Configuration as a method of structuring complex knowledge
METHODOLOGY AND LOGIC IN STRUCTUAL SYSTEM RESEARCH
1. In contemporary scientific knowledge the centre of the stage has been taken by the problems of structural-system research into phenomena. These kinds of problems can be seen clearly in biology and linguistics, sociology and physics, geography and criminology, mathematics and pedagogics, logic and psychology. Some of the more recently developed scientific disciplines have, from the beginning, formulated their subjects of research in systems, as specifically structural. Among these are cybernetics, semiotics and system-technology (research into and design of large systems).
However, attempts to bring out the structure of phenomena have run across serious difficulties. These difficulties appeared to have much in common despite their specificity. Consequently, specific directions of research developed; their goal being to determine and systematise that aggregate which would disclose the structural analyses made in the various fields. It was necessary that these researches yielded a certain apparatus of concepts and other means which would reveal the structures of phenomena more easily and would enable one to find descriptions expressing the specific systems of these phenomena. 'General System Theory', for instance, appeared in this way; the beginnings of which can be found in works of L. Bertalanffy. One can discover a similar direction of analysis in cybernetics. It is to cybernetics particularly that the term tern' is obliged for the extremely wide use at the present time. Nevertheless the repeated use of this term has so far not brought about essential progress in research into structural objects. Not going into a detailed analysis of existing ways of approaching these problems, we limit ourselves to the same starting point: the endeavour to uncover the so-called 'common signs', which would characterise all systems in general or at least sufficiently broad classes of systems. The central problem is, therefore, the reduction of a multitude of characteristics of systems, accumulated in various fields of knowledge, to smaller numbers of very general characteristics and laws (vide Studio Filozoficzne by Anatol Rapoport, 1963, Nl).
One characteristic of the search for 'common signs' amongst structures and systems, is an attempt to get away from the apparatus and methods of this or that concrete science Ц biology, mathematics, thermodynamics, or cybernetics. In connection with this, the problem of reduction of knowledge stands out as a problem of bringing together all the divers subjects of knowledge, and joining these with the apparatus and laws of the corresponding original scientific discipline which reason or other is acknowledged to be more general or more abstract.
This way of generalizing the analysis of structures and systems as a whole is transferred into a special scientific scheme, while the structure and systems themselves stand out, in point of fact, only as the new object of a corresponding discipline. This new object is, in principle, similar to previous objects of the discipline. The laws of that discipline must be extrapolated to the new object. Such extrapolation should then bring out the general characteristics of systems described in the language of any science. Such an approach into systems research appears to be in. essence purely naturo-philosophic, in as much as it leaves the researcher on the level of only objective descriptions and virtually closes the way to uncovering cognitive specificity in the formulation and solving of a problem.
2. But numerous facts (some of which we will examine later) show that the usual naturo-philosophic approaches and method reasoning cannot lead to fundamental progress where research into structural objects is concerned. In research of this kind, science runs into not only a totally new field of reality, but also a completely new and previously unknown cognitive situation. Evidence from scientific practice shows that situations cannot be 'solved' simply by the application of methods of research now available. With increasing acuteness and urgency, the development of systems research itself raises the question of the necessity to originate cognitive methods to solve its own special problems. Moreover, one can rightly assert that it is precisely the absence of such specific cognitive methods which, in present times, is the basic obstacle to the development of structural-system research and, consequently, is the basic problem. Researchers are beginning to realise this fact. Thus, the core of research into self-organizing systems (which has expanded on large scale in the last few years) is not in the descriptions of various systems as such, but the search for, or creation, of special cognitive constructions (mainly models), which could be the means for the analysis of self-organization (see, for instance Self-organizing systems published by MIR 1964 'Principles of Self-organization'. Transactions of Symposium, Pergamon Press L., N.Y., 1962).
However working out new means of research is the prerogative not of theory as such but of methodology and logic.
As a special and relatively isolated subdivision in the general system of theoretical research, logico-methodological analysis becomes possible only when comparatively extensive and diverse knowledge has already been accumulated about the object.(One can say that this analysis appears is response to the collection of diverse information, since one has to work with it in a special way). That specific subject, in which the logico-methodological analysis proceeds, is similarly defined: it is directed to this information by the interrelations; by problems of research and methods of research; and finally, by the relation of information, problems and procedures to the objectives.
In this way the logician and methodologist are interested not in the object as such, but in another, noticably more complicated formation; the system of interrelations between the object, problems of studying the object information about the object and the actions of the researcher. Moving in this topic, the logician and methodologist cannot construct the theory of the object but they can give recommendations for future procedures of research in that subject.
It is obvious that some methodological information is being made use of in any scientific investigation. But when concerned with comparatively simple subjects, methodological analysis is carried out, as a rule, intuitively, and the researcher usually has no chosen logical knowledge for that expressed purpose. But when research into complex and very complicated objects begins, the situation changes drastically. It is then not enough to have only intuitive methodological notions; the ideas must take the form of discreet and precisely stated knowledge. Therefore in structural-systems research, both specialised logical and more general logico-methodological analyses become an indispensable organic part of the process of research into an object.
THE PROBLEM OF SYNTHESIS OF KNOWLEDGE RELEVANT TO AN OBJECT
3. One of the most important problems of structural-system research, which demands specialised means obtainable only in logico-methodological analysis, is the synthesis of diverse knowledge (particular systems knowledge) about one object into a single theoretical system.
Each piece of knowledge about the object is worked out while a specific problem is being solved. It mirrors the object from one side, singling out in it a limited group of qualities, indispensable for the solution of the one practical problem. That which is important for solving one particular problem, is often found to be unimportant or insignificant for the solution of other problems. Therefore the emergence of new practical or theoretical problems firstly compels us to approach the object from new points of view, to distinguish new properties and form new bodies of knowledge; and secondly, questions about the relation of the newly arisen practical problems to already existing knowledge, force one to, clarify its possible use in solving these problems, or to obtain further information about the object.
With the accumulation of a substantial number of such onesided and particular pieces of knowledge, there arises a special theoretical problem of uniting them into one many-sided body of knowledge about the object. Solution of this problem has not only an abstract, but also a deeply practical meaning; it allows us to rationalize; to condense the accumulated knowledge, and thereby leads us to economy in working with it.
But how can one unite, in a single system, onesided pieces of knowledge about the object that allows us to work more efficiently together with the solving of particular problems?
Often they are connected, purely mechanically, with the conjunction, 'and'. Then the studied object emerges as the sum of those sides, or qualities, which have been singled out before. In a diagram the process can be illustrated thus:
The methodological base of such a notion (although it usually is not expressed in such an obvious form) consists in this: each of the properties bound up in the information is treated as a substantial part of the object while the actual structure of the object is understood as being 'composed out of these parts. Therefore the formalistic bonds, established in the scheme of knowledge, are simply transferred 'inside' the object itself and are represented as its 'structural' bonds.
It is in just this way that an attempt is usually made to build a theory of thinking as a whole. For suppose there is taken on one side the notions of thinking proper worked out in the history of logic and psychology, and on the other the notions formed in linguistics about language. 'Thinking' and 'language' in such an analysis are considered as two 'parts' of a single object, and the problem consists in determining the character of the objective bond between them. But in spite of a long history of research, the situation has not gone beyond the formulation of very general propositions, such as: language is not possible without thinking no thinking without language (for more details about this problem see G. P. Schedrovitsky, 'Verbal thinking and its analysis', Questions of linguistics, 1957, No.1)
This state is also generally characteristic of contemporary theoretical biology in its attempts to build an integrative theory of life. As we know, the structure of contemporary biology consists of a tremendous number of diverse disciplines, all of them investigating 'life'. But each conducts a one-sided research, with which it constructs a specific theory that is particular relative to an integrated theory of life. When this integrated whole is spoken of, it is represented as a sum total of parts, as uniting into oneness the different biological levels. Thus, the 'levels' themselves are understood purely ontologically Ц that is as levels of the biological reality itself, as its strata Ц and their logical nature is not taken into account. So far, this approach has not hindered the development of |research on any 'level', but it does not allow us to construct a satisfying theory of life.
Lastly we will consider semiotics. Efforts to construct a general 'theory of symbols became particularly intensive from the end of the nineteenth century, and up to the present time have engendered a series of different concepts; logical, logico-philosophical, logico-psychological, linguistic and psychological. But none of these succeeded to any degree in constructing a satisfying (i e. uncontradictory and sufficiently comprehensive) theory of symbols which would ensure a solution to extant present day practical problems. From our point of view such a result is quite natural. It ensues from the very character of the ennumerated concepts; all are analysing the symbol from one side only, or at best from two or three, and have no methods which enable one to examine the symbol as an integral formation. This is quite natural from the point of view of the history of research. Logic, psychology, linguistics, anthropology and other sciences consider the symbol not as an independent subject of research, but only as an outer material; or, at best, as an element of another subject of research: knowledge and science, processes of drawing conclusions and processes of thinking, the activity of an individual in solving a problem or communicating with other individuals.
In constructing each of these subjects it was enough to take into account only some, and not all sides of the symbol; accordingly, the methods of analysis characteristic to each of these sciences made it possible to understand only these separate sides of the symbol. They did not make it possible to reveal the symbol as a whole. For instance, when exposing the logical structures of reasoning it is necessary to examine the material of symbols in relation to the meaning residing in them, i.e. their objective content; it is not necessary to take into account the relation of this material to the activities of the individual and their development. From another point of view many regularities of the speaking activity of individuals can be determined without reference to an analysis and description of the content and meaning of symbols.
But when the problem of constructing a general theory of the symbol arose, such methods 'failed'. Indeed, because of its objective nature, the symbol can be singled out as an independent subject of study only when it is considered in the entirety of all its fundamental functions. The main point is this: the origin and development of semiotics, from the first ideas put forward by Lock, Leibniz and Kondiljak, was dictated by the necessity to overcome the one-sidedness characteristic of logic, psychology or linguistics in their approach to the analysis of a symbol, and to synthesise the methods of all these sciences. However, this tendency was subject to a completely distorted presentation: first of all in linguistic researches, but also, to a somewhat lesser degree, in logic and psychology. In each of these disciplines semiotics was conceived of as a simple broadening of the corresponding science; as an application of its concepts and methods to a new field of objects (see, in particular, 'Symposium of structural study of symbolic systems', Theses of reports, M., 1962). As a result, logical, linguistic and psychological approaches to the elaboration of semiotics formed independently, isolated from each other, and each of them strove to embrace the whole field of symbolic reality. (See G. P. Schedrovitsky, 'Of the method of semiotics research of symbol systems. Semiotics and the Eastern languages, M., 1967.)
And so the picture becomes the same as in the other examples we have cited; separate, particular pieces of knowledge about an object being used purely mechanically to bring out a general theoretical system taking the pieces of knowledge to be parts of the object itself. Therefore the structure of the object, in the final summary, is always regarded as isomorphous with that system of knowledge which can be obtained by the direct unification of already existing pieces of knowledge, obtained independently from each other.
4. However, taking into account the specific nature of knowledge and its relation to the object of study, then one approach to the problem of how to synthesize separate pieces of knowledge which are all related to one object, can, and in our view must, be completely different.
The main point is that abstractions from afar do not always present distinguishing aspects of the object of study. As a rule these form differently. The knowledge arrived at through the solution of particular practical problems can be made use of as projections which are 'taken' from the object at different 'angles'. This is shown in Diagram 2: the shaded circle is the object, the lines (A), (B), (C) represent knowledge which fixes the sides of the object. If such a representation of abstraction is fair (there is no fundamental objection to be seen against it) and existing knowledge presents projections taken of the object, clearly a purely mechanical unification of these projections cannot give a representation of the real structure of the object. Such attempts to unify 'views' of the object obtained in such a manner are as unpromising as the attempts to get an idea about the structure of a detail by simply joining two of its graphic projections.
But how, in such a case, is one to achieve a synthesis of the different, one-sided aspects of an object?
5. The basic methodological approach to this problem demands first of all precise differentiation of concepts about the object and subject of research. Such differentiation is a matter of principle in any methodological analysis, while in the methodology of system research it has a special significance, being in essence, the starting point of all work (see G. P. Schedrovitsky. 'Problems of methodology in system research', published by Znanie, 1964. 'Methodological problems of system research', General Systems, 1966).
The concept of 'object' is very complicated and, at the present time, is used very loosely. Essentially different from each other are:
(1) Object of operation
(2) Object implied (that to which symbols refer)
(2) Object of study;
by another principle one must also distinguish the opposition between
(a) real object of operation, and
(b) ideal object.
In the context of this article it will be enough for us to define the object of study as the supposed reality, which is studied, assimilated, and which always differs from existing, historically organic relative knowledge in the present moment.
The subject of research, on the contrary, is formed by research; this reality, created by science itself, exists only as much as the knowledge about the object exists. In beginning the study of some object we consider it from one or several sides. The selected aspect becomes 'deputy' for the whole object. In as much as it is knowledge of the objectively existing it is always objectified by us and as such forms a subject of science.* If we take advantage of the presentation in Diagram 2, then the subject and its relation to object can be presented as in Diagram 3:
As a whole it is the subject. The circle represents the object, and its shaded part represents the 'side' of the object singled out in the research. Line (A) represents knowledge in connection with this 'side'.
In specialized scientific research the subject of knowledge is looked at as being 'adequate' to the object. This is quite right and is substantiated so long as the movement of research is inside the frame of a given subject. But when there is a build-up of several different subjects of research with regard to one object (as shown in Diagram 2) or when this object appears in the subjects of different sciences, such an approach becomes a hindrance to the synthesis of these different subjects, creating paradoxes or contradictions in the development of knowledge. The only means of avoiding the latter is a logical analysis which examines the subject of research as a result or product of activity, or as a product of human reflection, which is unrelated to the object and not as reducible to the same; existing within special methods of science, as a special creation of human society, subordinate to a special conformity with the laws of life and not coincident with the law-governed life of the object itself.
The character of the subject depends not only on which object it reflects, but also on what, for this subject, has been formed for the solution of what problem. The problem of research, and the object, are the two factors which define (not deterministically, but teleologically) how, with the help of the ways and methods of research, the subject indispensable for solution of a given problem will be formed. The indicated peculiarity of the formation of subjects of study and their corresponding knowledge about objects leads us to see that systems of lawful representation strongly differ from the reality of structural objects. This divergence should not be taken as some abnormal, inadmissable occurrence. On the contrary; any formal system of object representation is a special operative system in which and with which one acts quite differently from the way one would act with the object itself. As is known, operative systems are created with this in mind, and in such a way that working with them should differ substantially from direct operation with the object. Therefore, in principle, we cannot and must not strive to force the system representation to coincide with structures of the object. Obviously the direct opposite is needed; that this non-coincidence should be realized as a principle and that from this the solution of methodological problems should originate.
Projections in diagrams are not the presentation of parts or details, but this does not hinder their use in the least because special procedures exist which allow one to pass from one projection to another, for instance, to an axonometrical projection or from the projection to the detail itself in the process of its preparation. The same can be said about different types of radio-technical schemes (block schemes, principle schemes, installation schemes) , where one is given rigid rules on the transition from one scheme to another. Therefore the main thing is that there should be procedures of transition between different operational structures, which will stand simultaneously for the existence of connections between them.
6. It is not difficult to notice that these procedures can exist and. 'work' only when one has matching, especially for the adapted projections; far from it, that one can establish procedures for finding connections between any projections taken at random. Consequently, any method of synthesising abstractions proves to be rigidly connected with a specific method of obtaining them. We can pass from one set of graphic projections to others and 'build' the object from projections just because these projections themselves were obtained in a special way, this being a| necessary preliminary to the connecting procedures. In another way it can be said that procedures of abstraction and procedures of synthesis of the knowledge obtained by their means must be organically connected with each other, and must form a single cognisable mechanism.
7. The above principle can be put into a formula and be adapted; to any theoretical knowledge and ideas which we wish to unite. Taking this into account, we must first of all come to the conclusion that by itself, the fact of the presence of several theoretical ideas which have been obtained independently from each other while solving various problems does not yet give sufficient grounds for formulating a question about the possible connection between these representations.
This can be explained in terms of the drawing showing multi-subject research into an object (Diagram 2). Let us suppose that the projections; (A), (B), (C) and so on were 'taken' from the object without any strict rules, being determined by the 'nature' of the object and the procedures of the ensuing synthesis of resulting projections (on the whole such a. situation is fairly typical to a number of branches of contemporary science). In the presence of these conditions some parts and elements of the object will be reflected several times in different projections and consequently, in a different context. This will lead to a 'doubling of essences' confusing the object's characteristics in as much as the same 'points' will be presented in different ways, with varying functional emphases. On the other hand, some elements and sides will not be reproduced at all and this will create an important 'void' in our representations.
It is obvious that if one uses this type of analysis and description no procedure of unification can give the necessary results. Something of this kind happened in pre-Marxist political economy, when V. Petty, A. Smith, D. Ricardo and other researchers tried to construct a general economic theory mechanically connecting the existing categories, such as goods, labour, capital, cost and so on. All attempts to unite these categories came to nothing, and could not lead to anything, because the definite aspects of economy which were objects of research 'passed' through several concepts and descriptions, and other important aspects were not 'grasped' at all. In that state, as they existed before K. Marx, these concepts and ideas could not be brought into a single system, because they were developed without reference to the problem of synthesis. In contrast, the success of Marx's analysis can be explained from the methodological point of view by his construction, in principle, of other points of origin, which, from the beginning, allowed him to develop the structure of the subject of research and in this context present his concepts and categories as a system. The methodological base of this research was, as we know, the method of ascent from the abstract to the concrete, in which the means of forming abstractions and the means of forming their synthesis in the process of ascent were combined in a single knot. (See A. A. Zinoviev, 'Ascent of the abstract to the concrete' on the material of Capital by K. Marx, Thesis, M.G.U., 1954, and by the same author 'On the problem of abstract and concrete understanding' Filosoficky Casopis, 1958, č2.)
In fact, this example contains the answer to the question of how the synthesis of the different theoretical concepts must be realized if these concepts are received 'chaotically' Ц i.e. not because of each other and without any orientation to the ensuing synthesis. Obviously in such a situation the first step should consist in reconstructing the initial concepts themselves, freeing them from similar or repetitive content, and supplementing them with other ideas which prove necessary from the point of view of synthesis, etc.
An attempt to follow such a cycle collides at once with an obvious paradox; for the initial abstractions to really form a system and tie up with the problem of synthesis, the researcher must have a representation of the actual structure of the object in the initial set up and align all existing one-sided projections with that representation. In other words, the construction of a theory of a structural object presupposes the availability of a representation about the structure of the object. But this is really so, and it is not possible to solve the problem by any other means. The point is that the initial structural representation is not yet the theory of the structural object; it lies in a special plain of knowledge about the object, i.e. methodological, and fulfills a special methodological function in the process of research, being only a means for the construction of a theory.
In the plan of analysis this conclusion signifies a great deal. It sets in motion that line of study which must be realized if one is to achieve a synthesis of existing knowledge about the object. First of all the conclusion underlines the fact that it is not possible to resolve this problem while remaining in the plain of existing knowledge alone. It shows that analysis of the abstractions Ц and, taking a broader view, of all the procedures Ц by which the existing knowledge was obtained, must enter into the motion. It shows also, that it will be necessary (and this is an indispensable condition of the preceding demand) to perform special work of reconstructing the structure of that object of which the already available knowledge is a projection.
It follows that to re-create the structure of an objectї means to construct a special new subject, and in constructing it to bear in mind the strictly determined specific problem, i.e. to bring about the synthesis of available knowledge. The cycle of research can be shown in a diagram thus:
Here K represents a new lawful form which reproduces the structure of the object. The group of continuous arrows shows theoretico- methodological movements in coming from existing knowledge (A)(B) (C); while the discontinuous lines indicate that the available pieces of knowledge (A), (B), (C) are 'projections' of the object, i.e. its new presentation.
The diagram visually demonstrates that in solving the problem of synthesizing different knowledge about one object, it is necessary instead of searching for some common bonds in the plain of that knowledge, to reproduce the structure of the object, and then proceed from it to reestablish those 'twists' of abstraction, which have lead to the available knowledge. Only in this way can one get the necessary connection between different representations of the same object.
8. But what does it mean to reproduce the structure of the object in a higher form than available knowledge about it, and thereby supplement this knowledge? In our view it means the introduction into the system of the aggregate of knowledge of a structural model of the object.
Models and their specific functions in the development of knowledge require specialized methodological analysis. In this instance we refer to a special model with a special function in the system of a developing theory; it is a representation of the object created especially for the purpose of unifying existing knowledge. (Hence, one can say that it is precisely the accumulation of unifying knowledge that in reality sets and defines the character of the model to be introduced.) This model simultaneously explains existing knowledge and serves (together with the logical description of derived abstractions) as its substantiation.
We will call representations of the object, created with the indicated purpose in view, 'configurations' (compare with B. A. Lefevre and his idea for means of presenting the objects as systems, in the book Theses of reports from symposium. 'Logic of the scientific research' and 'Seminar of logicians',
From this point of view, if one examines the researches already mentioned in the fields of biology, semiotics, theories of thinking and other sciences, one can affirm that their development at present turns on the absence of adequate structural models for the objects studied, i.e. absence of 'configurations'.
Indisputably, attempts to construct such models are being undertaken, with increased intensity. But just because of the absence of an evolved method, from the beginning the researcher not infrequently chooses a methodological approach with no prospects; instead of constructing a configuration which does logically base and explain the existing diversity of knowledge, he accepts one of the available system-representations-of-the-object as an initial position and thus closes the door when it comes to revealing the actual structure of the object.
But, in many fields of knowledge there are possibilities for the construction of effective structural models of objects under research. Models of that type underlie cybernetics and system-techniques and they have played a big rôle in the development of present day genetics, though the methodological part of the work in these fields has by no means been fully realized.
A long term plan, now listed in our programme of work, presents a structural model of education. Pedagogical researches, as conducted up to now, are not capable of exerting a decisive influence in overcoming the increasing gap between the system of education and the demands of science, production, and those leaving school. This suggests that the sphere of pedagogical research is too narrow. Indeed, these researches are conducted almost exclusively as psychologico-pedagogy, and it is difficult to discover a common system of theory even in this one area. But education is in itself a complex structural object. It certainly is not limited to problems of the psychological development of an individual. It is absolutely necessary to widen the subject of pedagogical research; It must embrace the problems of aims in teaching and education, the shaping of the man of the future (i.e. that 'product' which is needed by society in 10 to 20 years' time), dynamics of small groups, content of teaching problems of discipline and development, rôle of the teacher in teaching and education, stimuli of the educational activity and so on. To embrace this very wide range of problems it is necessary to synthesize the ideas and methods of a number of scientific disciplines; sociology, logic, psychology, linguistics, ethics and aesthetics. But such a synthesis perforce brings one again to the construction of a configuration (see Questions of philosophy, 1964, No.7).
Semiotics faces analogous methodological problems. In our opinion semiotics lacks first and foremost an initial structural model of sytobols as a whole object which would unite all the ideas about the symbol worked out in logic, psychology, linguistics, sociology and ethics (see Semiotics and Eastern Languages, M. 1967).
Referring to biology: it is our opinion that only through the construction of a special configuration can the problem of explaining 'life' be solved. It stands out in the same way as the problem of synthesizing the different levels of description of biological reality. For instance, it is understood in just this way by A. St. Derdi (see his book Introduction to sub-molecular biology' Ed. Nauka, 1964, p. 16). In recent times many biologists have come to the conclusion that an explanation of the essence of life which takes as the initial point one of the existing levels of description, will hardly succeed. Generally speaking, one can say that the structural model is indispensable to the biologist; this model must arise, speaking figuratively, as 'perpendicular' in relation to the existing levels of description. This means that a logical analysis of the origin and development of each of the existing levels is indispensable for the construction of the model. Such an analysis becomes the premise for creating a configuration, which will serve as an initial point in the construction of a general theory of life. It would also be possible to show the meaning of the construction of a configuration in some other branches of contemporary knowledge, for instance, in mathematics, in the research and planning of large systems, in criminology, linguistics, economico-mathematical modelling and so on. But this requires specialized discussion. Right now it is important to underline the organic need for the use of logico-methodology in the research into structures of complex objects, concerning which different system representations are constructed, with specific ways of subdividing the object, with specific units of analysis in each such representation. It would be an oversimplification to imply that, by itself, the idea of a configuration model could solve all questions connected with creating a structural-system theory. The construction of a model and the ensuing research need to overcome considerable difficulties, but, in our opinion, the construction of the model makes the work more hopeful.
9. Correlating the existing information about an object with a newly formed configuration often leads to a substantial reconstruction of the configuration. Such a reconstruction constitutes one of the most important aims of methodological work, for it makes possible an immediate unification of existing concepts within the domain of the initial description. This is shown visually in Diagram 5. The double arrow depicts logico-methodological reconstruction of existing pieces of knowledge about the object, while the line A'B'C shows their synthesized system. Such a unification has an indisputably practical value; it brings about a peculiar 'flattening' of different system representations arranged on different levels that cannot be brought directly together in a single 'linear' theory of the object. This considerably facilitates operations with the system of know-ledge and, in particular, is an indispensable condition of its formalization.
10. When the unification of knowledge is accomplished configurations become unnecessary and can be omitted in the system theory. But usually the configurations are kept, and moreover, begin to live and develop by their own 'logic', sometimes even becoming a special stratum of the theory. The explanation of this is that the models, constructed with the aim of synthesizing existing knowledge about the object, can be, and often are made use of in other methodological functions; especially as models of an object which allow the means of solving emerging practical and theoretical problems to be outlined. The most important specific problem of methodological analysis is the formulation of recommendations concerning procedures of research and the description of an object Here the methodologist proceeds from the practical and theoretical problems which may arise or are contemplated, to answer the question of which subjects of study need to be formed and how they should be used, for the solution of these problems. Speaking figuratively even before the beginning of a special research into an object, the methodologist must build a plan of this research; lay down groups of assembly, subsections and decide on the method of work to be used in each of them. He must move in a special, methodological, stratum of knowledge. The picture looks like this; it is as if we start constructing the whole building of information from the 'top floor', and then 'suspend' all the rest from it, including the foundation, i.e. the theory of the object proper. This is shown visually in Diagram 6. D and E show the new specific information about the object already obtained through the methodological plan that has been compiled.
Such a description made in advance of the necessary procedures of research into the object, is made possible because methodological analysis always has a greater togetherness than any corresponding specific analysis. It transfers concepts, principles and schemes of subdivision from one explored region into other not yet known. Thereby the methodologist appeals to general logical principles and concepts, in which the whole experience of our research activity is accumulated, as it is in the specialized ones, but more so in general subjective knowledge. But always both will stand out as methodological in relation to the object under direct study and not as specialized knowledge of a subject. An essential element of such work is the confrontation of the existing model of the studied object, with the models of objects already mastered, and the new problems arising with regard to this object, with those problems which were determined for other objects (for this, of course ,the problems themselves must be presented in sufficiently subdivided states). New subdivisions are introduced on the grounds of such confrontations of the object under study, while its model is continually built up and subjected to further extrapolation. If the results of such a confrontation point to similarities in problems and objects Ц those previously learnt and those being studied Ц then we can transfer those schemes of subdivision and analysis which have been worked out or applied while working with the mastered object to the new object. The results of this transfer are exactly those which are established in the methodological plan of research.
Like the structural model of the object, the plan presents one subject with its own special principles and methods of construction and extrapolation. This is, in point of fact, also a configuration, but with a slightly different function and different structure, than the configuration or structural model. This is a distinctive block-scheme of the object; it fixes all those subjects of study, which must be formed to make possible the solution of the set problem and, what is most important, determines the sequence of their examination. If the configuration model must represent the object as a single whole, irrespective of the different problems of its study, then the configuration of the second kind Ц the plan Ц must, on the contrary, present the object divided into a number of subjects, whereby the division itself and the method of binding the elements are determined by problems which must in this instance, be solved. In this sense the plan is not so much an 'image' of the object, as a scheme or 'rule' which regulates the activity of the researcher.
It is important to emphasize that these two types of configuration are closely related. The structural model is constructed on the basis of particular subjects of research into the given object and is a distinctive 'metasubject' in relation to them, embracing the subject as a whole. In as much as the object cannot be studied in 'one dimension', it must be subdivided into separate elements. This new subdivision, based on a specifically constructed correlation between the structural model and the emerging problems, is brought out in the plan. Therefore, subdivision in the plan does not reproduce the former division into subjects of study. On the contrary, it must be new, obeying the logic of the metasubject's unfolding on the grounds of new problems of research.
The plan, which determines the subjects of study and procedures of analysis, is an extremely effective methodological means which allows research to be organized in a definite direction, and coordinates activity towards a single aim. The need for coordination, especially in the complex syntheses of sciences, is confirmed by many examples; in particular by experiences in pedagogics and the new rapidly developing discipline of 'systemotechniques'.
Another interesting example, which illustrates the importance of such methodological development in research into complex objects, is the 'morphological method', created by the well-known Swiss astronomer
F. Zwizzi (working in
The plan of anticipated research is a secondary product of methodological analysis, setting a general scheme of the theory under construction. On this basis the third part of the methodological work can then unfold, which is directed to the clarification of procedures of research and the description of each selected subject. But this constitutes a subject of a special analysis and is outside the scope of the theme of our discussion.
The successful development of structural-system research depends to a considerable degree on the elucidation of specific cognizable means. The problem of constructing such means can be solved only by methodological analysis. It has its own specific apparatus of concepts supported by the general logical theory of thought, and its own lawful means for the representation of objects of knowledge (examples of some representations of this kind have been cited). The application of these means will make possible the construction of lawful abstract objects, appearing as models of real objects and the construction of schemes of reasoning which could not be obtained by the analysis and description of direct empirical manifestations of the object of research. One of such specifically methodological means of research is the 'configuration'.
Owing to the postal strike final corrections could not be made to this article. We hope to amend errors in the June issue.
* At present we leave aside the purely social aspects of the problem, which have exceptionally important significance in forming the subject of science this side demands special analysis, while here it is enough for us to give only the logical characteristics of the process.