Національна Академія Мистецтв України Інститут культурології

Perhaps it is textual antinomy that the declared similarity between text and proposition can come back to. The conjecture on it has been suggested by V.G. Admoni. There are three most important structural features common for text and sentence that include first of all the effect of tension: it reveals itself in the special role of the final phase that corresponds to expectations both in text and in propositional structure¹¹⁴. The next principle ensues from here and can be defined as the plot inherent both for text and sentence. In particular plot is here conceived as the extended predication¹¹⁵. At last, the third feature consists in the semantic multipart structure where the coexistence of different meanings with different lines of references takes place¹¹⁶. At the same time it is by no means proposition that can compress a text. As the descriptive result of such compression one would become rather a simile of a score than a propositional structure. Besides, these statements betray essential deficiency: they confuse proposition with inference as the logical backgrounds of (respectively) sentence and text. In this regard it is by far not every sentence that can possess its plot. It is only in hypotaxis with its expansion of predicative relation to the scope of taxis where the traits of plot can be found. Obviously one can transform a simple sentence into such expanded structure but only after such transformational operations it would be correct to say of plot.

These V.G. Admoni’s statements are to be confronted with the analogous enunciations of O.E. Mandelstam who compared text to a single word reflecting thus in the reverse form the famous simile by A.A. Potebnya (cited further)¹¹⁷. Thus the controversy arises, whether word or sentence (resp. notion or proposition) can be correlated with textual integrity to represent it in a descriptive compression. The answer should be negative to the both members of this alternative: text represents inference irreducible to other logical forms. Therefore it seems to agree with a much more reserved approach uttered by G.O. Vinokur who stressed the ubiquitous references that penetrates all lexical stuff of a text without further conclusions as to propositional or lexical analogies of such referential net; the conclusion would be the lexical combinations (of collocations and colligations) that build up the continuous verbal tissue¹¹⁸. Moreover it is the universal reflection where the sources of this organic entirety of poetical texts are to be found so that all the meanings become motivated through their reciprocal signifying relations and due to this motivation text gains integration¹¹⁹.

Textual integration radically contradicts to propositional or lexical integrity. In opposite to words and sentences where the entirety is predestinated with paradigm a text must still be built and synthesized. Lexical stuff is given beforehand in a vocabulary and sentences are determined with grammar whereas the unity of text doesn’t arise as something evident. This situation correlates to that in logic: there are propositional calculus with its truth vs. falsehood values and the calculus of classes with respective rules whereas inferences aren’t given immediately from premises remaining irreducible fully to the rules of logical quadrangle or figures of syllogism; otherwise they would turn to become tautologies. .
1.2.2. The Inferential Basis of Textual Integration as the Consequence of the Foundation’s Paradox
A text can be ultimately represented as a hypotactic construction where sentences become subordinate clauses. The coherence of a text can be then reconceived as an all embracing compound sentence. Separate sentences turning into subordinate clauses, the integrative schemes of text become transparent.

Such transformation becomes still more important for the communicative text representing dialogues of partners (conversations). Each communication presupposes the existence of some axis providing the mutuality necessary for the interchange of information. One can detect the filament of such axis as far as the utterances of partners create a kind of conclusion in the form of the so called “heap” or sorites (). It entails also the existence of an arbiter capable of representing such uniting textual structure together with communication’s partners. Therefore text as message represents deductive procedure of building inferences or entire syllogisms and coming to conclusions that are to be substantiated with textual stuff. The case of sorites is here especially demonstrative as it goes about interpretative abilities to restore the correct order of speech¹²⁰.

Thus a text is to be regarded as the representation of inference or conclusion (syllogism) and not as a mechanical accumulation of propositional sentences (judgments). Text demonstrates thus its irreducibility to a plain sum of sentences. In its turn the inferential nature of text as the integral entirety entails also the inapplicability of predicates’ calculus to its description as far as syllogisms can’t be reduced to the procedures of this calculus. It ensues from the relationship between subjects and predicates that turns to be much mote complicated than it is usually supposed in this calculus precluding thus any possibility of applying it¹²¹. In its turn predicate calculus would presuppose very ambiguous properties of implication so that correct inferences can’t be adopted within its system¹²². The difficulties are still more complicated when it goes about the representation of propositional structures within the terms of calculus¹²³. The same concerns propositional calculus that necessarily involves the upper level of generalization¹²⁴. In particular one can’t here account for the inferential role of predicate as far as it turns out to be inseparable from subject¹²⁵. Another aspect of inferential irreducibility to predicates’ calculus is connected with temporality that propositions can abstract from. In particular as far as it goes about pragmatic applicability of meditation within what is called “teleological syllogism” (G. Tarde) one has to reckon with motifs and initial collisions to be overcome in inference [Ишмуратов, 1987, 69]. One can add still that it lacks in propositional calculus 4 figures of the 19 acknowledged as correct in syllogistics [Гокиели, 1, 304].

The seeming “readiness” of syllogistic figures and the pretended “paradigm” of inferences show in reality essential difference in comparison to propositional frames. Inferences would coincide with tautologies and bring no new knowledge if such frames were valid here. The irreducibility of text as inference to a set of ready propositions is proved with the refutation of the opposite viewpoint that has been wittily defined by L.P. Gokieli as “the viewpoint of sum” («точкой зрения суммы») [Гокиели, 1, 49]. The organic system of inference would then be replaced with a mechanical aggregate. In its turn if all inferences were closed and built as ready paradigms of tautologies than the already mentioned paradox of foundation would appear that concerns the famous Goedel’s theorems on incompleteness and incompatibility of axiomatic systems. Each separate proposition needs it substantiation. Therefore it always implies an inferential conclusion standing behind as its demonstration. It is why proposition (or notion) as such can’t become the initial point of any cognitive procedure¹²⁶. .

Thus it is the deductive process represented within the text that remains constant. Vice versa its division into propositions turns out to become variable. Inference integrates text, and propositional division remains unimportant. This last statement is in particular connected with the paradoxes and controversies ensuing from the rejection of propositional forms by G. Hegel. Though the statements of a doctrine are always presented as a sum of propositions their contents can by no means be reduced to such sum¹²⁷. In particular it goes about the fact that propositions (sentences) “don’t constitute, incarnate or reveal … contents or truths … The propositions … must be comprehended as those signalizing rather than constituting, inducing rather than building up the contents” [Soll, 1975, 47]. No need to say that proposition can be transformed into an enumeration where predicate can occupy, say, an initial place and thus be marked as the central element of field structure.

This contradiction between inference and proposition is suggested by L.P. Gokieli to be solved with the concept of the already mentioned “radical conclusion” («коренной вывод») where there would not be place for the confrontation of premises and conclusions. Such concept promotes in removing reductionism proper to axiomatic approach so that negations’ examination becomes the initial point of inference¹²⁸. Thus the self-substantiating conclusions will be built where negation becomes principal inferential device. Respectively instead of reduction to absurdity or the demonstration from the contrary (as the means ensuing from implication and presuming the foundation paradox) one takes regression to infinity as the principal refuting device¹²⁹. Therefore “radical conclusion” can’t be identified with implication as it doesn’t presuppose the existence of separate premises¹³⁰. Respectively the negations and alternatives build up the initial point of inference precluding the implicational division of antecedent and consequent. For instance the Descartes’ principle will be then conceived as the refutation of self-negation¹³¹. This doctrine of radical conclusion comes back to Hegelian criticism of implication and of attempts to divide it into a set of propositions and the separation of premises as axioms¹³². The alternative to reductionist’s approach is here to be found in the connection between the foundation and reflection as its source¹³³. Besides, it is to stress that the paradox of foundation presupposes transcendent transitions as those to be conceived as limits¹³⁴. The necessity of such transition is caused as the counterpart of phenomenological reduction within the demands of approaching reality.

In its turn predication discloses its irreducibility to intersection in the same way as inference turns to be inexhaustible with implication. It is again negations that determine the formation of proposition. The statement “a cat goes” means not only that and animal doesn’t jump, run or lie as well as that it is not dog, ape fox etc. The negation concerns also the relation between the notions “to be a cat” and “to go”. It is contrast if semantic fields that is fixed in each collocation. Intersected classes do reciprocally deny each other. Here is to be taken into account also that all conjunctions (intersections) can be represented and replaced with the combination of implication (inclusion) and negation (in particular as the negation of implication with negated consequent). Respectively subject vs. predicate relation reproduces the asymmetry and anisotropy of that of antecedent vs. consequent. The unfitness of propositions to represent relations demonstrates it especially sensibly¹³⁵: for instance it remains indefinite what is to be subject for the relation “more” and what preferences are to be given to one of the terms; in the same way the attempts to identify relation with abstraction would fail (as in equality¹³⁶). Thus the boundaries between proposition and inference are transgressed so that propositions disclose their dependence upon the inferences¹³⁷.

This reasoning shows that from one side inferences are irreducible to propositions; from another side they are also irreducible to syllogisms being something much broader. That syllogisms don’t exhaust the inferences is proved also with the problem of mediate inference that don’t suit the rules of reduction. In particular “it leads us… to a relation between two concepts… by the introduction of a third concept” whereas syllogisms introduces still additional restrictions: “… it is conceivable at all events that the mind can establish, between its concepts, relations other than that of subject and attribute” [Coffey, 1, 385-386]. Subsequently it follows that “the question is not whether these relations … cannot be reduced to the subject – attribute relations… But the question rather is… whether there are combinations as of those other relations which… yield mediate inferences”[Coffey, 1, 387]. In spite of syllogistic figures the thought remains open for unforeseen arguments. Therefore in opposite to propositions capable of being represented as closed frames it needs open and movable semantic nets.

The statement on the irreducibility of text to a set of proposition and on the ensuing propositional transformability has been posed for the first time by Yu. S. Martemjanow who comes to the conclusion in favor of the priority of textual foundation that would not be that of propositional nature¹³⁸. The reason is in the essence of textual structure determined with the order of the inferential reasoning¹³⁹. The statement has been exemplified with the opportunities to represent the same proposition with a divergent series of simple sentences¹⁴⁰. Respectively the approach to disclose textual invariant structure (the so called canonic form) has been suggested where it would be restored with the recurrent steps of gradual propositional simplification¹⁴¹. In particular the transition from the formal (potential) structure of proposition to the actual division of sentence has been suggested to be conceived as a special kind of transformation called as modulation with the obvious reference to musicology¹⁴².

It would be here appropriate to stress the irrelevance and inessentiality of the opposition between simple sentence and clausal structure as well as between hypotaxis and parataxis. One can easily trace the opportunity of mutual conversion so that a textual passage can be represented with equivalent different propositional versions. Thus for instance the transformation is admissible of a simple sentence into a compound one: <The tree grew near a river ↔ There grew a tree. There was a river nearby>. Still more persuasive is the case of such transformation into a complex sentence with subordinate clauses (underlined are the conjunctive words introducing the clauses): Al this while I sat upon the ground very much dejected (D.Defoe. Robinson Crusoe, Journal) ↔ *It was ground that I sat upon when this while lasted because I was very much dejected. The generalized notion of taxis reflects just the case of such relative indiscernibleness of parataxis and hypotaxis. It attests again the priority of inference in respect to proposition in textual structure as well as the inessentiality of proposition for textual coherence.

Meanwhile the question arises as to the criterion of the equality of textual metamorphoses as well as on the nature of textual invariant. As far as such invariant can’t coincide with any metamorphose it must be an abstraction, therefore the question would arise as to the degree of such abstraction. The necessity of such abstraction has been included in N. Chomsky’s definition of deep structure with its “kernel” sentences¹⁴³. Respectively the dual structure of textual generation was suggested: that of basal system generating deep structure and its transformations in representational surface structure of propositional metamorphoses [Хомский, 2005, 89]. It implies reductionism in regard to textual representation that is supposed to be achieved with a finite number of transformational transitions.

The problem of invariants and their nature is of crucial importance for generative linguistics. It is here to remind that within the first cases of transformation applicable for descriptive purposes it was even the constant lexical units involved in the procedure that were regarded as invariant so that actually only the word order was to be changed¹⁴⁴. Meanwhile as far as the task becomes the generation instead of descriptive interpretation (opposed on the basis of action vs. contemplation) the transformations are conceived on a wider scale. In particular this task comprises now the necessity of producing texts that have not yet been manifested¹⁴⁵. In its turn the opposition of deep vs. surface structure appears as the correlate to that of the morphological inner vs. outer forms so that the transformations turn into transitions from one level to another so that they again become interpretation (in the sense of representation) of deep structure’s abstractions in the concrete and detailed surface structure¹⁴⁶. Such controversies of generation vs. interpretation entailed the necessity of reconsidering the nature of transformational invariant now to be regarded as the semantic abstract attributes instead of the earlier “kernel sentences” with their syntactic priorities¹⁴⁷. As the consequence invariant is now taken for an abstraction comprising only separate points of attributive space.

Besides, it is also to be taken into account that it is not all transformations that could be esteemed as equal so that the invariant would be restricted only with some abstract attributive parameters: for instance the transformation in the side of substantive (the S - transformations) come to the loss of information appertaining to temporality, aspect, actuality so that the transition from propositions to nominative sentences always generates enunciations endowed with more abstract contents than the primary statements. In its turn, the transition from substantives to propositions (especially with the involvement of denominative verbs) demands additional information concerning verbs (the V - transformation) so that respective transformations aren’t reversible. Therefore the mentioned transformational space is to be esteemed as inherently asymmetric and void of equivalence. It would suffice to compare the transformations <The wall is grey> S→ <Grey wall> V→ <The wall was grey> to notice the ensuing inadequacy of contents. Obviously the admissibility of such transformation within the retained textual identity can be demonstrated with the cases of single substantives replacing clauses in a hypotaxis as in the examples of N.D. Arutiunova: such cases are enabled as the consequence of the existent context of complex sentence where the temporal meaning is transferred to substantives with their referential environment¹⁴⁸. Such is the case with the notorious aphoristic statement “aprés nous le déluge” (after us (let be) the Flood) where “us” (nous) implies in reality the clause “(after) we’ve left this world”. Thus admissibility of the mentioned transformation is not universal but restricted with conditions arising just from textual integrity irreducible to transformational procedures.

Here it would seem appropriate to remind a simple example of A.F. Koni: «Стоит переставить слова в народном выражении кровь с молоком, чтобы увидеть значение отдельно взятого слова, поставленного на свое место» (It would suffice to permute the words in a folklore locution ‘blood with milk’ (meaning “robust, healthy”) to persuade oneself of the importance of a single word put in its appropriate place (the permutation will give here the designation of a milk polluted with blood)) [quot. Алексеев, Макарова, 1985, 81]. Still more this demand of textual precision for the identity of contents concerns the artistic text as it can be attested with a shrewd observation of L. Tolstoy «Как ни странно это сказать, а художество требует еще гораздо больше точности, precision, чем наука» (Strangely to say would it be, but the artistic work demands still more precision than science) (from the letter to L.D. Semenov, 1908) [Толстой, 1955, 3, 526].

Meanwhile the adherents of transformational-generative approach have suggested the widening of transformational procedures taking the deep structure of the obtained textual versions for identical and retaining the same semantic invariant – as, for instance in Yu.D. Apresyan’s statement¹⁴⁹ with the ensuing example <он женоненавистник> (he is a misogynist) = <женщины – предмет его ненависти> (women are his hatred’s target). It seems too obvious that to regard the cited sentences as semantically equivalent would be incorrect and precocious: the first represents an abstract idea (a person can be a misogynist as don’t hate any particular woman) whereas the second presupposes the manifestation of personal attitude without both inner character’s inclinations and the selection of a particular target for this manifestation. It is known, for instance, that the notorious Malthus has had a large family in spite of the declared misanthropy. The same objections concern the sample cited by V.G. Gak: <Люди не могли пройти через узкую дверь> (People couldn’t pass through the narrow door) = <Узкая дверь не пускала людей> (The narrow door didn’t let people pass) and <Они услышали, как рубят лес> (They’ve heard one to fell wood) = <Они услышали стук топора> (They’ve heard the knocks of an axe) [Гак, 1998, 449]. No need to remind that to can and to let mean absolutely different processes so that the effects of the mentioned narrow door have different meanings. As to axe’s knocks they don’t imply the process of felling trees. At the same time obviously there can be detected something common between the cited transformational versions though its essence is rather a problem to be detected.

One can only say about a relative identity of some elements of textual contents retained within the cited transformations. At the same time this identity can by no means be represented with a kind of “kernel” sentences. Here the contradiction of generative – transformational approach seems to be detected: at one side it declares textual irreducibility to propositional structures, at another side it presupposes reticently the necessity of textual deep structure’s reduction to the finite number of transformations. Reductionist’s viewpoint does thus remain ensuing essential restrictions as to the possibilities of the approach.