Вы находитесь на архивной версии сайта лаборатории, некоторые материалы можно найти только здесь.
Актуальная информация о деятельности лаборатории на lex.philol.msu.ru.
Melnikov G.P.- Systemology and Linguistic Aspects of Cybernetics


| << back | | continuation |



3.4. THE CATEGORIES OF THE SYNTAX AND MORPHOLOGY OF NATURAL LANGUAGE AND OTHER OCCASIONAL COMMUNICATIVE SYSTEMS.

Cognominative syntactic meanings and senses.

Our scheme for the formation of communicative and thought units and the interlinkings between them enables us to examine natural language as one of the realisations of the principles of meaningful intercourse, and therefore to transfer conclusions about the nature of linguistic categories to cybernetic occasional communicative systems.

The semantics of linguistic units which are named syntactic are very often regarded as the capacity of these units to signify the relations between content components designated as "lexical", nominative "material" and "semi-sign" language units. In such treatment there is always a danger of misunderstandings. We should therefore direct our attention to the following.

Concrete and abstract patterns of these denoters, like the denoters themselves, can join combinations by changing into higher level units. One of the characteristic features of these amalgamated units is a network (structure) of links and relations between components. But since amalgamated patterns can also take the role of senses as reflections of denoters, the links structures and the relations of components of complex patterns can be included among those features on which basis a communicative classification of patterns is produced. Consequently a communicative abstraction, changing into a meaning, as one of the aspects of a moneme, can be an abstraction of not only the obvious features of each of the patterns of a certain combination, but also an abstraction of the obvious particular features of the link methods of complex patterns' components.

However, since both amalgamated and non-amalgamated patterns remain independent representatives of certain communicative classes in the communication act, the communicative role of abstractions personifying general features peculiar to each representative within the boundaries of some particular class also remains invariable. Therefore both types of meanings examined are only distinguished from one another by the level of complexity of the denoters, and in the communicative relation (and consequently in the linguistic relation) these two types of meanings and accordingly the two types of monemes and morphemes are identical and represent one category of meanings, which can be called denominational or nominative [206, p.345].

The more detailed a pattern of a sense it is desirable to reproduce in a listener's consciousness, the fewer chances a speaker has to select such a single meaning, and having perceived this meaning a listener will be able to make use of it as a hint of a sense with all the details implied by the speaker. In such cases as we have already seen it is essential to choose not one but several components or fragments in that sense, and to select for each of these a separate, most suitable nominative moneme. Then the listener will perceive (through Identified meanings) hints of several characteristic junctions of the named sense and will be able to recreate a pattern in his consciousness which corresponds more fully to that complex sense implied by the speaker.

In other words the sense will in this case be represented by a complex sign consisting of several denominational, nominative monemes.

However if it is taken into account that a multi-moneme nomination can correspond to its purpose only with the condition that the listener has not only correctly identified all the listed or enumerated nominative meanings, but has also correlated the roles of these meanings correctly, as a means of hinting at fragments of the single integrally whole sense, then, in receiving a moneme with such a multi-component meaning, a listener must correctly understand the actual significance of each meaning in any concrete act of multi-moneme nomination, and must understand the place of this meaning in the unique structure of communicative relations of the monemes used.

Consequently the need may arise in the communication process to in some way help the listener to understand these special moneme characteristics which are external in respect of the sense mentioned, as members of a single whole - a multi-moneme name in which each moneme has its share of the nominative function, and is a cognominant, i.e. a co-designator of a set sense so cognominant moneme relations themselves (a group which must include for example determinant relations) can become an object of nomination, an intent of a speaker, i.e. a special variety of sense, which can be called cognominative syntactic sense.

In representing a structure of certain relations and significances of cognominant monemes, a cognominant syntactic sense, like any other, has a structural pattern nature. Therefore in order to express linguistically a cognominative syntactic sense it is always possible to use the similarity between the pattern of the cognominative structures and significances of the moninative monemes used, and the meanings used of any other denominational nominative monemes, in order to name the corresponding cognominative syntactic characteristics of any moneme or their cognominative combination by means of a morpheme with a denominational meaning. In this case a linguistic sign remains the representative of a nominative meaning, but a meaning made similar in order to dictate a cognominative (e.g. determinative) syntactic sense. Therefore it can be said that the communicative emploi, later defined as basic function, of a moneme may sometimes not coincide with its communicative role.

If certain standard cognominative syntactic patterns are extracted from a large number of concrete multi-moneme nominations and, having changed into independent abstract patterns, they enter into a usual association with specialised linguistic signs, then it is only in such language that cognominative syntactic monemes will appear with what is being signified - cognominative syntactic meanings, and with their signifiers - formal syntactic cognominative morphemes, i.e. with morphemes having syntactic emplois, and not just roles. An emploi of this sort can be, for example, an Indication of the fact that a certain complex linguistic sign is used attibutlvely, i.e. as a precise definition of another sign. 

Communication, Predicative Sense and Predicative Meaning.

Although a complex type of nomination utilising monemes which include several co-designator monemes to express complex ideas, leads to the rise of a special syntactic sense (cognominative), it is ultimately directed at effecting the same communicative act which is also realised by means of a one-moneme sign, if all that is necessary is to name the sense. Consequently, syntactically expressed attribution (i.e. definition) remains a variety of nomination and is distinct from the designation of the relation, different in nature, between sense units, i.e. from predication. If we do not define the essence of this phenomenon we cannot be sure of the practical use of the ontological scheme of the relation of language and thinking, and we cannot understand what a sentence is, although it is absolutely clear that without a sentence as an expresser of a "separate, completed thought" there cannot be meaningful intercourse between either people or machines.

If all thinking - both concrete and abstract - is "pattern" in nature, we can understand by the term predication a minimal act of recognition, reduced to the formation of a certain new pattern which is presented to a person as a precisely defined variant of a certain pattern which has already existed, i.e. previously established knowledge. If this act of recognition is stimulated by another person by means of language, then the section of speech flow leading to such a result is called "communication" (but not a sentence). Communication has as a rule two components: theme and reme; these are evident according to the character of the sense expressed by them which depends on what a speaker considers is new for the listener (i.e. forming for the listener a new, or, more precisely renewed pattern as its sense), and what is considered only an indication of what this renewal relates to, (i.e. having as its sense an old renewed pattern).

Consequently, the elementary thinking act of recognition can be characterised (according to our conception of the thinking activity mechanism) as an act of the displacement of the pattern of a real or imagined denoter by a more precise - in the opinion of the recognizer - pattern of this denoter, while the communicative act, directed not at simple nomination, but at linguistic transfer of an element of the speaker's knowledge to the listener, i.e. the communicative act, stimulating predication, is ultimately reduced to nomination of the pattern of the corresponding denoter in such form as it was previously known to the listener, and to a subsequent stimulation of a more precise pattern of this denoter in the listener's consciousness. It can also be said that the theme of communication names the initial knowledge of the listener, while reme provides in some way or other the renewal of this knowledge. In cybernetic terms the sense of the theme is the operand and sense of the reme is the operator, and the result of such an operation is the transformed content of the operand of necessity achieved only owing to the thought act stimulated by the senses of theme and reme but already being effected on the basis of a concrete or abstract pattern of extralinguistic thought, and not verbally. It should also be noted that the resulting sense of the communication received by the listener has no linguistic expression in the signs of this communication.

Very many methods of forming a new pattern, i.e. reme sense, on the basis of the Initial, i.e. the named, sense of the theme, can exist in principle. In a particular case for the nomination of a new pattern the nomination of an old one with an Indication of those features which distinguish the new pattern from the old might be necessary.

So we see that communication, embracing as a rule two complex signs for the nomination of theme and reme, must reflect a special type of integral wholeness of pattern, formed by means of these two nominations - a wholeness reflecting the fact of the change of the one pattern into the other, that is, a wholeness representing something in the nature of a transfocal pattern of the evolution of a certain fragment of knowledge.

But a sequence of monemes used in two such nominations to express one act of predication can prove to be such that for a higher guarantee of the listener correctly understanding the speaker's intention, the latter will require clarification as to what monemes are components of the theme, and what the components are of the reme, i.e. what the predicative significance of the monemes used is, or what the predicative structure of the communication is. The variety of structural relations between monemes in the communication act will be called the predicative syntactic sense [125-126].

It is clear that the predicative sense, like the cognominative syntactic sense, may not have specialised monemes, and then, if necessary, the predicative sense must be expressed through nominative or cognominative (in their role) monemes. But the most typical predicative structural characteristics of communications, i.e. predicatively significant networks of relations of moneme sequences, can also be abstracted, and become usual units with specialised linguistic signs fixed to them. Only in this case does language have predicative meanings and contain predicative monemes (and consequently morphemes), for which the expression and the designation of predicative senses changes into their emploi, i.e. which change into formal expressers of predication.

But since any linguistic elements, including those whose function, i.e. emploi, is the expression of the fact of predication, sometimes have roles which do not coincide with their emploi a problem arises: is it possible to determine a factual, real theme and reme of communication independently of how the communication is formed?

Although concepts of theme and reme themselves are not generally recognised at the present time, nor generally known in linguistics, the question of the methods of isolating them, which is expressed in linguistics literature in terms of seeking logical stress, the psychological predicate etc., is not new [18; 30; 31; 51; 82; 135; 140-142; 148]. 

Principles of the division of communication into theme and reme; the nature of the parts of a sentence. If we start with a definition of the essence of predication it is easy to conclude that to establish the boundaries of the reme and theme of communication one must be sure above all that a speech section is not a simple chance selection of words, but a communication. If a listener has perceived the content of this speech section and has understood that it is a communication, the question must then be asked: what previous knowledge must the given communication replace with new knowledge, i.e. what old knowledge is negated by means of this assertion expressed in the communication.

On the basis of these reflections we propose the following methods of establishing the boundaries between theme and reme. First a diagnostic context is employed to show that the sequence of linguistic signs is a communication "(it is true that....) (in the sense that it is false that....)" If this is indeed a communication, then it is possible to isolate without difficulty a reme by means of new diagnostic context, having inserted such Indicator words into the communication as "...(in fact)...(and not)..." That speech section which must be concluded between these indicators is a communication reme, while the remaining components are the theme.

Let us look at an example. Let the speech section be: "They will be very busy tomorrow". Let us suggest that starting with the concrete conditions of communication we have established that the empty places in the first diagnostic context can be filled up in this way: "(It is true that) they will be very busy tomorrow (in the sense that it is false that they will be very busy today)". Consequently, this speech section is indeed a communication. It is then clear that if indicator-wards can be inserted into it thus: "They will be very busy (in fact) tomorrow (and not today)", then "tomorrow" is the communication reme, and "they will be very busy" will be the theme. [128].

It can be seen from this example that a division into theme and reme may not coincide with the division into subject group and predicate group. Apart from this, which is extremely important, in any communication without exception there must be a division of the communication into its main parts, if the communication according to traditional terminology, "express a separate concluded thought". Frequently in those works where the problem of the so-called "actual division" is examined, its presence is only recognised when the sentence has "logical stress". But if the parts of the communication may not coincide with the parts of the sentence, then what are the parts of the sentence? This question can also be solved on the basis of our scheme.

The essence of the difference of a communication from a sentence in the light of the ontological language scheme developed is explained in the following manner.

The positional and lexical realisation of the division of an utterance into relation groups depends on an associative sense definition of the role of each sign within the boundaries of the theme and reme group. In this case the corresponding kernels of theme and reme should hardly be called subject and predicate.

In other words if language does not contain special signs, whose emploi is partly made up of reflection if not of the emploi then just the occasional role of communication signs as expressers of theme or reme content, and this occasional role is established on the basis of the sequence order of signs and their senses, then in communication there are not even any parts of a sentence. For example, the Russian expression "His sister is a Doctor of Science" is a communication, but not a sentence, because having theme and reme in its composition this sequence of words contains no morphemes of formal emphasis of the "themeness" or "remeness" of the corresponding parts of the sequence. But in the Tuvinian communication "O-lar otuz-laz", the "they are sitting" element "lar" in the reme points to the formal link of the reme with the theme "olar", and therefore both the subject ("Olar") and the predicate ("oturlar") stand out formally in the communication [126].

Such an understanding of "predicateness" as a category different from "remeness", and of "subjectness" as a category different from "themeness", leads to the following definition of the concept "the sentence".

We can best name as sentences only those communications which have predicative meaning, i.e. whose wholeness is farmed by specialised predicative monemes. Naturally such formation is most frequently used in those cases where the actual role of predicative monemes is created by a need to designate a predicative sense. Monemes with the emploi of expressing the sense of a theme change the kernel of the theme into a subject; monemes with the emploi of expressing the sense of a reme transfer the kernel of the reme into a predicate. The theme group and reme group are thus transformed into a subject group and an object group.[126; 127; 206].

However if monemes appear with a particular emploi, the possibility sometimes arises of also using these monemes in roles which do not correspond to this emploi. As we have already seen, nominative monemes can prove to be actual expressers of cognominative or predicative syntactic senses, The presence of predicative monemes makes their use possible "by compatibility" in the role of expressers of a non-predicative sense, either nominative or cognominative.

Other parts of a sentence are defined in a similar way. For example, definition takes place in a sentence only when there are words in it which do not simply bear cognominative sense or make a sense expressed by another word precise, but are capable of receiving an accordant indicator exclusively as an element testifying to the cognominative emploi of this word and bearing no other sense hint.

In this light the Turkic two-word combination "tachta masa" (i.e. wooden table), contains the attribute "tachta" which is not formally a part of a sentence which can be defined, even if this combination appears in the role, for example, of subject in a sentence. The cognominative syntactic sense in this case does not have any expression via attributive syntactic meaning [127; 130]. 

The structure of a link text. A sufficiently clear differentiation of the levels of signs, meanings, and senses of not only nominative, naming units but also of purely grammatical, cognominatively and predicatively syntactic units, enables us to touch upon the problem of the connection of units in speech sections larger than communications, i.e. the problem of the "structure of a text". For example, it is obvious that if a communication process is effected according to a worked-out ontological scheme, then the one intent or plan can be divided by various methods into communicative portions. Therefore the number of communications per one portion of "intent thought" cannot in principle be the same in intercourse in various languages, and it can also depend both on the individual style of a speaker and on the particular features of the intercourse situation, genre etc.

In particular if communication was an effect or consequence of a need to inform the listener about the occurrence of a complex process, the dynamics of the change of content, which must be conveyed by linguistic means, can be correlated with the dynamics of an actual external process, which has been observed by the speaker (denotative level), or with the dynamics of reflection in the process of getting to know this process which is being observed (cognitive level), or, finally, with the dynamics of the plan or of the programme for revealing this process to the speaker (communicative level). The non-convergence of these three planes requires an awareness of the means of co-ordinating the units of one of them when they correspond to a single unit of the other, for example, when one integrally whole sense unit is expressed by several portions of signs in a chain of communications. The establishment of the most typical structures of sense components and sequences of signs corresponding to these structures enables us to define more precisely not only the nature of communication units and sentence categories, but also to give definitions of the criteria for the change of parts of a sentence to parts of speech which are clearer than those in linguistics literature. For the moment let us dwell on one aspect of the problem of text structure.

If in acts of communication between people (or indeed between machines, working according to the principles of orientation in the environment and meaningful communication shown in our scheme) a listener has perceived only the theme of the communication and has not yet finished waiting for a formulation of the reme, then since the sense of the theme has been understood, in the consciousness, as an identifying and forecasting apparatus, the hypothesis concerning a possible continuation of the communication, i.e. concerning a possible reme with a given theme, can now come into being. And this act is nothing more than a prognosis of the next step not only of content but of the form of expression in the text's structure. If a communicant transmitting a communication has prognosed reflexively such a reaction of a listener, he cannot waste linguistic means for a detailed check of the prognosis, but can manage with a minimum of meanings hinting at the sense of regular steps of predication Imposed by the text on the listener. Thus possibilities of ellipses arise, i.e. the missing out of many speech current units obvious to the listener when the content is steadily revealed by means of these speech units and also the possibility of using parts of a sentence not just according to their emploi of being the theme or reme.

When the principles of steady, uninterrupted prognosing are observed, such units as conjunctions have a communicative interpretation (and are not thrown out of the text as, for example, in the description system of text structure in I.P. Sevbo's book [157]). The role of conjunctions is "commenting" (on the part of the transmitter) on the correspondence of the receiver's prognoses and his needs for new information in respect of the continuation of the information current. For example, from the received information section there follows something completely contradicting the content further conveyed. In this case the conjunction "but" is used. "I am very ill but I will definitely come to your place". The convergence of prognosis and actual continuation can be prompted for example by the conjunction "and": I am very ill and I will not be able to come to your place today".

It is easy to conclude from such a view of the function of conjunctions in a text how important they are in making the sense linkage of a sequence of communications evident. A speaker inserting conjunctions lets the listener constantly know how he is presenting to the listener the rate of revelation of the sense of the text. But a listener, continuously assured of the pertinence of the conjunctions, is confirmed - by the fact - in his conclusions as to the correctness of his interpretation [18].

Thus with the transfer of information to another person in natural language the conjunctions system fulfils to a large extent those functions which in remote control technology are called "quitting signals", enabling the communicants to be sure that information passes along communication channels without distortions. Only on the level of the interaction of unique fragments of senses in a unique sense field is such prognosing reflexing, and inter-correction of the communicants possible. But when these sense processes are a result of advance reflection effected in machines, they can only be realised, as we have seen, on the basis of a multi-parameter and multi-layer comparison of the patterns of the external environment, of transformations of one pattern into another, and the classification and extraction of these patterns according to the types of activity and the creation of patterns of standard types of of transformation of patterns etc. It could be shown that these are, above all, those processes which in abstract form are reflected in the particular features of the tensor approach to solving complex tasks, and that this tensor approach is most naturally evident in the principles of modelling and prognosing when solving complex cybernetic problems by means of computational environments and systems [52; 75].

Parts of Speech and "actual division". If cognominative and predicative sense is expressed in a particular language by means of special auxiliary morphemes which have been defined as proper syntactic ones, and if in the language there are morphemes with cognominative nominative, and not syntactic, meaning, then nevertheless in such a language according to the criteria already established there exist only parts of a sentence. And only if there are morphemes in the language in which both syntactic and nominative meaning is presented simultaneously we can speak of parts of speech existing in the language, for example, not simply of the syntactic category of definition, but also the morphological category of the adjective. The fact that a noun and verb are present is determined in a similar way. If co-ordinates of theme with reme merge with nominative elements of static meaning (in theme) and dynamic meaning (in reme) then these parts of speech develop in language: a noun and a verb [127].

The Interpretation of the reme as a part of communication leading to the appearance of new content, representing a renewed pattern expressed in its initial form by signs of the theme makes it understandable that the process of transmission of new knowledge by linguistic means can be multistage. Any reme, insofar as by means of it new knowledge has been formed from the old for a listener, prepares the ground or conditions for the following communication because the knowledge that has just been transmitted can be named in this new communication as something already known, i.e. it can become the theme content of a new communication. In this case we have the simplest variant of the communications link in a sequential, linear, sense, text-structure: the resulting sense of the preceding communication becomes the theme sense of the following one. Hence it is clear that there are potentials in language for the rise of a class of specialised signs to take the role of the theme, whose content are specifically for the role of the reme, the theme of which is the resulting sense of the preceding communication. Signs which have the specialised function of carrying out the role are represented as a category of adverb. The utterance "we read beautifully" can be interpreted in such a case as a short method of expressing two succeeding communications: 1) we (theme) read (reme); 2) all this (i.e. all that we read, the fact of our reading) is beautiful.

A whole situation, and not just a separate word, is understood as being part of a certain class, and an attribute of this situation conforms with one determined according to the implied class. In the Russian language a situation becomes the most indefinite class - the neuter gender; therefore Russian adverbs are formed in the first place from the neuter gender of the adjective "to read a book beautifully", to "run home quickly" etc. [131]:  beautifully : Krassivo; quickly : bystro.

The rise of other grammatical categories can be explained as a peculiarity of their standard position and function in the texts' sense structure, examined from the point of view of "actual division", i.e. the interlinkage of several elementary communications, shrouded in the form of a single syntactic unit - a sentence. So, for example, if two linked actions of the same functioning person are the content of two communications and the sign of this functioning person very frequently takes the role of the theme, then in order to reduce the number of signs a single sentence is constructed, in which the functioning person is expressed by the subject, one of the actions by the verb-predicate, and a second by a special verb form representing in Russian a special part of speech - the gerund. For example, two communications, "the man was walking down the street; the man was limping" is put forward as one sentence. "The man was walking down the street limping" - the very fact that the gerund is used testifies formally that the implied theme to the concealed reme, represented by the gerund, is in a typical case the theme of the first reme, represented by the verb-predicate.

Here it becomes clear that "-ing forms" in English, for example, only occasionally coincide in their syntactic sense with Russian gerunds.

It could be shown that one can give a definition for such grammatical categories as the case by a similar method of showing the true composition of communications and showing the systemic relations between the cases.

However in the light of what has been said it is natural to ask this question: haw can the fact be explained that in some languages for example special grammatical meanings for the expression of the communicative emploi of linguistic signs have been created, while in other languages the emploi remains unexpressed, and the corresponding auxiliary or link information is expressed by non-auxiliary meanings, only descriptively, or pluralistically.

Systemic linguistics, i.e. linguistics based on the achievements of systemology, explain this by distinguishing the determinant of linguistic systems.

If we start with the division of all the languages of the world into four most strongly distinctive types, i.e. if we start from a typological classification of languages going back to Humboldt and remaining generally accepted until now, then the internal determinants of these languages can be formulated in the following way.

Inflective structure occurs as a result of the tendency to manufacture and unify beforehand linguistic signs for all senses, socially meaningful for members of the linguistic group having transformed through this these senses and the signs expressing them into usual ones; and other types of languages are attuned for the occasional optimisation of signs which are being formed for actual senses.

An agglutinative structure brings about occasional optimisation through economic use of auxiliary morphemes [91; 93; 109; 126; 127]; Root-isolating languages achieve the same effect by wide use of morphemes with non-auxiliary meanings to express link senses, while incorporating languages have a special, deeply adapted technique which allows them to manage with the minimum theme information and to reduce a communication to an occasionally formed "word-sentence" as the expresser of a reme.

An analysis of these four basic types of natural language structure from the viewpoint of the external determinant bring us to the conclusion that the internal determinant distinctions are a result of environmental external determinant distinctions and the main factor in these environmental distinctions is on which level of generalisation of the units of consciousness the minimum mutual informedness of the speakers is to be found before a standard act of communication [90; 95; 96; 113].

If we distinguish just three levels of these units - the level of current patterns capable of the occasional actual function, the level of individuals' generalised patterns. basically the sense of proper nouns, and finally the level of generalised gender and aspect patterns, appearing mainly as usual senses of common nouns, then we can show that incorporation occurs when the speakers are usually deeply mutually informed on all three of the levels of consciousness mentioned; root-isolation is best in those cases when insufficient mutual informedness on all these levels is typical, agglutination is effective when the informedness on the level of current patterns is weakened, and in reflection when there is reliable informedness only on the level of gender and aspect. If we start with the concept of the external determinant of languages we can manage to extract internal determinants and from internal determinants - when the material reserves and the system function is known -there result the particular features of the structure of the linguistic system on all its tiers: phonetics, phonology, morphology, syntax, and semantics [73; 87; 88; 90-96; 99; 101-103; 110; 125-127; 130; 169; 204; 205; 206].

As far as natural languages' external determinant distinctions are concerned, their effects are more evident when "normal" i.e. acoustic and articulatory languages are compared with languages which are visual and gestural for deaf and dumb people, and tactile and contactual for blind people.

In concluding this section we can emphasise that the ontological language scheme we have been examining enables us to look, from completely different angles from those accepted at present in cybernetics literature, at problems of meaningful and formal description of speech processes, and of the correlation between artificial and natural languages, and to evaluate the possibilities of perfecting machine translation and to determine the perspectives for effecting dialogue between man and machine in natural language.

further >>






| contents | | main page | | further >> |