Національна Академія Мистецтв України Інститут культурології

These features of frame as enumeration are to be born in mind while applying it to semantic analysis just as an enumerative derivative questionnaire, though the conclusion as to its effectiveness for the identification of the described lexical stuff seems to be overestimated⁵⁵². One admits also the indefiniteness & ambiguity of frame but relies upon its being useful for the presentation of something of nonverbal nature⁵⁵³. It is the limitations of listing procedures determined with filling up the ready slots that entail the refusal from the initial aspirations of frame analysis to reveal and represent the latent contents because it by no means can be contained within the borders of a plain questionnaire⁵⁵⁴. As to an artistic work it is evident that latent and implicit (background) side plays primordial role, and subsequently the devices of the frame analysis can’t be accepted as satisfactory. To sum up, there are no particular reasons for special separating frames from other forms of appositive & enumerative listing structures. The stress would be better put upon the ability of such structures to become devices for the self-description of a text as they are always implicitly present in a latent form within the textual borders. In particular it concerns the predicate arrays describing a plot in the form of taxis.

The same concerns some folklore genres such as riddles that present list of attributes of the object to be guessed. For instance the riddle «одного не знаю, другого не вижу, третьего не помню» (the solution – рождение, рост, смерть) meets perfectly the features of a predicate frame representing a plot (here – the chief moments of a human life). [Садовников, № 2135] Another riddle «кости лубяные, лубки костяные, всходы живые» (solution «яйцо») [Грынблат et al., № 1220] can be regarded also as a fold of a myth of cosmogony, be the role of an egg as a cosmic symbol taken into consideration. That a riddle allows and presumes frame representation can easily be seen in the so called cadre making where the initial situation undergoes transformations so that the particulars of the solution be referred to the respective details of the description presented in the riddle⁵⁵⁵. The parallel between cadre and frame seems to become meaningful. It is worth mentioning that such notions as frame and slot were adopted from cinema to visualize verbal material. At the same time as far as riddles are concerned it would seem more reasonable and appropriate not to use the notion of cadre, because there existed for ages another term which perfectly corresponded to riddles’ conditions. It goes about the so called emblems that belonged to baroque literature as a peculiar genre uniting visual and verbal images. It was emblems that reflected riddles’ situations most adequately.

Frame questionnaire in this respect reproduces a somewhat widened syntactic scheme in the manner of the famous R. Kipling’s verse (from “The Elephant’s Child”): “I keep six honest serving men, / They taught me all I knew. / Their names are What and Why and When / And How and Where and Who”. This approach was further developed by Ch. Fillmore in his “case grammar” where the possible environment of a predicate had been envisaged in the manner of questionnaire⁵⁵⁶. The chief restriction of frame method becomes here evident in its attachment to propositional structure. Appositions come then into play as the results of the dissection (dismembering) of a text with the ensuing isolating abstraction where the meanings of isolated fragments are changed.

In its turn the recent achievements of computational linguistics supply analytical apparatus with the so called tags or tagging lists.⁵⁵⁷. It is obvious that tags are compatible with the thesauri of the so called in situ types, i.e. with the descriptions of texts based upon the methods of distributional analysis where the combination of lexical units and their presence in those or these places of text are taken into account [Дабралюбау et al.]. It is easy to show that the notions of frame and tag are reversible: if frame describes the environment of the concept to be found (as in riddles) the tag presents possible environment associated with the supposed concept. At the same time in both cases it goes about listing structures. In particular it becomes evident that frame entails tag as the valences of key words (attested in frame’s slots). At the same time from one side tags are to be seen as the generalization and expansion of frames represented as the arrays of descriptors that fill respective ordered set of slots reflecting so an algorithm of a narrative. From other side tags as listing structures or indices determine the semantic compatibility (or the so called “semantic isotopes” after Greimas) i.e. the presence of a respective semantic feature⁵⁵⁸. Thus a tag unites both paradigmatic aspects of a semantic field and the syntagmatic enumerations that contain lexical units used for mapping a situation. Isotopic places determined with tags can be regarded as a mediating structure laid between paradigmatic taxonomy and textual strings. It is their intermediary position that makes them especially valuable for the descriptive tasks. In contrast to text any enumerative structure (listing) is void of those latent references and implied contents that determine textual coherence. Such structures as zeugma or homogeneous subordinate clauses or paratactic rows of sentences with syntactic parallelism are loose in their inner connections; they can in some cases change the order of sentence without abusing the sense. Such interchanges were impossible in a text: the same sentence at the beginning and in the end would have perfectly different meanings determined with the syntactic perspective. Such reduction of a text to a list means also the replacement of the actual referential net of this text so that another net would prevail instead, namely the net that an isolated expression bears.

Frames and tags are usually considered as the intermediary steps towards codification (programs, algorithms) to reproduce text. Meanwhile the aim of listing transformations of text is not registers (indices) and summaries only. The procedures have their goal in detecting the supposed generative algorithm although the possibility for the reproduction of the original text would be excluded from the beginning on. A textual summary (for example, the description of a narrative plot) would admit very different versions of textual reintegration and regeneration deviating very far from reproduction. Such virtual program admitting very wide range of reintegration’s versions is supposed to exist as textual analytical epiphenomenon (satellite) that becomes disclosed and revealed in data representation.

It would be also reasonable again to remind that the borderline between propositions and appositions is by no means insurmountable. Each enumeration or apposition can easily be transformed at least into existential or indentifying (defining) propositions together with immediate inferences. Listing structure then is to be regarded at least as a row of the statements on the existence of designated objects and their definitions. For instance the apposition “rain, street, bridge” can be easily converted in the propositional structure “It rains (or not). There’s street. There’s bridge. As far as it rains there’s no snow. The street (and not a road) can be situated in an inhabited settlement. The street must cross a river or a valley or a road”. Such inferences are already the consequences of “logical quadrangle”. The chief deficiency of dismembering a text into index is that it approaches chaos with the disappearance of ties between the entries of such derivative index. It is not merely the loss of information: it is the very transformation of text into chaotic entity that prevents indexation from being a descriptive device apt for representational tasks. It concerns in particular the so called “chaotic enumerations” (the term of L. Spitzer) admissible only as an intermediary step in producing a textual score. A perfectly different case is to be observed with collocations. Actually they can be regarded as incomplete sentences presupposing conversion into questions demanding answers. It is propositional transformability that marks the properties of collocations. Then the separation of a collocation enables generating a kind of cento built of the parts from different sentences. Thus the tasks of textual description & compression converge. In particular it goes about compression instead of excerption that is to be carried out so that the dilemma of compression vs. excerption arises that’s to be solved with descriptive methods. The preference for dynamic compression ensues already from its cumulative effect that static excerptions are void of.

Meanwhile the descriptive device of incomplete sentences (as implicit questions) is widely used in theatre in comic dialogues when one of the partners begins the sentence and another ends it trying to guess the possible continuation. Such is, for example, the conversation between brother and sister from A. Salacrou’s play “L’archipel Lenoir ou il ne faut pas toucher aux choses immobiles” (The Archipelago Lenoir or one oughtn’t to touch the immobile things). Marie-Thérèse: Tu ne vas pas tout de même pas reprocher à maman ... (Nevertheless thou won’t reproach our mother …); Victor: … ce que tu as pu comprendre, bien que nous n’en jamais parlé (… that thou hast been able to comprehend although we have never discussed it). Another sample can be found in E. Rostand’s “L’aiglon” (Eaglet, 1.1). Marie-Louise: “Un frère qui …” Bombelles: “Fils d’émigré, reste émigré”. (‘And the brother who is … The son of the emigrant and remains emigrant’ These words concern the Emperor’s son). Description thus becomes a cento made up of the cited collocations and observer’s inferences.

As far as it goes about purposely performed elimination of some textual segments and textual abbreviation in general a lot of information will be lost as the result of such transformations aiming only at semantic references to be taken into account. The preliminary step towards data representation can be said to be a transition to a “telegraph style” with its “jingloisms” and “wellerisms” (to mention Dickens’ “Pickwick Club”) resembling archaic languages of isolated and incorporated types. One excludes pronouns, personal names, auxiliary verbs and particles (replacing them with other auxiliaries). Besides, one abstracts from personal intentions (in particular with making predicates impersonal) and localities / temporalities so that purified lists would arise. Propositions are to become substantives that presume the reversed process of the formation of denominatives. In this sense each list of substantives as nominative sentences can be converted in an equal list of denominatives (in particular with the aid of auxiliary verbs) so that the equivalence of the Nominal and Verbal versions of the list would become evident, the importance of denominatives consisting in their opportunities of the return to text⁵⁶⁰. One could compare such data representation with an abridged literal interlinear translation for a language of isolated type when it goes of listing structures of isolated lexical units. At the same time lists can convert in artificially constructed words-sentences that convert in their turn in composed neologisms. Such composita represent a seeming interlinear translation into a language of incorporated type (as in the L. Sterne’s work). Of course they include both proper textual lexical units and those added as supplementing interpretative comments.

As far as the above discussed standard descriptive methods are concerned they come back to the works of I.A. Melchuk (who developed the works of L. Tesniere) and are usually labeled as those of [Sense ↔ Text] model [Мельчук, 1974 (1968), 278]. Textual elements are divided here into the polar confronted sets of generalities vs. particulars (actants vs. circonstants according to L. Tesniere). While transforming this approach into [Theme ↔ Text] model (where the musicological methods of L.A. Masel’s “integrative analysis” have been applied) [Жолковский, Щеглов, 1996 (1970)] there has been chosen the experience of proverbial studies (with the examination of possible alternative to the described textual statements) as the paragon. While dealing with a text represented as a string of propositions one has to bear in mind that each sentence here becomes only one representative of a set of equivalent judgments that can be possibly obtained reciprocally from each other as the conclusions inferred after the rules of logical square. Therefore each statement from a string of the kind is to be conceived together with a set of adjacent immediate consequences. Accordingly the first step in data representation would consist in supplementing the existent statements with the implied adjacent ones so that a set of sets would arise in the resulting array. Such would become a plausible interpretation of the mentioned models. Meanwhile it becomes obvious that such set of the presumably equivalent sentences presupposes the existence of the previously determined invariant that is supposed to represent the already known sense. This reticent conjecture on the preexistent knowledge of the sense is the principal deficiency of the [Sense – Text] model. Therefore it seems more preferably to apply the attributive approach (developed by A.V. Bondarko) where one deals only with the features of attributive space without involving the conjectures on the preexistent semantic invariant. As to the equivalence of the transformative textual versions it would become necessary to seek for the respective invariant as a problem represented with the attributive features and not as an autonomous separate entity of the known sense.

All these and similar procedures of compiling and compressing text can be esteemed as auxiliary artificial devices that disclose the inner textual potential structure. Meanwhile with the transition from potentialities to actualities the duality of proposition vs. apposition disappears together with that of generalities vs. particulars. Messages becoming from textual propositions are also appositive structures in the sense that they don’t depend upon direct predicative net (taxis) of the text being attached to the actual predicates (rhemes) that drift together with communicative “flux and reflux”. In its turn it is particulars and not generalities that become rhemes making thus the whole field structure of centralization shift. Thus the whole map of the preliminary textual description must be corrected and reinterpreted. As an essential corrective additional moment for such reinterpretation the communicative intention must be regarded⁵⁶¹. Together with the involvement of intention the new situation comes: the shift of centers in actual structure entails also the drift of compatibility as the meanings of separate locutions are being reviewed. As far as referential net of the whole text endures modifications the references of these locutions concern otherness (that arises already with the segregation of textual segments introduced with proposition). Textual filament of these references doesn’t remain the same, so the locutions concern other objects than those supposed to do as the elements of potential structure. Direct designations become circumscription of something other when taken actually. Thus to represent analytically actualities one has to interpret text as a periphrastic transformation and respectively to apply motivational net to study it. In particular one would find in the completive structures the means to represent in the fullest scope periphrastic textual metamorphoses removing at the same time the dualistic contradiction between proposition and apposition.

To render the nodal textual structure that would correspond to the concept of taxis there has been suggested the term clausal structures where it is stressed that an enunciation is an implication of a series of premises and vice versa “an empty clause has neither premises or conclusions” [Ковальски, 1990, 30]. Thus the specifically English grammatical term clause designating the distinction of subordinate proposition from that of autonomous status is applied to the nodal vs. linear (enumerative, listing) structure. Respectively one can say of predicative perspective that can be defined as the interrelationships of the predicates as demonstrated in subordinate clauses. Obviously the English-speaking term clausal structure corresponds to the already discussed concept of syntactic perspective. In particular one could say also of predicative perspective as far as the taxis is concerned. Thus the nodal structure with its subordinate relations that replace coordinate relations is determined with the generation of perspective. It is to stress that the presence of clausal structures is a well attested fact of poetical practice and can be exemplified with the samples of the figure of zeugma in such widely known verses as R. Kipling’s “If” (“If you can keep your head / When all about you / Are losing theirs …”) and K. Simonov’s “If Thy Home is Dear to Thee” («Если дорог тебе твой дом…»). Such clausal structures can be very visibly represented in graphical way with the device of multiplied bracketing (the device known in particular in computer techniques as the alternative to the so called Polish notation). For further subordination there can be involved also the generalized concept of taxis. Due to this growth of nodal structures separate propositions are to be regarded within the whole syntactic perspective as the integrating textual force. The dependent state of propositions turned to clauses betrays their integration in nodal structure. Obviously nods in L. Tesniere’s sense represent field structure and respectively the centralized semantic space belonging both to text and to descriptive representational structures as those of glossary. Thus ultimately a text can be represented as a single propositional structure of hypotaxis or as a zeugma.

Meanwhile neither zeugma (clausal structure) nor the above discussed indexation & intitulation (enumeration) are valid enough for textual data representation. Artistic text is a huge collocation (to remind the mentioned ideas) that needs respective representation. It means that while retelling a paragraph the principal locutions of the style must be retained becoming thus the prototypes for the determining the attributive invariant. Then one has to use both clausal structure and indexation with the cited locutions as the insertions in the observer’s own hypotaxis, evident preference of the clausal structure being their capacity to become abridged and compressed ultimately till a single sentence.

At the same time to represent a text in a single proposition with the outlook of clausal structure turns out to be the same as to represent it with a thorough enumerative structure of apposition. Here propositional and appositive structures coincide in the ultimate scope of representation. It will become evident when one carries out an experiment of converting a text into an index (thesaurus, glossary). The seemingly disparate enumeration of lexical units or phrases obtained in such a way reveals its stratification in the already discussed two extreme levels of generalities (predicates) and particulars (circumstances, complements). Such approach of disjoining generalities and particulars (and actually of purifying a text from details) doesn’t actually contain essential novelties. It is worth reminding the anecdotic enunciation of G. Clemenceau: “Un sujet, un verbe, un complément direct, c’est suffisant pour écrire. Si vous tenez à ajouter un adjectif ; c’est plus delicat, venez me demander conseil” (Subject, verb and direct object are sufficient to write. If you insist upon adding an adjective, it is more delicate, so please consult me). Such approach being applied consecutively, one would obtain ultimately abstract constructions of the (S)HE DOES IT – type applicable for the representation of any proposition. One can say here of the mentioned