tone, that is, its intensional content, that caused him intellectual discomfort:  "It is easy to see... that thou art not used to this business of adventures…"

Dolezel considers the dialogue of two characters of the novel representing two different interpretations of reality. The question is: which one of the two is right? Right is he whose point of view coincides with the author's. Since the author said beforehand that Don Quixote and Sancho Panza approached the windmills, and not the giants, Sancho Panza is right. In the fictive world of "Don Quixote" (a novel) windmills exist, but giants don't. Giants exist only in Don Quixote's imagination. Therefore giants and windmills have different ontological status, though in both cases this status is fictive. Thus in possible worlds of fictional narrative being can have gradations.

Dolezel considers these gradations an exclusive characteristic feature of fictional narrative. We have already written why we think that it's not so: why such gradations of being are inherent in all possible narratives. However, now our task is more detailed examining of ontological difference between windmills and giants. It might seem that no great perspicacity is needed to realize how much actual is comprehension of this difference for hagiography. After all, "Don Quixote" was written by Servantes as a parody on chivalry novels, whereas chivalry novels emerged from hagiographic novels like "Romance of Alexander" and its versions that we've mentioned above. Giants can be met, for instance, in some Scandinavian sagas retaining characteristic features of hagiographic legends along with characteristic features of chivalry novels that influenced them (for instance, in "Saga of Ingvar the Far-Traveled": this saga is, in particular, also a hagiographic legend, since it tells of establishing the cult of Ingvar)…

The solution of the problem of ontological status of various fictive objects suggested by Dolezel may be acceptable for specific cases of literary analysis, but in general it's a simplification. It's based on the principle of giving credence to an author's conscious will. However, such principle can do only as a basis for working out an approximate solution. On the whole, as we know, although the process of artistic creative work as well as creating a narrative in general is always intentional, it may be controlled to a certain extent, and may not, by an author's consciousness (see above). Therefore an author's conscious point of view, even if we happen to become aware of it (in Servantes' case, of course, this is so), is undoubtedly not an only way of interpretation of the narrative possible world he created. In the postmodernist literary tradition (Borges, Pavic) an author openly renounces the choice between his characters' different points of view on reality. However, in any literature a reader retains his right to his own, different from an author's, point of view.  

If it were otherwise, creating fictional works wouldn't be like mental experiment and scientific research but more like publicism: that is, chewing over for the public a content well known to an author beforehand.

However, the world of fictional narrative is beyond all doubt richer than any simple scheme. In regard to the windmills of Don Quixote, Sancho Panza and Servantes this fact is well illustrated by the analysis made by the contemporary Mexican artist Octavio Ocampo, though only in the way of graphics:

Ocampo takes account of many different points of view at once. And not only those of Sancho Panza and Don Quixote, but also such that makes no difference between the giants, the windmills and the quarreling characters themselves. Therefore it can be said that Servantes himself represents the external world of the possible world of his characters. That's why, probably, the artist depicted him in the upper right corner of the picture and in a different color, in a way resembling the Russian icons featuring mundane scenes (like the Assumption) where angels are depicted as belonging to another reality. Nonetheless, Servantes' characters and Servantes himself became a part of inner possible world of other people, therefore the artist place them beside Servantes, as part of the world external to the possible world of 'Don Quixote".

For Dolezel the dividing line between artistic fantasy and reality runs accurately through an author, that is, Servantes. However, the artist shows that to him and to us, including Dolezel himself (but as a reader and not as a literary theoretician), Servantes is a part of an envelope of the possible world of "Don Quixote": and an inseparable part, as it seems. Dolezel's (and our) opinion that Don Quixote is more fictive than Servantes is based on casual facts we happened to learn regarding the circumstances under which the novel was written. However, without this "backstage" knowledge we would think that Don Quixote of the novel had a historic prototype just like Napoleon in "War and Peace", whereas Servantes, on the contrary, is but a legendary figure just like Hermes Trismegist to whom the authorship of Corpus Hermeticum is ascribed.

One might object that even if Servantes were a mythical figure, "Don Quixote" would be related to some author anyway: and we readers have to believe this author in regard to internal organization of the possible world he created. We ought to agree with this argument but not without limitations: in other words, if we trust an author no more (and no less) than a scientist interpreting his own experiment.

As is well known, a huge amount of experiments were interpreted correctly, from the point of view of contemporary science, by those who didn't conduct them at all. For instance, one of the two discoverers of oxygen, Joseph Pristley who published his discovery in 1774 never accepted the oxygen theory of combustion (developed by Antoine Laurent Lavoisier) that emerged as soon as next year and was based on the materials of his discovery. Pristley did much to secure the old theory of combustion on the basis of flogiston for several more decades, whereas Lavoisier in the face of a guillotine was bereft of a comfort to see the triumph of his nearly principal idea in chemistry… 

Coming back to the author of "Don Quixote", now anonymous, we should acknowledge that he cannot be more trustworthy to us than Pristley the great chemist or some medieval annalist who writes down the events in accordance to his own comprehension so that giants, along with other characters-in-the-play, may feature there quite openly (as in the case of Snorri Sturluson (1179-1241), chief Icelandic annalist of Norwegian history and an author of Heimskringla). It's possible that the author of "Don Quixote", for instance, just didn't know that wandering knights he tried to mock with his Don Quixote image are able to penetrate the essence of things with their sight and see what other mortals can't. Therefore the author just reserved the point of view of a philistine who can't see and discern the danger coming from giants pretending to be windmills. 

The last case may be considered rather close to hagiography. An attitude to "holy fools" [yurodivy] usually described in their "Lives" rather resembles that of many of "Don Quixote"'s characters and readers to its main character. Sometimes this attitude, hypercritical or negative, is preserved for us by sources quite independent from hagiography and even rather extensive at times. While abstaining from judging this example by its essence, let us mention Ivan Yakovlevich Koreysha (1783—1861) who became in 19th century a ‘mythopoetical’ character, either main or minor, of innumerable amount of historical, publicist, fictional, hagiographical and satirical works. In particular, he is depicted, not without skepticism, by Dostoevsky as yurodivy Semen Yakovlevich in "Demons" (Semen Yakovlevich is a collective image on the basis of Koreysha enriched with an element of the name of yet another popular yurodivy of the time, Semen Mitrich); while Leskov dedicated to him, under his own name, quite a satirical story "A little mistake". Koreysha had been favored posthumously with a number of short "Lives" and one rather extensive see the excellent collection in: Юродивый Иван Яковлевич Корейша / Сост. А. Ф. Киреев (М., 1894). Moreover, Koreysha had been honored with quite a rare occasion to become a character of ????? ("speech of blame", in ancient rhetoric a genre strictly opposed to panegyric at whose heels it also passed on the sly into hagiography). For reasons that we'll explain below psogos as an (anti-)hagiographic genre is very rare, and only those of saints who really annoyed some church party were "honored" this way: but even this used to happen rarely (the classic example is "anti-Life" of St. Maximus the Confessor written in 7th century in Syriac language by a monothelite Sergius of Rashayn). The first essay in the book (“The twenty-six false prophets and false holy-fools from Moscow”) - "Двадцать шесть Московских Лже-пророков, Лже-юродивых, Дур и Дураков" (М., 1864; many reprints, including the recent ones) by I. G. Pryzhov has been dedicated to Korysha and actually became his "anti-Life". Pryzhov, for all his disgust for his character, doesn't forget to mention a couple of miracles ascribed to him without questioning their factual side.  

Koreysha had been quite a historic person yet he could be a fictional but collective image, in which case we could draw a very close parallel between him and Don Quixote. If so, which one of Koreysha's biographers is ought to be likened to Don Quixote's biographer, Servantes?

It's obvious that the answer will depend, by the way, on how we generally regard noble wandering knights and their deeds: and also those who dare to mock them without understanding their deeds. It's quite possible that we may find reasons to interpret the encounter with windmills otherwise than Dolezel did. After all, it's no less possible that there was real fight with giants behind this encounter: although, of course, it wasn't a literary collective image of Don Quixote who fought them but historic knights whom Servantes' character imitated.  

Most flippant objection retained by those wishing to fix a dividing line between real and fictive on the author of a narrative, be it only fictional, is an a priori statement that giants supposedly don't exist. We call it "a priori statement" because there can be no possible experience referring to which such statement might be regarded as a posteriori. Yet being a priori still doesn't make a statement false, therefore we have to ask a contra question: if giants don't exist, what does then exist?
That's what we are going to consider in the coming pages.

We'll have to discuss what it means: "exist" and "don't exist". Won't giants turn out to be an "approximation" of something just as we suggested above considering "flogiston" an approximation of the notion of oxygen? Indeed, why should giants be worse than flogiston?

Russell didn't acknowledge that being may have gradations, therefore to him giants and flogiston had equal rights to be real, and these rights were equal to zero. However, if we're able to realize that at least our notions of objects have varying measure of ontological status, then the reality of giants is ought to be discussed with all the seriousness and with maximum of logical stringency. It's particularly advisable for hagiography where, along with giants, dragons and other no less "approximate" creatures can be found in plenty.

We hope that we made it clear enough that Servantes' opinion regarding the story with windmills concerned us only in the way of appreciating his tendentiousness as an interpreter of the mental experiment "Don Quixote" (namely as interpreter and not as an author: authorship of experiment doesn't matter since now the possible world of "Don Quixote" may be observed by anyone), and also in the way of understanding the intensional content of the narrative he created. As for what matters, an author's opinion has no privileged status. 

Constructing reference and principle of rational charity

Literature: Donald Davidson, Inquiries into Truth and Interpretation(Oxford UP, 1984) (especially the essay On the Very Idea of a Conceptual Scheme (1974), p. 183—198).

Thus, the world of narrative, as well as the world in general, appears to us as it's depicted on Octavio Ocampo's picture: we may see there giants and windmills, flogiston and oxygen, electrons prior to 1913 and electrons of 1913-1926 issue, as well as electrons with contemporary modifications: and all this depends exclusively on the point of view.

And it's not because Berkeley is right indeed and there's nothing in external reality besides our sensations, but because our conceptualization of external reality is always approximate, in accordance to Quine's principle of ontological relativity and Putnam's internalism. However, "approximate" doesn't mean "arbitrary". It's always coordinated with present experience: though only along the edges of the same fabric stretched across the frame of the loom, to which Quine likens all the totality of human knowledge (see above).  

Of course, according to Servantes, Don Quixote's giants turned out to be incompatible namely with the data of his experience, that's the point in the encounter with giants. Yet once again, we can't affirm that we have to believe to Servantes who could (theoretically) misrepresent the facts referring to the real prototype of Don Quixote; and even if Servantes were right, we would have to assume logically that the reason of Don Quixote's defeat in the fight with windmills is that he nonetheless wasn't a real wandering knight (like Koreysha who, according to Pryzhov, wasn't a real yurodivy for Christ's sake). Probably, a real wandering knight's encounter with giants would end as laws of the genre demand. It's possible also that our last assumption is wrong: maybe, wandering knights eventually failed to win over the giants, therefore by the time of Servantes they became completely extinct and their memory became an object of mockery, just as it happened in the Bible to Semitic pagan gods after the triumph of monotheity.  

Anyhow, the randomness of our interpretations is limited on one hand with our experience and on the second, with a bid for minimal harm done to the fabric of our knowledge as a whole. As for the rest, sometimes we find it hard to choose between giants and windmills, as well as between oxygen and flogiston: and this hard circumstance is compensated only with the fact that this uncertainty of the choice of interpretation lasts only until our experience forces us to settle down. Thus, during Lavoisier's life his experiments in favor of the oxygen theory of combustion were approximately counterpoised by Pristly's experiments in favor of flogiston. However, just as this example shows (and Thomas Kuhn worked out in detail), no scientific experiment can prove "decisive" on its own but only due to the subsequent convention depending, in its turn, on changes in the global outlook: the very "fabric" of our knowledge Quine wrote about.

Here arises a question of relation between our empiric data and those "essences" that we start to perceive in the external world. Quine answers this question thus (in "Two Dogmas of Empiricism"):

"As an empiricist I continue to think of the conceptual scheme of science as a tool, ultimately, for predicting future experience in the light of past experience. Physical objects are conceptually imported into the situation as convenient intermediaries -- not by definition in terms of experience, but simply as irreducible posits comparable, epistemologically, to the gods of Homer. Let me interject that for my part I do, qua lay physicist, believe in physical objects and not in Homer's gods; and I consider it a scientific error to believe otherwise. But in point of epistemological footing the physical objects and the gods differ only in degree and not in kind. Both sorts of entities enter our conception only as cultural posits. The myth of physical objects is epistemologically superior to most in that it has proved more efficacious than other myths as a device for working a manageable structure into the flux of experience".

True indeed, as the famous Russian post-modernist poet said:
Там где эллину сияла красота,
Мне из черных дыр зияла пустота…
[Where the Hellene saw a shining beauty
Emptiness gaped at me out of black holes]

Quine writes of nothing else but our problem of giants. Everyone, it seems, would agree that Homer's gods and giants are of the same line. Even in hagiography, in spite of the compulsory monotheity, Homer's gods often appear to be real (as demons).
Human thought, be it scientific or commonplace, is busy exclusively with creating "myths" ensuring better orientation in the "flux of experience".

In the same paper Quine gives yet another example: emerging of the conception of irrational numbers. Initially mathematicians are interested only in the real numbers; mathematical theories are developed so as to deal with them, but theories lack necessary simplicity. Thus, it's impossible to take functions of certain arguments: for instance, square root of two. This creates a difficulty that may be overcome in two ways: either by complicating the theory so that it could explain why square root of two cannot be taken, or by postulating special logical objects that would form a set of values of such functions of such arguments. What eventually emerges is a set of irrational numbers: one more "myth". With that, we keep putting real interest only to real numbers but the laws of algebra allowing us to deal with them may be essentially simplified if we keep pretending that irrational numbers exist along with rational ones. A theory was simplified at the expense of ontology that was complicated as one more kind of essences was admitted.

Quine's idea is that our knowledge in its development always strives to simplify theory, or using another comparison taken from the same article, to make a "fabric" of our knowledge as smooth and continuous as possible. This feature of the development of knowledge isn't peculiar to science but inherent in human thinking as such:

"Science is a continuation of common sense, and it continues the common-sense expedient of swelling ontology to simplify theory".

The same laws of thinking become apparent in fictional narratives as well, which is no surprise, considering what we've said above of literature as "well-conspired science". It's these laws that are responsible for the most important of intensional functions, and especially for what Dolezel called "saturation". However, we'll have to discuss it in detail a bit later on.

It's important to notice that intentional constructing (both conscious and unconscious) of objects of reference with a certain claim for ontological status isn't a feature peculiar namely to literature, although it's often ascribed to it. It's a peculiarity of human thinking as such, including a scientific thinking: what it always observes is only "swelling ontology".
In this regard analytical philosophy returned to the views of Giambattista Vico (1668—1744) on primary status of poetical way of expression in regard to prosaic (formulated in his "Principles of a New Science of the Common Nature of Nations", 1725). "Poetical views on nature" (we take a liberty to paraphrase A. N. Afanasiev) are peculiar not to Slavs only but initially inherent in humans in general, and science is just one of these views, no less poetical then all the rest. That's what the modern anthropology asserts as well: "Metaphor <…> isn't a late decoration of language but one of its fundamental varieties. <…> It forms an initial mode of discursive thinking" (C. Levy-Strauss, Totemism); said in the context of discussing the analogous view of Rousseau).

Formulating this conclusion in the terms introduced in previous pages we must say that an act of reference is always accompanied by an action of intensional functions. Therefore any object of reference, and not in quantum physics alone, is, to a certain extent, a function of an "observer". Reference acquires unambiguity due to the fact that we "accomplish" (or almost build up) in our mind its objects while directing our own efforts towards the simplicity of result for the benefit of our general outlook. In methodology and history of science it's called an action of research programs, and in art, an action of intensional functions. However, we would suggest usage of the last term far beyond the limits of poetics by applying it to all that we've included in narratology, that is, to historic narrative and even to the narrative of natural science as well.

If we speak of intensional functions in science, in regard to research programs, these are functions selecting facts and developing theories so that they would correspond to research programs, that is, intensionsal content of narratives of natural science (that is, scientific theories). Thus, in regard to science intensional functions are mechanisms used by research programs to build up their scientific theories.

We will call the act of reference that takes account of the action of intensional functions constructing perceptible objects, a constructing reference. Let's make clear that a constructing function is not a kind of special reference but only result of interaction between reference in the traditional sense and intensional functions.

We have yet to answer a question: how does mutual understanding occur between various "observers", that is, say, between an advocate of the "myth" of physical objects and an advocate of the "myth" of Homer's gods (the word "myth" should either be taken in quotation marks in both cases, or not, in both cases as well). This question was answered, while Quine was still alive, by his disciple Donald Davidson in his penetrating article "On the Very Idea of a Conceptual Scheme" (1973) that can be reckoned as a necessary sequel to Quine's "Two Dogmas of Empiricism". There so called principle of rational charity has been suggested.

Davidson illustrates this principle's formula and basis by the following example:

"…If you see two mast sailboat passing by and your companion says, "Look at this beautiful yawl", you face the problem of interpretation. It's natural to assume that your friend mistook a sailing boat for a yawl and so came to a wrong conclusion. However, if he's got a good sight and a proper point of view, it's more likely that he uses the word "yawl" not the way you use it and therefore he made no mistake in regard to the presence of a spanker on the passing yacht. We should constantly strive to keep interpretation off a blow, preserving rational theory of opinions. As philosophers we are particularly tolerable to systematical verbal confusion, wishing interpretation would give results. This process consists in constructing a vital theory of opinions and meanings out of sentences regarded as true".

Davidson's principal idea consists in reconciliation of two supposedly incompatible states of mind (for instance, recalling Quine, myth of physical objects and myth of Homer's gods) through the problem of interpretation (translation). We regard with "rational charity", if possible, to any formula concerning reality.

Regretfully, the principle of rational charity as it was formulated by Davidson appears to be fraught with its own difficulties even in the eyes of those who were ready to accept Quine's ontological relativism. Quine himself pointed them out.

Quine's first printed reaction on Davidson's article was wary and didn't exceed the limits of clarifying certain terminological confusions (it's curious that one of them was concerned with the term "conceptual scheme" principal for Davidson's article): that was an article "On the Very Idea of a Third Dogma" (W. V. Quine, Theories and Things (Harvard, Mass.,—London, 1981) 38—42). Only in one of his last works Quine resolved to discuss Davidson thesis by its essence (W. V. Quine, Pursuit of Truth (Harvard, Mass.,—London, 1990) 197—202).

It's evident that Quine meets Davidson's approach with big sympathy: he opposed it to the "sectarian" approach of adepts of one truth and calls an approach "a la Davidson" "ecumenical" (however, both terms are joke). Davidson, according to Quine, helps to understand that "limited by our human terms and means, we get to know this world in different ways". It may be compared with the problem of measuring the diameter of impenetrable sphere: we can measure it by inserting the sphere into appropriate holes of known diameters, or by measuring with a tape the length of circumference of its section and, subsequently, calculating the diameter; however, what we can't do in any way is get inside the sphere and measure its diameter from the inside.

However, Quine immediately acknowledges that so far he's been hesitating whether he should accept Davidson's approach or not; moreover, up to date he's still hesitating, so that for all the sympathy for Davidson he can't resolve to join him. Quine, by the way, has only one objection: since we all perceive the same external world, it's hard to believe that the only reason for discrepancies as serious as those arising between our interpretations is a problem of translation. Quine can't prove it yet he deems that apart from difficulties with translation of terms exclusively discussed by Davidson, there must be more serious problems about it all.

Though fully accepting the principle of rational charity we are nonetheless inclined to agree with this objection. We accept this principle as such but only if author's argumentation is essentially strengthened (in the light of the ontological ideas presented above). Since it's quite obvious that the question of reality of giants, goblins and sailing boats is one of the most important for comprehending hagiography, we are going to formulate our own point of view to which we immediately proceed.

Possible worlds of Fuzzy Kripke (FK) type

Literature: Nobu-Yuki Suzuki, Kripke frames with graded accessibility and fuzzy possible world semantics // Studia Logica 59 (1997) 249—269.

The principal difference between our position and position of most representants of analytical philosophy, including Quine and Davidson, is that we don't accept the thesis of the real world as just one, and no more than just one, of the possible worlds. On this thesis semantics of possible worlds in Kripke's edition is founded, and Quine never questioned this approach. Within the limits of analytical philosophy strictly opposite concept of possible worlds was formulated by David Lewis, yet we wouldn't accept it as well. Above we suggested our own version with reference to the recent works of A. M. Anisov who suggested an approach to semantics of possible worlds on the basis of fuzzy logic. Now, having observed both the problems of reference through use of the apparatus of fuzzy logic, and the problems interpretations in the light of Quine's ontological relativism, we can formulate our approach to semantics of possible worlds more strictly. In consequence, as we hope, fairly distinct reasons to trust Davidson's principle of rational charity will become apparent.

Thus, in our opinion, the real world shouldn't be considered as one of the possible worlds but as a totality of possible worlds. Here we disagree with Kripke (and also with Quine, Davidson and Putnam) and agree with David Lewis.
However, we disagree with Lewis in regard to his thesis of mutual impenetrability of possible worlds suggesting that possible worlds aren't accessible to each other.

Lewis needed this thesis to save the concept of objectivity of "objective reality" (here is, by the way, the brightest illustration of Quine's observation that a human easily reconciles with swelling ontology for the sake of simplicity of theory!). If each one of the possible worlds is obliged to have distinct ontological boundaries, then it's clear that there can be no "diffusive areas" between different possible worlds.

However, our approach to semantics of possible worlds was based on generalization of the Copenhagen interpretation of quantum theory: that is, the interpretation where uncertainties of values featuring in mathematical formula of Heisenberg's principle are taken for fundamental physical qualities of reality itself. Therefore according to our approach to semantics of possible worlds no possible world possessing full ontological unambiguity (presuming absence of uncertainty in "Copenhagen" sense) is possible at all. Thus, we initially invest in our semantics of possible worlds the very thesis that Lewis, just as Kripke and Quine, as well as Popper and Lakatos, equally tried to avoid… Historically multiworld semantics was applied to the philosophy of natural science only when something was needed to put against the "Copenhagen": whereas we, on the contrary, are formulating namely "Copenhagen" views through semantics of possible worlds.

Theoretically there may be two sources of uncertainty: either stochastic (probabilistic) character of processes or fuzziness of the objects (notions) themselves. According to the "Copenhagen" interpretation, stochastic character of observation results described by Schroedinger's probabilistic wave function proceeds not only from the probabilistic character of laws of classical physics (in Bolzmann's sense) but also from the uncertainty of reality itself that subsequently, when fuzzy logic appeared, was named "fuzziness". That's what allows to relate to the real world the version of semantics of possible worlds suggested by Anisov in his "logic of uncertainty" using the apparatus of fuzzy logic.

Regardless of Anisov, formal structure of the universe of fuzzy worlds that we discuss here was described by Nobu-Yuki Suzuki (gratis to P. Zusmanovich who offered me a possibility to become acquainted with this work).

Suzuki's system of possible worlds emerges by the way of modificating Kripke's system of possible worlds. Interrelations of Kripke's possible worlds are described by the means of so called Kripke frames. Frame F is a pair <W, R>, where W is a non-empty set of elements (possible worlds) of the frame F, and R is a binary relation to W called an accessibility relation. If for the elements x, y I W the ratio xRy holds true, this is called "x accesses y". Suzuki replaces R with partial function r assuming values in the interval [0, 1]. This operation with Kripke's modal logic is performed by analogy with the way L. Zadeh introduces the notions of his fuzzy logic by modifying the notions of classic logic. Function r thus becomes a factor of accessibility of y from x. Suzuki designates this fuzzy Kripke's modal logic he created as fK, where f means fuzzy, and K refers to the commonly accepted designation of Kripke's modal logic.

Guided by Suzuki's fK logic and using the expression of another Russian logician who also applied the apparatus of fuzzy logic to Kripke's modal logic (although he didn't specially consider the problem of semantics of possible worlds) we may call the type of possible worlds we've described "Fuzzy Kripke", FK (see: А. М. Миронов, Нечеткие модальные логики // Фундаментальная и прикладная математика 9 (2003) № 1. 201—230).

If the real world is a totality of possible worlds (however, of those endowed by fuzziness, that is, having no distinct boundaries), then between different possible worlds different degrees of intersection. Possible worlds behave analogously to the terms of linguistic variables in fuzzy logic. Distinct boundaries "a la Lewis" between fuzzy possible worlds are initially impossible. Fuzzy worlds just have to intersect, that is, to create "diffusive zones" between them.

If "diffusive zones" appear, then a profound – ontological – reason to trust Davidson principle will appear as well. Of course, neither Davidson himself nor Quine could arrive at such reason since both of them didn't believe in ontological uncertainty but spoke only of the uncertainty of reference. However, both of them intuitively exceeded the limits of their own views: Davidson, by formulating the principle that deserves real trust only being backed by ontological uncertainty, and Quine, by letting feel the necessity of ontological substantiating Davidson's principle.

It's necessary to say that "a basic accord between Bohr and Davidson on fundamental semantic issues and on the type of common sense realism", in spite of "a wide chasm", that "separates their interpretative perspectives", was noticed at last not long ago (E. Mackinnon, Bohr and the Realism Debates // Niels Bohr and Contemporary Philosophy / Ed. by J. Faye and H. J. Folse (Dordrecht—Boston—London, 1994) (Boston Studies in the Philosophy of Science, 153) 279—302, esp. 300). Yet it would be better if Niels Bohr were compared not only with Davidson but also with Quine's philosophy from which Davidson himself started.

The fact of ignoring Bohr in the contemporary philosophy is justly explained, among other reasons, by his own fault, since he wouldn't master the language of philosophy of the day: but most of all, since he was too opposed to the principal currents in 20th century philosophy, including the analytical (see in detail: D. Favrholdt, Niels Bohr and Realism // Niels Bohr and Contemporary Philosophy, 77—96). Thus, in his polemics against Einstein – Podolsky – Rosen Bohr, in fact, faced also philosophical opponents in the person of Russel and Wittgenstein who equally insisted on fully objectivist comprehension of reality. Bohr also was incompatible with logical empiricism (Carnap's, for instance) founded on differentiating analytical and synthetic notions. When Quine refuted these notions in his "Two Dogmas of Empiricism" (see above) he didn't refer to Niels Bohr whose words he almost repeated.

Logical empiricism asserts that only analytical affirmations can be true since they're determined by properties of language and not of ontology. Russel's famous example is "All bachelors are unmarried". Synthetic affirmations concern ontology (for instance, "All humans are mortal") but they cannot be entirely reliable. Both Quine and Niels Bohr starting from completely different positions arrive at the same conclusion that no affirmation can be reliable at all, and yet no notion of affirmation can be independent from ontology. Therefore Kant's differentiation of synthetic and analytical reasoning makes no sense.

Regretfully, no one hitherto ventured to draw as much as a sketch of comparative historical analysis of development of philosophical thought of Quine and Niels Bohr; nor we can take this task upon ourselves.

We will just take liberty to formulate for the further needs one humble conclusion: our intuition telling that the difference between humans and goblins (and also giants and windmills, Homer's gods and physical objects etc) should be substantiated not only by terminological but also by ontological uncertainty, may prove true if we consider our world a totality of possible world of the FK type.

Ontology of narrative and intensional function of access

Above we spoke of "our world". Yet what is this world if not the world of certain narratives? After all any form of our cognition of the world take the form of a narrative: whether historical, fictional or that of natural science… Therefore we can't speak of

"ontology in general" or of "the world as such": if we do it, we do it on the expense of explicit or implicit assumptions. In practice we always speak of possible worlds of certain narratives. When, for instance, Dolezel opposes his worlds of fictional narratives to the "actual" physical world he actually means just the possible world of positivist tradition in natural science of 19-20th centuries. This is the world of extreme ontological flatness, the world of Bertrand Russell's logic.

The physical world we speak about is the world of Niels Bohr described in accordance with the logical principles of Quine, Putnam and Davidson. It's as ontologically multipronged as a world of fictional narrative according to Dolezel.

Dolezel notices (р. 152; following the monograph: M.-L. Ryan, Possible Worlds, Artificial Intelligence, and Narrative Theory (Bloomington, IN, 1991), written, in its turn, under Dolezel's influence) how ontologically multipronged is the fictional narrative where different characters may be related to different private possible worlds: mutually incompatible but imparting to the whole narrative ontological profundity and aesthetical attractiveness. However, this quality of being multipronged isn't peculiar to the fictional narrative alone.

Coming back to the example of Servantes, let's recall what we've said of an author as just a bearer of one of the possible opinions, be it even very authoritative, regarding the comparative reality of windmills and giants (see above). All these possible worlds, whether of Servantes, Sancho Panza or Don Quixote, interact with possible worlds of readers: and not only such readers as Servants could foresee or those who appealed to, but any possible readers. Thus the narrative of a fictional work appears to be a part of more complicate narrative structures existing for every reader.

Of course, for many readers a narrative structure describing the real world is the same. The very possibility of creating Servantes' novel proceeded from the existence of such structure: Servantes appealed to the audience of readers for whom, just as for himself, giants and even wandering knights didn't possess a status of real objects. However, in the days of Servantes the very fact of existence of such audience was still fresh enough to be reckoned as non-trivial (and consequently giving a reason for creating a fictional work): the old audience of readers of chivalry novels still existed hence self-identification of the new audience through opposition to the old one was still necessary.  

Thus, a possible world of a narrative, being a complicate system of several (at least) FK-worlds, is perceived by a reader (and also by an author) inside their own FK-worlds.

This perception is carried out so that for each separate FK-world a specific "level of accessibility" from a reader's possible world is determined. We call here a "level of accessibility" what in the previous considerations we've called a "diffusive zone" between different FK-worlds.

The very expression "accessibility" regarding possible worlds belongs to M.-L. Ryan (see esp. Ryan, Possible Worlds… 32—33 and M.-L. Ryan, Possible worlds and accessibility relations: a semantic typology of fiction // Poetics Today 12 (1991) 553—576), who revised Kripke's idea (see: Kripke, Semantical considerations of modal logic). Ryan took from Kripke an interpretation of probability as a ratio of accessibility from within the actual world of a certain logical object found in a possible world; whereas from David Lewis she took an idea of possibility of "decentralization" evolving due to the principally equal rights of all possible worlds: meaning that none of them can claim the status of central in regard to the whole system.   Without accepting Lewis' concept on a whole, Ryan considered it productive for narratology where it's not always expedient to count off a reader's "actual world". Therefore Ryan defined an accessibility of a given possible world as its accessibility from the world located in the center of a given system of worlds, meaning that the center may be chosen in different ways. Taking into account (which Ryan doesn't) that there can be no "actual world" as something unique, there's no sense to speak of accessibility of possible worlds apart from the concept of decentralization.

The possible world of Servantes coincides with the possible world of the audience he aimed at, and its direct extension is the possible world of Sancho Panza. Here a density of "diffusive zone" is maximal, as well as a level of accessibility. At the same time, in regard to the possible world of Don Quixote a tough barrier is erected: this possible world is claimed a result of mental disease. However, even this barrier isn't absolute. The most important effect of the whole Servantes' novel is based precisely on getting over it: that is, on the fact that it's namely in the possible world of Don Quixote, which is in fact a result of ailing change of consciousness, that things far more important to our reality than the ability of telling between giants and windmills, become apparent. Thus the novel exploits one of the topics traditional in the world literature: the topic of seeming insanity concealing a profound wisdom. This reminds us once more of the fact that "etymologically" "Don Quixote" originated, through chivalry novels, namely from hagiography.

Considered examples show that a possible world of each narrative can be characterized as a whole through the characteristic of ties (levels of accessibility or diffusive zones) between its private FK-worlds and FK-worlds of an author and his audience, whether planned or unplanned.

This general characteristic of narrative possible world is defined by intensional function (see above) that Dolezel calls "authentication"; he dedicated to it a whole chapter in his monograph (p. 145—168). The above cited example with the dialogue between Don Quixote and Sancho Panza is taken exactly from this chapter. 

Through all the preceeding pages we've been explaining in detail the difference between Dolezel's and our comprehension of this dialogue. It should have make clear why we weren't ready to accept the name "authentication" for the respective intensional function: in fact, this function as such doesn't define the measure of reality or fictivity of a given private FK-world, but only the degree of its accessibility from an author's FK-world that an author would like to share with his audience.

After Quine with his "myths" of both Homer's gods and physical objects, the question of possibility to define within a narrative any absolute affirmations regarding reality or fictivity of any possible objects should be finally removed. A narrative tells us only of compatibility of certain objects with the "fabric" of our knowledge as a whole: or speaking more precisely, not so much our knowledge as an author's knowledge that he strives to share with us.

The last affirmation is quite obviously applicable to the problem of interpretation of historic and hagiographic sources. We'll discuss this in detail along with considering the respective intensional function.

And in the meantime it's important to give a more adequate, from our point of view, name to this function. The name we suggest is an "accessibility function", that is, a function determining the level of accessibility between different FK-worlds.

With this theoretical characteristic our introduction into the ontology of narrative may be at last concluded. Now we may proceed to describing both the real action of intensional functions and the internal structure of possible worlds where it takes place.

Intensional functions in action

We already spoke of intensional functions but only by the way of describing them in general. Now we are going to describe them in detail.

Dolezel describes in detail only two of them: the abovementioned "authentication" and "saturation". At that he reserves that "…we don't know how many sorts of intensional functions exist" (p. 143); however, there are reasons to suppose that the two that he emphasized may be considered basic. We'll take a liberty to develop Dolezel's approach a bit by introducing certain specifications: first of all, those concerned with extending the notion of "intensional function" to all three basic types of narratives.

Here is the list of basic intensional functions as we see it:
1) The function of accessibility ("authentication", according to Dolezel);
2) The function of homogenization (Dolezel describes as a special function only one of its manifestations, "saturation": however, its other possible action, on the contrary, is "consolidation"; ontology is able "to shrink" and not only "to swell", since our goal is making the "fabric" of our knowledge homogeneous and not more extensive);
3) The function of symbolization of descriptors (exceptionally important to hagiography, but not only to it).

In regard to any narrative intensional functions are like building technologies in regard to a building. The way of applying a certain technology depends on architectural project. Roughly speaking, the simpler the project, the less technological it is. In regard to narrative the role of architectural project is played by the logical scheme of its plot described by the means of modal logics.

Description of "building technologies" of narrative only in regard to the fictional narrative has been first associated with discussing general problems of reference (and this is Dolezel's unquestionable contribution). Nevertheless, others spoke of this problem in this or another way in regard to other types of narratives. Thus, in regard to historic narratives similar problems were discussed by Ankersmit, in regard to narratives of natural sciences, all those who described the activity of what Lakatos called research programs. Empirically determined rules of critical hagiography also refer namely to the sphere of action of intensional functions. Particularly great experience of observations of intensional functions has been accumulated, though without special logical analysis, in the theories of mass media, information war and propaganda, and also in political and electoral technologies. 

All the mentioned areas, in this way or another, lay in the field of narratology, and each of them can use its theoretical conclusions for its own ends.

Accessibility function

Function of accessibility of various FK-worlds of narrative is nothing else but criterion of "reality" of what narrative tells about. There's no point to speak of reality as such since it is, in regard to narrative, Ding an sich. We can only appreciate a degree of accessibility of reality of each separate possible world from our own reality, while differentiating between the opinion of an author of narrative from our own reality (see above; we have already explained in detail why we shouldn't take an author's position a priori even if a narrative is fictional).

Accessibility of each separate FK-world depends on density of diffusive zone between this world and the possible world of our notions of reality.

If we define for possible values of this density an interval between 0 and 1, its values will always be somewhere in this interval and mere values of 0 and 1 will be, correspondingly, their infimum and supremum, that is, their unattainable limits.

If density of diffusive zone with a given FK-world were zero, then we could learn nothing about it, since it wouldn't comply with our experience and our abilities of comprehension. And if density of diffusive zone reached 1, then our possible world would be just identical to possible world of narrative, which is, of course, impossible (we aren't going to discuss here superstitions claiming that possible worlds of scientific theories are in fact worlds of our "actual" world; those who still don't deem it obvious enough may get some explanations as additional examples follow).

A notion of density of diffusive zone between a given FK-world of narrative and the FK-world of our perception is actually a "logically correct" definition of the intuitive notion of "measure of reality" of a given possible world that we've introduced above. 

"Logically correct" definition differs from an intuitive one inasmuch as it's a definition of a mere notion of reality given through a notion of FK-world. Poceeding from this notion of reality makes possible speaking of its various degrees.

Something, however, becomes impossible: and namely, it becomes impossible to insist on impossibility of existence of Homer's gods since we agreed to insist on possibility of existence of an electron. Both Homer's gods and electrons are notions exceeding the limits of experience, not only daily experience but experience as such. It may be argued that electrotechnology was developed through the use of the notion of electron and not of Homer's gods. However, it would be necessary to recall what notion of electron was used in electrotechnology: it was an electron prior to 1913, a small round particle that was fit for electrotechnology but appeared to be completely unfit for atomic radiation. Proceeding from the contemporary notions we would say that electrotechnology was developing on the basis of the myth of electron but not of the "real" notion of an electron. However, somebody would further object, Homer's gods were unfit even for electrotechnology: they were of no use at all. And it's here that our opponent would prove wrong: if a certain concept is of no use for electrotechnology, it still doesn't mean it's of no use at all. Homer proves that his gods appeared to be very useful for history (they gave a wonderful account of both motives and a course of the Troyan war; this explanation became outdated nearly three thousand years later, which is not so bad for a scientific theory) and, for instance, for ethics. It's true that the ancient civilization left us without gods fit for development of natural sciences, not even for such "celestial" science as astronomy. However, the matter is just that gods also may be chosen in different ways: while Greek gods were useless for astronomy, Babylonian gods were quite fit for this purpose. Everyone's got astronomy in accordance with his gods.  

Therefore if somebody by no means wants to believe in Homer's gods because of their lack of correlation with natural sciences, he might be offered Babylonian gods that were at least tested by astronomy: and by the best one in the world for millennia. These are the very gods that nowadays are trying to transform into the superstrings theory, black holes, mole burrows, big bang and other, no less remarkable things.

However, we started walking in circles, repeating what we've already said above and eventually, repeating Quine. Instead, we must start thinking of other things, and namely: why such easily interchangeable theories as Babylonian gods and superstrings theory, humans and goblins, windmills and giants, can act in certain narratives without hindrance, whereas in other narratives objects of reference cannot be so easily interchanged. For instance, such affirmations as "there are two chairs in the room" and "there are three chairs in the room" can be claimed equal in the sense of Davidson's principle of rational charity only by the way of much more complicate (and therefore deserving far less charity) tricks.

Accessibility function and truth values

The difference between the two groups of affirmations mentioned above consists in the fact that in the first case affirmations are much more theoretical, whereas in the second case they are much more empirical. Theoretical affirmations refer to FK-worlds that are most remote from us (remoteness refers here to accessibility level, that is, the least density of diffusive zone between the respective FK-world and our own), whereas empirical affirmations refer to the least remote ones. 

Empirical affirmations claim most of all that they determine the state of things in our immediate vicinity. They are closest to the frame of a loom upon which Quine stretches the fabric of our knowledge.

Dealing with narrative we must know to distinguish its more theoretical parts from the more empirical (in general, we must take into account that narrative may be heterogeneous and consist of different parts: thus, a chronicle may contain both chronological index of facts and historiosophical reasoning of an annalist). Sometimes it's obvious, sometimes not. And sometimes it can be obvious only if we know to discern relevant signs in accordance to special rules. Let's consider this in detail.

Let's take for a start most evident case: various narratives concerned with electrotechnology.

Basic formulas of physical laws referring to electrotechnology were developed as early as in 19th century: in spite of the fact that 19th century notions of electricity carriers were quite comparable, from the modern point of view, with Babylonian notions of gods moving the celestial bodies. Comprehension of physical nature of laws expressed by 19th century formulas has changed in most radical way, yet there was no need to change neither Ohm's law nor Maxwell's equation. The most theoretical of theoretical levels, the level of metatheory, suffered changed with ease without causing any troubles to users of electric appliances.

However, let's imagine another electrotechnical narrative: a manual to a given electric iron. Introducing substantial changes into such narrative without changing the iron itself is fraught with imminent consequences.

This example forces us to make the following observation. The more "theoretical" is a given narrative, that is, the less is a density of diffusion between FK-world of a narrative and the possible world of our consciousness, that is, the lesser is our accessibility level – the lesser are our possibilities of applying truth conditions to this narrative. Its possible world is so diffusive to us that we lose criteria of truth and falsity in regard to it.

That's what Quine meant, as we think, when he equalized, from epistemological point of view, physical objects and Homer's gods. It's not for nothing that Quine, in this case, didn't refer to Frege who simply deemed it impossible to apply the notion of "truth value" to imagined objects. From Quine's point of view, any theoretical object is as much imagined as it is theoretical: taking into account that any object, to a certain extent, is theoretical. Therefore the sense of a notion of "truth value" in regard to affirmations about objects grows inasmuch as empiricity of these objects grows.

The higher is a degree of diffusion of a given narrative world with our own, the more the sense of a notion of "truth value" in regard to affirmations about this world grows in our eyes. It's nothing else than just another formula of a well-known requirement of a common sense: we can judge the verity of certain phenomena only inasmuch as we have an access to it.

We have no access to the world of Homer's gods and heroes, as well as to the world of 19th century electrons and Niels Bohr's electrons, or to the world of giants and even windmills of "Don Quixote", therefore we are competent enough only in the way of perceiving them as various modes of theoretical comprehension of reality. We have our reasons to consider them more or less apt to our certain needs, and after all, our every specific reasoning involves a number of simplifications and is always approximate: but no more than that. In order to assert absolute truths like existence or non-existence of Babylonian or Greek gods a religious revelation is needed, since ordinary human abilities never suffice in such matters.

Let's consider now a possible world that is close to us (in a sense of density of diffusive zone) and whose structure is described with involvement of false affirmations. For instance, that may be a narrative world of a manual for electric iron containing wrong directions in regard to the choice of voltage. It's easy, though maybe a bit too much expensive to get convinced that this manual is fallacious.

The FK-world of the Impossible Iron corresponding to this manual may be interpreted in two ways: either as a part of our world whose description features a wrong affirmation, or as another world completely isolated from our own, where an affirmation in regard to applying a certain voltage to an iron cannot be verified in our world in any way.  However, even in the last case the words "completely isolated" bear no absolute meaning. The world of the Impossible Iron contains irons that would be similar to the irons of our world but for one detail; therefore the degree of diffusion between this possible world and our own varies from zero notably.

Another version of the World of the Impossible Iron is, for instance, a historic narrative containing (as we know perfectly from more reliable sources) a factual error. This example may be useful for those who still can't see the direct link from the Impossible Irons to history and hagiography. 

The example shows that interpreting a factual mistake as mistake depends, most of all, on the possibility of identifying the possible world of a given narrative with a possible world where we place our empiric data. This, in its turn, depends on the degree of empiricity or, on the contrary, theoreticity that we attribute to a given narrative.  

If we perceive the manual for an iron as a manual for an iron that we can take in our hands and try to use, then we interpret the manual as a predominantly empiric narrative referring not so much to its own as to our own world. However, we can perceive this instruction poetically and symbolically without sinking to its utilitarian interpretations. It's possible to give examples of instructions that were initially compiled as quite poetical texts:

Нечаянный сон — причина пожаров.
Не читайте на ночь Уткина и Жарова.
[Occasional sleep may be cause of a fire,
Don't read Utkin and Zharov at night of retire]
(V. Mayakovsky)

The topic of using proletarian poets Utkin and Zharov as a stuff for reading has been long forgotten together with poets themselves, not to mention that the danger of fire in consequence of an occasional sleep is also long gone (probably, together with oil lamps), as anyone who has a habit of sitting days and nights before the computer should know well.   Thus, extensional content of the given poetical narrative lost its actuality long ago. However, the narrative itself retained its actuality, which is supposedly unquestionable, and it's concerned with the narrative's intensional content (having emerged mainly due to the function of symbolization of designators "Utkin" and "Zharov"). Therefore we can say that we face a poetical text and just an "instruction" (one poet's hostile gesture towards two others).

Therefore visible relation of a possible world to our possible world is, in general, a question most important for interpretation and not obvious at all. This is actually the principal question concerned with an accessibility function.

Theoretical as empirical and general as particular

Thus, inasmuch as narrative shifts away from the empiric "edge" of the fabric of our knowledge Quine speaks about, verity of statements constituting the narrative becomes increasingly hard to check. A notion of truth value doesn't lose all sense, but it becomes increasingly hard to use practically. Roughly speaking, it's possible to divide the whole content of a narrative into two components, according to the character of knowledge they contain: circumferential (closely associated with empirical experience) and abyssal (not associated with experience directly but offering explanations of this experience). Above we spoke of these components of narrative's content as "empirical" and "theoretical". This division into components isn't inherent yet in the narrative of myth, whereas secondary sacralization inherent in hagiography (see above) manages to overcome it. However, all the "ordinary" narratives - fictional, historic, of natural science - emerge precisely on the basis of this division of the fabric of our knowledge into circumferential and abyssal areas. As for hagiography (and apocalyptics), it wouldn't appear as well if there was no such division to overcome.

Now we have to consider how both abyssal and circumferential content of narratives of various types is concerned with accessibility functions.

The simplest case is myth where differentiation of circumferential and abyssal content makes no sense, and therefore functions of accessibility of the possible world of respective narrative are distributed uniformly:

A respective feature of mythological thought was first adequately described by E. Cassirer:

"For the modern science the unity it aspires to is the unity of the law of nature, a law purely functional; for astrology [whose respective feature Cassirer relates to the most ancient features of mythological thought – B.L.] it is the unity of transparent content that remains unchanged, unity of structure of the universal whole. It likens the world to a crystal: even if we broke it into smallest particles, they would still retain the same form of organization. On this various classifications of primitive tribes are based, where tribesmen, plants, animals, natural phenomena etc. are submitted to the same system. (The Philosophy of Symbolic Forms. Volume Two: Mythical Thought. New Haven: Yale University Press, 1955).

 In mythological consciousness everything is equally associated with experience and everything is equally abyssal. There is an unlimited access to the whole reality of mythological narrative, since possible worlds of perceiving consciousness and possible worlds of mythological narrative can't differ by definition.

In the rest of narratives the difference of abyssal and circumferential content becomes apparent (it is designated on our schemes with horizontal dotted line).

Thus, circumferential content of the fictional narrative, that is, the world of fictive characters, may be ultimately inaccessible. Dolezel wouldn't speak of inaccessibility but of fictivity of this world. However, the more general features of this fictive world we consider the more similarity it reveals with the possible world we are in. This similarity may extend not only to characters' personalities but also, say, to the topography of St. Petersburg: always recognizable in Dostoevsky's books. Diffusive zone between our FK-worlds gradually becomes denser inasmuch as it recedes from individual features of the world of a fictional work. Very conditionally (and having renounced representing this transition gradually as it is) we may represent the above said by the means of this scheme:

Thus, in order to raise the question of reliability of biography of Myshkin from "Idiot" we'll have to move into the possible world of a narrative. This is possible since both in that world and in our world there exist, for instance, similar laws of spiritual advancement: however, the Prince Myshkin himself exists only in that world. However, there's no need to sink into the possible world of the narrative in order, say, to find Rogozhin's house: it's would suffice to open the novel at the correct page and then to walk the Sadovaya and the Gorokhovaya streets. However, in this house, momentarily recognizable by its description in the novel, neither Rogozhin nor even any trace of Nastasya Philipovna's murder will be found, since Rogozhin itself exists only in the world of the novel. The laws of spiritual advancement that Myshkin suffered from belong to the diffusive zone between our two FK-worlds. Rogozhin's house belongs to the same zone. Of "Raskolnikov's house" even more can be said, since its mere name exceeded the novel and is now placed on the bronze memorial plate on its wall; the diffusion between the world of "Crime and Punishment" and the world of its readers proceeds to deepen with every decade.

Let's emphasize once again the difference between our approach to reality of possible world of fictional work and Dolezel's approach by recalling the already cited example of Napoleon: Napoleon of history and Napoleon of various fictional narratives. Traditional literary criticism (not using semantics of possible worlds) considered Napoleon a historic figure, whereas Pierre Bezukhov was referred to as a fictive one. Dolezel justly objected that such violation of the principle of ontological homogeneity of the world of a fictional work doesn't allow an adequate interpretation of its ontology. However, Dolezel's own solution consisted in declaring both Pierre Bezukhov and Napoleon of "War and Peace" equally fictive. This solution was logically consistent but rather obviously contradicting both regular intuition of literary criticism and Tolstoy's explicit task: creating in his novel alternative historiosophical scheme. As we've noted above, Tolstoy's Napoleon may be considered fictive only if Tarle's Napoleon and all the work of historians in general is to be considered fictive as well. Dolezel's approach doesn't allow differentiating between fictional and historical narratives, therefore declaring the possible world of one of them purely fictive must automatically extend to the possible world of the other.

Our approach based on the concept of possible worlds as FK-worlds allows counting individual objects in several possible worlds at once (if they are located in diffusive zones between these worlds). Therefore the contradiction caused by the problem of affiliating the objects designated by the rigid designator "Napoleon" with either the world of real history or possible worlds of various fictional works is resolved: for all the difference of respective objects in different possible worlds, these objects always have a certain area of intersection. And this area isn't just "designator", that is, mere "designation", but a certain reality. However, it's necessary to remember that no historical conception ("narrative substance", using Ankersmit's term) of "Napoleon" can be fully identified with the historic person of the same name.

Contrary to fictional narrative where circumferential content may be almost inaccessible is any sort of bare description of facts or items: for instance, inventory or report. Such narrative presumes no room for generalization, theories etc. Of course, that doesn't mean that it's completely shorn of abyssal dimension, since even when just naming objects we operate with common notions and classifications: but by contrast with fictional narrative our unquestionable right is to consider (conditionally) the abyssal part of such narrative empty:

Fictional narrative and narrative as inventory or report are two opposite extremities.  Criteria of verity actual in our world in regard to this narrative are possible to apply either only to abyssal or only to circumferential content. That actually means that accessibility functions admit us to only one of the two areas of any of these narratives.

Ordinary narratives claiming neither artistic fiction nor inventory of objects or events distribute their content between circumferential and abyssal levels in different ways. Such narratives correspond to accessibility functions that lead them to empirical (superficial) content as well as to theoretical (profound):

It's necessary to emphasize that different accessibility functions lead to different parts of a narrative (that's why arrows on schemes don't continue but are limited by their area of a narrative). The reason is that in all the forms of post-mythological consciousness, even in natural sciences, there's no direct and unambiguous correlation between empirical sphere and theory (we spoke about it many times, see above).

In mythological thought such correlation existed: theory unquestionably defined not only interpretation of experience but experience itself as well. Apart from drawbacks obvious to modern humans there were evident advantages about this state of things. 

Thus, once ethnographers were amazed to learn that members of one of the tribes of South American Indians quite seriously identify themselves with green parrots, their totemic animals. Actually ethnographers didn't realize too soon that totemшс beliefs presume namely self-identifying with totemic animal and not just some symbolic correlation. Adequacy of reality of such state of mind can scarcely be considered absolute, yet it can be essentially higher than adequacy of reality perception by less primitive humans. The last striking example is 26 December 2004 tsunami: closest to the epicenter were Andaman and Nicobar Islands populated by Paleolithic tribes. These tribes escaped the disaster together with their fauna since they felt the danger together with animals. In Sri Lanka which was far more distant from the epicenter of tsunami, only animals were saved that, contrary to humans, managed to escape in a due time. Many of the deceased Europeans would probably prefer, if there were a choice, to identify themselves with some green parrots, acquiring thus the ability of reacting to danger adequately and realistically. Therefore even totemic beliefs quite deserve to be regarded in accordance with Davidson's principle of rational charity.

Let's observe now what happens to accessibility functions in consequence of secondary sacralization inherent in hagiography (see above for the notion of secondary sacralization).

A possible world of hagiographic narrative doesn't transform into mythological world therefore its respective knowledge is still divided into abyssal and circumferential areas. Hagiographic narrative may be likened to mythological and fictional since it's supposed to deliver, first of all, a certain profound knowledge of its reality that is identified with our own. Therefore it's always presumed that accessibility functions lead to an abyssal area of content of hagiographic narrative.  
As for circumferential part of content, outwardly it seems to resemble historical or scientific narrative, whereas structurally, by the character of accessibility functions it resembles myth: but in fact it differs from both of them. Here are schemes for, respectively, historic (on the left) and epic (on the right) hagiography: 

Similarity to myth consists in the absence of separate accessibility functions for circumferential and abyssal spheres. This means that unambiguous link is restored between empiric facts and their profound content, that is, between causes and sense. This is an essential feature of any sacral knowledge.

However, accessibility function in circumferential area may have, so to say, an apparent part (designated by dotted part of arrows): an access may not be granted but only simulated, as it happens in fictional works given by their authors a form of someone's diaries of chronicles.

For instance, it so happens in epic martyrologies built only on hagiographical coordinates, that is, proceeding only from the martyrium and the date of the feast. All the remaining (out of the limits of hagiographical coordinates) extensional content of such works has only semblance of empirical description of the private case of a specific martyr, whereas in fact it serves the theoretical description of the whole: of the fact that the given martyrium is a centre of the mattyr's cult.  The general is expressed her through the particular (pars pro toto), the theoretical is expressed through the empirical: in accordance with the standards of mythological thought and in accordance with literary canons of such specific branch of the genre of epos as Passions epiques.

For all that, an apparent part of an accessibility function differs essentially from a circumferential part of accessibility functions in mythical narrative: that's why we ventured to call it a "simulation" of accessibility. Regardless of how Lives would be perceived by various categories of their readers, ecclesial culture on the whole always presumed critical attitude to their extensional content. On the level of intellectuals it might be conscious criticism (to the point of putting certain hagiographical works in the index of forbidden books), whereas on the level more popular it might be fairly free modification of extensional content leading as far as to modifications of biographies of saints and even to forming collective images of saints out of several prototypes taken from the older hagiography. Secondary sacrality of hagiographica narrative never became immovable (or more precisely, hard-to-move) sacrality of myth. Therefore real functioning of hagiographic narratives always had much in common with fate of ordinary historiographical and sometimes also fictional works.

Most similar to ordinary (secular) historiography is, of course, the hagiography of historical type (according to Delehaye). Historians are so often deceived by this similarity so that they forget about the difference, but it seems to be unclear even to them, that self-identification of a monument as hagiographical work warns them about. This difference, nonetheless, consists of the fact that a hagiographical narrative, in contrary to historical, doesn't offer independent functions of access to their empirical content (on the schemes, similarity between accessibility functions for abyssal and circumferential content of narrative is designated by continuous arrows on both sides of the horizontal dotted line). While speaking so we, of course, make it rough, since there are quite ideologized historic narratives that resemble hagiography and even creative myth, as well as hagiographical monuments whose compilers more or less digressed from their direct task and were carried away by historic pictures as such. However, the general tendency is just as we have formulated it: the empirical (circumferential) in hagiographical narrative (ideally, as ever) is unambiguously intercorrelated with the theoretical (abyssal). This unambiguity, in its turn, needs what we have called a constructing reference that is gained at the expense of action of various intensional functions. Therefore if historic narrative (sometimes in spite of historiographer) strives to be openly interpreted, hagiography ever strives to take the constructing reference to the maximum.

The more explicitly the tendency to constructing reference is expressed in hagiographic narrative - that is, the more it resembles an epical narrative - the more it resembles an epical or a fictional work in regard to how truth values of statements it contains should be defined. 

The naive historiographical criticism of hagiographical monuments (in Russia best known by the monograph by V. O. Klyuchevsky (1841—1911) "Ancient Russian Lives of Saints as historical source", 1871) is based on the analysis of separate episodes and facts in regard to their "plausibility" or "implausibility" in the context of historian's general notions of historic reality but without taking account of the specific place of these facts in the very structure of Life. Delehaye has shown that the last parameter is most important, since namely it allows understanding what rules determine in each case the search of truth values: whether they comply with our world only in the area of abyssal (theoretical) content or in the empirical area as well. Delehaye's conclusion that the last parameter (historical reliability on various levels) correlates with the genre nature of Lives is confirmed by Ryan (see above) who points out that the perception of text, be it fictive or plausible, is mainly concerned with our initial expectations preceding the acquaintance with a text: and these expectations derive from our notions of the genre nature of a given narrative.  

The truth values that may interest historian must belong to diffusive zone between FK-world of historian and FK-world of hagiographical monument. Diffusive zone is only where accessibility functions act: them but not their simulations.

In order to understand where exactly accessibility function act, one must understand how other intensional functions act at their side. Before proceeding to considering these functions, a short summary is necessary regarding accessibility functions.

Accessibility function: basic summary of discussion

In spite of the fact that the very notion of accessibility function was set as early as in Kripke's 1963 classical paper, basic for all the contemporary semantics of possible worlds, and in regard to narratology it was elaborated by Ryan in 1991 publications (see above), we had to revise it rather essentially and to suggest our own version.

However, it couldn't be otherwise: inasmuch as we initially couldn't accept the assumption most important both for Kripke and for Ryan (and for Dolezel as well) that so called "actual world" is an only (ideally defined) world known to us. Elena Semino (see above) wrote of the problems concerned with this assumption as early as in 1997; however, she didn't exceed its limits in the logical schemes she proposed, but only corrected it within the margins of her cognitivist approach (which allows analyzing psychological aspects of creating and perceiving narrative as such but not its internal logical structure). Semino's psychological and psycholinguistical approach expected to be complemented by logic and ontology.

We have tried to build up such complement on the basis of Quine's principle of ontological relativity and Davidson's principle of rational charity (having imparted to the last ontological basis which it missed in its original version).

We initially proceeded from the suggestion that we don't know what the "actual world" we live in is like. That's why in particular we can't compare the knowledge of this world claimed by apocalyptics and hagiography, with any absolute truth. We can compare them only with information provided by other narratives (those of natural sciences, for instance) that, strictly speaking, are no less mythological, though after their own fashion. We emphasized the last circumstance, following Quine, not for the purpose of producing our own authorized noise amid the Luddite crowd of "deconstructors" of science and philosophy, but, on the contrary, in order to affirm a sort of profound equal honesty of all kinds of human knowledge of all times.  

As a result, we suggested considering reality as a totality of fuzzy worlds of various volume and density of areas of intersection.

Inasmuch as we have no preliminary knowledge of qualities of our "actual" world, we can't follow Ryan in defining the degree of accessibility of possible worlds of various narratives in accordance with basic features differentiating them from "our" world (such as presence in narrative worlds of the same objects that can be found in the "actual" world", the same logic, the same physics laws etc). All the worlds we can compare with each other are narrative worlds. Ryan herself noticed that the criteria she suggested for measuring possible distinctions between fictive and actual world aren't absolute. We can agree with her that they nonetheless may be useful for the analysis of literary fiction in the modern sense of word, but seeing no way they might be useful for the analysis of hagiography and historical literature we renounced discussing it.

Nevertheless we continued, along with Ryan, interpreting the question of accessibility of a possible world of narrative as a question of its reliability (or, speaking within the limits of the analysis of literary fiction, "realisticity"). Therefore the question of accessibility, as we saw it, concerned also the question of correlation of narrative's content with reality, that is, the question of practical applicability of a notion of truth value to statements of narrative.

Following Delehaye and Ryan we noticed the correlation of these rules with genre nature of narratives in general and hagiographical narratives in particular. In the terms we suggested, the most important question of appreciating hagiography consists in possibility of differentiating between accessibility function as such and "simulation of accessibility".

"Simulation of accessibility" is a term that may be actually applied not only to hagiography, but not to any other narrative as well. Thus, it's useless to apply it to fictional narrative. Even in regard to the most realistic and historical novel we are initially warned (by the mere genre it belongs to) that we deal with imagined objects and facts and, consequently, with an unconcealed absence of accessibility and not with its simulation. There's no point to speak of simulation of accessibility also in regard to the abyssal content of any narrative: it's always accessible to some extent (or, as in the case of narrative-inventory of report it may be reckoned, speaking roughly, that it's just not there).

However, in regard to hagiography a notion of simulation of access to a possible world of narrative is endowed with essential sense (it's also may be important for historic narratives when they reveal features of structural similarity to hagiography). Hagiographic narration is always conducted (or it seems that it's conducted) in a way as if full historic plausibility of each fact were being affirmed. In fact it's not so at all and hagiographical narration lets it be clear by the means of intensional function of accessibility. This function in different ways points at different objects of hagiographic narrative, and one must know to distinguish it from its simulation. Of course, simple rules like those formulated by Dolezel and Ryan for a fictional work (where, for instance, author's speech, apart from postmodernist exceptions, is a voice of an absolute truth) don't suffice here. Instead, we have to consider accessibility functions along with other intensional functions, chief among them are functions of homogenization that we are going to consider now.

Functions of homogenization and conceptual blending

Literature: ,G. Fauconnier, M. Turner, The Way We Think. Conceptual Blending and the Mind’s Hidden Complexities (N. Y., 2002).

Before plunging into the description of fascinating intensional functions, it would be necessary to recall thoroughly what was said above of intertextuality, its referential mechanism and its significance for any traditional literature to which hagiography belongs as well.

The matter is that the functions of homogenization are such that they are switched on when narrative is being read but afterwards the result of their action passes into next narrative created by a reader (or an audience – if we speak of verbal tradition) of the preceding one. It is impossible to understand any text referring to (any) Tradition, without taking into account that it's a product of reading (listening to) the preceding texts referring to the same Tradition. What couldn't be said explicitly (or even not meant) at the previous stage can be explicated (or "explicated" in quotation marks) at the next stage, presuming at that a certain interpretation of the Tradition, often quite disputable and even unquestionably debatable, but nonetheless far from its arbitrary distortion.

A particular but extremely important case of such transition of tradition is passing from narrative containing only description of empirical reality (historical or natural phenomenon or experiment) to narrative containing its theoretical comprehension (which is, in particular, always true for hagiographic narrative).

Functions of homogenization, as well as all intensional functions, are functions from texture to extensional securing the process that we called a constructing reference. By the means of functions of homogenization we comprehend the sense implicated in texture. The result of such comprehending may be put into shape in the texture of next narrative where it will already appear more explicit. Texture of any narrative built on the basis of intertextuality will assume the presence of multi-layered explications of narratives referring to previous stages of the Tradition and (what is important for both history and hagiography) to description (that probably never existed in a written form) of empirical phenomena.

Functions of homogenization were described, though from different point of view and proceeding from other ontological premises, by F. Ankersmit in regard to historiographical narrative (see above). He notices that "…the existing Nss have a tendency to coagulate into groups, whereas a "narrative space" between groups remains more or less empty" (Ankersmit, Narrative logic… (see above)).

The function of saturation draws more attention than the rest of the functions of homogenization (for instance, Dolezel deals with it alone), and justly: remembering the tendency of development of human knowledge described by Quine (see above): "swelling of ontology in order to simplify theory".

Obviously contradicting the "Occam's razor", human mind easily multiplies essences in order to simplify theory. This is the way we became aware of electrons, quarks, superstrings, Homer's gods and goblins. In the same way an investigator proceeding from certain cues can guess that one more, "unplanned" man was present where the crime took place. Therefore a contradiction to Occam here is more likely apparent: after all, what Occam prohibits is multiplying essences beyond necessity, and a notion of necessity (whatever Occam might think about it) is a thing very subjective, and it's precisely what Quine's rule regarding the "swelling of ontology" describes.

However, Occam's razor often works in the way of its classic application: but then it cuts off everything, worrying only for simplifying theory. We've suggested calling this type of functions of homogenization a function of consolidation. Differentiating between the functions of saturation and consolidation is convenient only for practical goals, since theoretically these are functions of the same nature and their mechanism works in a similar way while, however, bringing opposite results.
This mechanism isn't studied within the limits of narratology, but within the limits of psychology of cognitive processes. Fauconnier (known also for his works concerned with cognitive linguistics) and Turner in their recent extensive monograph named the psychological processes responsible for activity of our intensional functions a "conceptual blending", and the images evoked in our mind by these processes, "blends".

The processes of conceptual blending continuously maintain the integrity of the fabric of our knowledge and prevent it from tearing up. In regard to perception and transduction of narratives, intensional functions of homogenization are no more than a particular case of such "blending". Therefore let's hold that our choice of a term "homogenization" corresponds to the "conceptual blending" of Fauconnier and Turner, since blending is actually homogenization.

We are not going to consider psychological theory in detail, yet we will outline it in several words and illustrate it by an example indisputably remote from hagiography by its genre, but appearing to be quite close to it structurally.

The essence of a mechanism of conceptual blending is illustrated by the following scheme of an elementary process:

http://cogsci.ucsd.edu/~faucon/253/image.jpg

Initially there are two so-called mental spaces: Input I1 и Input I2. There is a correlation (of any sort) between some of their elements; on the scheme it's designated by continuous lines. It becomes possible to generate the third mental space: a Generic Space. This space will involve only those elements from the two initial spaces that were correlated between them and are identical. This space is a space of intersection of both initial spaces.

The generic space will define the internal structure of the fourth mental space: the blend itself.

Certain elements from the initial spaces will get into the blend bypassing the generic space. Other elements won't get into the blend at all. However, those that will get into the blend under the influence of generic space will be transformed in different ways. For instance, as the scheme shows, initially different objects may coincide, whereas initially identical objects may become different.

Let's consider one example from Fauconnier and Turner's book. This is an promotional poster calling no to spare money for kids education.  The inscription in large print says: "Joey, Katie and Todd will be performing your bypass". One feels an ominous pun here, since "to bypass" may also mean "to neglect". Already from the title the subconsciousness guesses that "your negligence" may be fraught with a bypass of corresponding quality.

As for the picture, it's scary beyond doubt: there is an ordinary surgery room, a patient and surgeons in overalls with surgery tools: but the surgeons are about seven years old: the very age for children to start their education… However, it looks like these ones did not start it properly at all.

The poster is by the American organization called 'Education Excellence Partnership' (first posted, as far as we managed to find out, in the New York Times at 7 November 1996).

In small print under the picture we read: "Before you know it, these kids will be doctors, nurses and medical technicians, possibly yours. They’ll need an excellent grasp of laser technology, advanced computing and molecular genetics. Unfortunately, very few American children are being prepared to master such sophisticated subjects.  If we want children who can handle tomorrow’s good jobs, more kids need to take more challenging academic courses.  To find out how you can help the effort to raise standards in America’s schools, please call 1-800-96-PROMISE.  If we make changes now, we can prevent a lot of pain later on". 

The effect of the poster is determined by a blending process.

In initial mental spaces there were, on one hand, the present situation where children don't get a proper education, and on the other hand, the future situation where a person the promotion is aimed at might need a bypass. Then this person will lie before these children on a surgical table.

In a final blend on the poster the future is actualized, that is, moved into the present. It seems that the children are still seven at the moment an operation starts and their surgery skills are those of children about seven.

A bewildered spectator looks at the poster more closely and then spots an inscription in even smaller print that eventually answers the “cursed” and ‘old’ question: "What is to be done?"

The core (though not an only one at all) peculiarity of a blending mechanism in this case is ignoring the temporal differences for the sake of actualizing of plot, that is, making a spectator perceive the future as the present. Such actualization is very characteristic for canonical texts of hagiography as well.

Full text of the article in Russian (975 KB)