In his standard textbook The Conditions of Learning,  Robert Gagné (1970) identifies eight hierarchically organized types of learning. The highest, cognitively most advanced type is called problem solving. In problem solving, "two or more previously acquired rules are somehow combined to produce a new capability that can be shown to depend on a 'higher-order' rule" (Gagné 1970, 64). Problem solving is dependent "on the store of rules  the individual has available" (Gagné 1970, 223).

Although Gagné's position was first presented quite a while ago, it has not really been surpassed or superseded by more recent theorizing within cognitive psychology. For example, Donald Norman in his textbook Learning and Memory  (1982) identifies three basic types of learning: accretion, structuring, and tuning. His structuring is a fairly close counterpart of Gagné's problem solving. It implies the formation of a new conceptual structure or schema on the basis of previously acquired knowledge and experience. As a typical example, Norman reports his own learning of the Morse code. Having trained himself a long time to receive individual letters in the Morse code, not improving noticeably in speed, he was adviced to focus on words and phrases instead of letters. A dramatic improvement occurred.

 "I already had a solid base of performance on the individual letters, and so I was able to benefit from the advice to enlarge the unit size - to restructure my knowledge." (Norman 1982, 83.) 

The similarity between Norman's structuring and Gagné's problem solving is obvious. The jargon has changed, but the substance remains the same.

At the first sight, problem solving or structuring seem to be satisfactory characterizations of the uppermost reaches of human learning. What more can one expect than insightful solutions to problems through a novel structuring of the subject's mental model or cognitive schema?

The problem is that problem solving and structuring are essentially reactive forms of learning.  Both presuppose a given context which presents the individual with a preset learning task. Learning is defined so as to exclude the possibility of finding or creating new contexts. However, it is this  very aspect of human performance - or rather the lack of it - that is becoming the central source of uneasiness and trouble in various fields of societal practice. In general terms, troubles of this type may be  named the difficulty of anticipating, mastering and steering qualitative changes in individual lives, in families and organizations, and in the society as a whole.

Symptomatically enough, Norman ends his book with a tirade on how badly modern technology matches human capabilities. According to him, system designers misuse and ignore the users: "they start with the machine, and the human is not thought of until the end, when it's too late: witness the control panels in the nuclear power plants" (Norman 1982, 115). Norman's solution is: techonological systems should be designed so as to make learning easier.

Pleas like this follow the traditional patronizing approach: the poor learners must be helped to cope with the tasks given  to them. The approach is self-defeating. Norman himself points out that it takes a long time to learn the mastery of a complex skill. At the same time, the contexts of the tasks and skills are going through profound qualitative changes which often render previous tasks and skills obsolete. Norman himself says 'when it's too late'. This lag can never be overcome by patronizing, by asking designers to plan more 'user-friendly' systems. It can only be overcome by enabling the users themselves to plan and bring about the qualitative changes (including the design and implementation of technologies) in their life contexts. 

If learning has nothing to offer in this respect, we have good reason to talk about the futility of learning. Both in theory and in practice, human learning actually seems to be doomed to the role of running after those qualitative changes in people's life contexts. While the learners are engaged in diligent problem solving and structuring in order to cope with changes that have shaken their lives,  there are already new qualitative changes quickly getting ripe to fall upon them. This stance is documented by Gagné as follows. 

"A great scientific discovery or a great work of art is surely the result of problem-solving activity. (...) Nothing (...) supports the idea that there is anything very different about the problem solving that leads to discoveries of great social import. (...)  But the major discovery, in contrast to the common garden variety, involves a feat  of generalizing that goes far beyond what may be expected in the usual learning situation. There is an 'inductive leap,' a combining of ideas that come from widely separated knowledge systems, a bold use of analogy that transcends what is usually meant by generalizing within a class of problem situations." (Gagné 1970, 227-228.)

Here we have two assertions. Firstly, great creative achievements are based on the same kind of inductive, combinatorial problem solving as any common act of learning by problem solving. Secondly, usual acts of learning by problem solving have practically nothing in common with  truly creative discoveries because in the latter the 'inductive leap' is so much greater. In other words, Gagné first denies that creation has anything qualitatively special in it. Immediately thereafter he points out that creation is indeed qualitatively special because it transcends the context given.

The outcome is rather gloomy for learning. 

"(...) because it is a method rich in reinforcement value, the solving of problems within structures of intellectual skills to be learned may create a love of learning, a 'thirst for knowledge' in the individual learner. But it is a vastly different thing to suppose that this kind of learning will necessarily predispose the individual to become a 'creative' thinker, capable of making great contributions to science or art. To be sure, the variables that produce genius are surely not entirely innate and must prominently include factors in the individual's experience, arising from his environment. But except as a method for acquiring prerequisite intellectual skills, 'practicing discovery' seems an unlikely choice of antecedent variable to be involved in the production of genius." (Gagné 1970, 229.)

This is a specimen of self-defeating circular reasoning. First the author tacitly assumes that the highest form of learning is practicing inductive combinatorial problem solving which by definition does not transcend the context given. Then the author triumphantly concludes that learning by problem solving does not lead to true creativity, i.e., to transceding given contexts.

In this book, I shall examine whether learning really is doomed to futility or whether this is an historical artifact of only limited and temporary validity, both in theories of learning and in the societal practices involving learning.

More specifically, I shall argue (a) that the conception of creation as inductive combinatorial generalization (albeit in magnified scale) is fundamentally false; and (b) that the conception of the highest form of learning as inductive combinatorial problem solving or structuring is also fundamentally false.


The alternative to reactive forms of learning is expansion which transcends the context given.  Because of its elusiveness, expansion is traditionally not considered a proper object of scientific investigation. It has very much remained  a domain of mysticism.

C. G. Jung made one of the important early attempts to incorporate expansion into psychological theory. For him, the key concept was the collective unconscious.

 "From this point of view the conscious personality is a more or less arbitrary segment of the collective psyche. It consists in a sum of psychic facts that are felt to be personal. The attribute  'personal' means: pertaining exclusively to this particular person. A consciousness that is purely personal stresses its proprietary and original right to its contents with certain anxiety, and in this way seeks to create a whole. But all those contents that refuse to fit into this whole are either overlooked and forgotten or repressed and denied. This is one way of educating oneself, but it is too arbitrary and too much of a violation. (...) Hence these purely 'personal' people are always very sensitive, for something may easily happen that will bring into consciousness an unwelcome portion of their real ('individual') character." (Jung 1966, 157.)

According to Jung, psychoanalysis may lead to annexing deeper layers of the collective unconscious which produces an enlargement of the personality leading to the pathological state of 'inflation'. 

"It occurs whenever people are overpowered by knowledge or by some new realization. 'Knowledge puffeth up,' Paul writes to the Corinthians, for the new knowledge has turned the heads of many, as indeed constantly happens. The inflation has nothing to do with the kind  of knowledge, but simply and solely with the fact that any new knowledge can so seize hold of a weak head that he no longer sees and hears anything else. He is hypnotized by it, and instantly believes he has solved the riddle of the universe. But that is equivalent to almighty self-conceit. This process is such a general reaction that, in Genesis 2:17, eating of the tree of knowledge is represented as a deadly sin." (Jung 1966, 156.)

 On the other hand, expansion may lead to self-knowledge and truly widened consciousness. 

"(...) the more we become conscious of ourselves through self-knowledge, and act accordingly, the more the layer of the personal unconscious that is superimposed on the collective unconscious will be diminished. In this way there arises a consciousness which is no longer imprisoned in the petty, oversensitive, personal world of the ego, but participates freely in the wider world of objective interests. This widened consciousness is no longer that touchy, egotistical bundle of personal wishes, fears, hopes, and ambitions which always has to be compensated or corrected by unconscious counter-tendencies; instead, it is a function of relationship to the world of objects, bringing the individual into absolute, binding, and indissoluble communion with the world at large. The complications arising at this stage are no longer egotistic wish-conflicts, but difficulties that concern others as much as oneself." (Jung 1966, 178.)

For Jung, expansion is achieved through the collective unconscious, which in turn is reached with the help of psychoanalytic therapy. The conception is somehow very static: the collective unconscious resides  somewhere deep beneath more superficial layers. The task is to get into touch with it, to seize some of its immense power. But how did the collective unconscious emerge in the first place? How does it develop? Can the individual participate in creating new forms of the collective unconscious? And above all: Is the collective unconscious only a mental, spiritual layer or does it have some kind of material basis and embodiments in people's societal and productive practice?

As long as these questions remain unasked and unanswered, the Jungian theory remains mystical.

In recent psychological theorizing, some attempts have been made to reintroduce expansion as a scientific concept. In his 'transgressive model of man',  Jozef Kozielecki (1986) distinguishes between protective and transgressive behavior. The latter "allows for moving forward: the person is capable of exceeding the boundaries of his or her material or symbolic achievement, that is, capable of creating or assimilating new values" (Kozielecki 1986, 90). Transgressive behavior is further divided into two types, expansion and creation. The former consists in the acquisition and assimilation of existing material or symbolic values (commodities, business, power, influence, knowledge). The latter entails the solution of new, unconventional problems.

Kozielecki gets into trouble when he tries to apply these distinctions in concrete cases.

"There should be no difficulty in classifying Columbus's voyage or Einstein's discoveries as typical instances of transgressive behavior. We are apt to hesitate, however, when asked to decide if the solving of the Missionaries and Cannibals puzzle is a case of transgression or not. Similar problems in classification crop up in every other domain of psychology, of course." (Kozielecki 1986, 92.)

To avoid such difficulties, Kozielecki puts forward a definition as broad as possible.

 "Any intentional action whose outcome transgresses the subject's past achievements is seen as a case of transgressive behavior." (Kozielecki 1986, 92.)

In other words, if the subject could not previously solve the Missionaries and Cannibals problem -  and then finally solves it - this should obviously be accepted as a case of transgression. In effect, there is no clear difference between any kind of problem solving or structuring and transgression. The difference between a problem and the context producing the problem is blurred - or rather, contexts are not considered.  Notice that Kozielecki speaks of transgression only in terms of an intentional and individual-psychological process, as 'exceeding the boundaries of his  or her  achievement'. Jung's powerful though opaque idea of the collective  and often not very intentional character of expansion is given up without discussion. Notice also the circularity of Kozielecki's definition: what transgresses is transgression. Very little explanatory power is left in our hands.

Another recent attempt is provided by Karsten Hundeide (1985). His key concept is perspective.  Using a spatial metaphor, Hundeide introduces a general theoretical idea of two developmental principles, expansion and contraction. When one is located in a definite position, there are certain things one can see directly. They occupy a central position in the field of vision. Other things are in the periphery, and still others are outside one's field of vision or perspective.

Correspondingly, when one is in a definite interpretive position, there are certain conclusions, judgments, and insights that can be immediately seen as plausible and evident. Others are impossible, irrelevant or implausible. Thus, in order to arrive at a definite conclusion or insight, one must be in the right position. If one is in a 'false position' in relation to a certain conclusion or insight, there is little point in elaborating alternatives from that position. Instead, one must redefine the situation or 'restructure the field,' as Gestalt psychologists put it. Such a redefinition of one's position may be of an expansive character.

 "This expansion may result from a confrontation  between positions, between the recurrent alternative one takes for granted  and a contrasting alternative.  In order to solve this conflict, the person may have to 'move back' to the more detached and abstract position (...). From this position both conflicting perspectives may be integrated and united.

 (...) There is also the opposite movement (...). I call this the contraction of perspective.  This term was chosen because it is a movement from a wider more inclusive position to a narrower one with fewer options. Contraction of perspective may take place under conditions of monotony, reduced variation, or the absence of contrasting alternatives." (Hundeide 1985, 314-315.)

Hundeide is very conscious of the difference between problem and context. He also recognizes a specific type of problems, namely conflicts or contradictions, as the source of expansive recontextualization. However, his expansive recontextualization suffers from the same weakness as Kozielecki's whole conception. It is reduced to an individual and mental process. Thus, it is onesidedly attributed the flavor of abstraction and detachment. Jung's insight into the collective nature of expansion effectively counteracts this type of cognitivist impoverishment of human development. 

"(...) the collective dream has a feeling of importance about it that impels communication. It springs from a conflict of relationship and must therefore be built into our conscious relations, because it compensates these and not just some inner personal quirk.

 The processes of the collective unconscious are concerned not only with the more or less personal relations of an individual to his family or to a wider social group, but with his relations to society and to the human community in general. The more general and impersonal the condition that releases the unconscious reaction, the more significant, bizarre, and overwhelming will be the compensatory manifestation. It impels not just private communication, but drives people to revelations and confessions, and even to a dramatic representation of their fantasies." (Jung 1966, 178-179.)

So Jung sees new kinds of communication as necessarily involved in expansion. But are only cognition and communication reorganized? Does the material practice remain intact?

In this book, I shall argue that it does not. To the contrary, true expansion is always both internal and external, both mental and material. More specifically, I shall argue (a) that expansive processes can indeed be analyzed and modelled; (b) that the gateway to understanding expansion is neither the concept of collective unconscious nor that of perspective but the concept of activity;  (c) that expansive processes are becoming integrated into processes of learning, i.e., that a historically new advanced type of learning - learning by expanding - is currently emerging in various fields of societal practice. 


This book is a report of extended theoretical research. For many people, theory construction is either inductive generalization from so called empirical facts or  purely speculative reasoning. In my view, theoretical research in its mature form is neither one nor a combination of these two.

I agree with Klaus Holzkamp's (1983) characterization of theoretical research. He differentiates between what he calls the level of categories and the level of specific theories. Categories are basic concepts with which the scientific paradigm or school defines its object, its inner structure and boundaries. Such categories "always include certain methodological  conceptions about how one shall proceed scientifically in order to grasp the object adequately" (Holzkamp 1983, 27-28). The research reported in this book belongs to the level of category construction.

 "Whereas the construction of categories as basic theoretical concepts may be regarded from a bourgeois point of view mainly as a question of arbitrary definitions and conceptual fixations, the 'historical'  category  analysis  we are proposing is a procedure based on empirical  material (...) in which scientific rationality  is extended to a problem field which used to be closed to it: the formation of basic psychological concepts.  The methodological difference between research on the level of specific theories and research on the level of analysis of categories is thus not that the former is 'empirical' but the latter 'speculative', merely 'deductive', or the like.  To the contrary, both research types are empirical,  but the material collected and used is in the first case of an 'actual-empirical' and in the second case of an 'historical-empirical' nature." (Holzkamp 1983, 50.)

So the research reported in this book is theoretical research aimed at the construction of categories, using a specific type of empirical data. This specific type of data typically consists of propositions and findings of previous analyses,  or more generally,  of previous representations of the object of research.

Such data may be predominantly either object-historical or theory-historical. Object-historical data consists of propositions and findings describing the development of the object of the research - in this book, the historical development of human learning and expansion. Theory-historical data consists of theories or theoretical propositions concerning the object, considered in their historical origination and succession - in this book, theories related to human learning and expansion.

In the construction of categories, also actual-empirical data is often useful and necessary. But here Holzkamp's distinction between the level of category construction and the level of constructing specific theories is essential. In research aimed at a specific theory, actual-empirical data is an indispensable and integral element of the research project. In research aimed at category formation for an entire paradigmatic orientation, actual-empirical data may play a suspended and more mediated role, as if gradually growing into (and simultaneously altering) the suggested categories from various concrete projects.

In any theoretical investigation moving on the level of categories, three methodological questions must be implicitly or explicitly answered. These three questions are: (1) how to select  the data; (2) how to process  the data into categories; (3) how to bring the categories developed into fruitful contact with practice.

In the following sections, I shall address these three questions, using two very different examples of theoretical research as points of comparison. The first example is the short but pathbreaking paper Toward a Theory of Schizophrenia  (Bateson 1972, 201-227), written by  Gregory Bateson, Don Jackson, Jay Haley, and John Weakland in 1956. The second example is the much discussed two-volume work The Theory of Communicative Action  by Jürgen Habermas (1981; in English 1984  [Volume 1]).

Incidentally, both examples are concerned with the theme of communication. However, the paper by Bateson & al. is aimed at a reconceptualization of the theory of schizophrenia, while Habermas's book aims at formulating a comprehensive theory of communicative action in general. It may look as if the paper by Bateson & al. would be quite specific and not belong to the level of category construction at all. However, its theoretical kernel, the single central category generated by the authors in that paper, has had an impact that by far exceeds the limits of a specific sub-theory. It has been instrumental in the reorientation of the entire field of family therapy (see Hoffman 1981) and it has inspired a variety of novel theoretical openings in other fields.  


In theoretical research, just like in all empirical research, the selection of data is crucial for the credibility of the outcome. Two dangers are constantly present. The first danger is data selection through blind chance or intuition without articulated justification. The second danger is the subordination of data selection to predetermined outcomes, i.e., use of data as mere illustration of conclusions fixed by the researcher in advance. In both cases, the typical critique focuses on the questionable representativeness or comprehensiveness of data.

At the beginning of their paper, Bateson and his collaborators explicate their database as follows. 

"The theory of schizophrenia presented here is based on communications analysis, and specifically on the Theory of Logical Types. From this theory and from observations of schizophrenic patients is derived a description of, and the necessary conditions for, a situation called the 'double bind' - a situation in which no matter what a person does, he 'can't win.' (...)

Our research in this field has proceeded by discussion of a varied body of data and ideas, with all of us contributing according to our varied experience in anthropology, communications analysis, psychotherapy, psychiatry, and psychoanalysis. We have now reached common agreement on the broad outlines of a communicational theory of the origin and nature of schizophrenia; this paper is a preliminary report of our continuing research." (Bateson 1972, 201-202.)

The data demonstrated in the paper itself consists mainly of (1) the philosophical Theory of Logical Types (adapted from Whitehead & Russel's Principia Mathematica),  as applied to communication, and (2) observations of schizophrenogenic family situations and schizophrenic patients. However, the data is presented in a rather brief and condensed manner. The whole paper consists of 27 pages in the 1972 book version. It contains 16 footnotes (of which two refer to personal communications).  No attempt is made at representativeness ofdata. The choice of data seems to stem from the authors' personal inspirations rather than from any systematic analysis of previous theories or of the history of schizophrenia. The whole paper bears the characteristics of a lucky hybrid: a goodidea that emerged in a group versatile, sophisticated and unconventional enough to embark on a challenging intellectual adventure. The credibility of the category generated (double bind) lies less in its database than in its immediately fascinating heuristic power and in the visions it opens.

Habermas's voluminous work is completely different in its relation to data. Thomas McCarthy, the translator of Habermas, gives the following characterization. 

"He develops these themes [of communicative action; Y.E.] through a somewhat unusual combination of theoretical constructions with historical reconstructions of the ideas of 'classical' social theorists. The thinkers discussed - Marx, Weber, Durkheim, Mead, Lukacs, Horkheimer, Adorno, Parsons - are, he holds, still very much alive. Rather than regarding them as so many corpses to be dissected exegetically, he treats them as virtual dialogue partners from whom a great deal that is of contemporary significance can still be learned. The aim of his 'historical reconstructions with systematic intent' is to excavate and incorporate their positive  contributions, to criticize and overcome their weaknesses, by thinking with them to go beyond them." (McCarthy 1984, vi-vii.)

In fact, Habermas pours a massive cavalcade of theories and concepts onto the canvas of his book. More specifically, it brings together  "the theories of action, meaning, speech acts, and other similar domains of analytic philosophy" (Habermas 1984, xxxiv) on the one hand and classical sociological theories on the other hand. In the 1174 pages of the book, there are 1242  footnotes (original German version; Habermas 1981). The reader is subjected to a virtual bombardment of sources. The credibility of the argumentation is very much based on the data. But it is not based on the professed representativeness of the data, rather on the internal connections and 'plots' found between and within the various sources.

In the present book, I follow neither Bateson & al. nor Habermas in my selection of data - and I follow both in certain respects.

I shall use three principal types of data in this book. The first type  of data consists of theories and theoretical propositions  pertaining to human learning and expansion. This type of data has the dominant role in the present work. In the selection and presentation of this data,  I am following certain structural steps or stages of argumentation.

First of all,  in each chapter (except Chapter 5, which is actually a methodological postscript), the construction of categories begins with an identification and characterization of the most advanced state of theorizing within the currently dominant paradigm.  With 'the most advanced' I refer to theorizing which either crystallizes the dominant conception in a very clear fashion or, in its aspiration to go further, tendentially exceeds the conceptual and methodological boundaries of the dominant paradigm and thus makes those boundaries or limits visible. However, such theorizing is also acknowledged as advanced within the paradigm - it is not generally disregarded as merely an eccentric curiosity.  Given the object of this book, the dominant paradigm is the cognitive psychology of learning and development. As its representatives, I am using Gagné, Norman, Kozielecki, and Hundeide in Chapter 1;  Bereiter, Langley & Simon, and Klix in Chapter 2;  Baltes & al., Brown, Riegel, Bronfenbrenner, Lerner, and Buss in Chapter 3;  Hallpike, Dreyfus & Dreyfus, Brehmer, Bruner, Miller, and Simon - and later a long list of others  - in Chapter 4.

Secondly,  to counter and problematize the propositions of cognitive psychologists, I examine and employ  certain classical theories  which put the problem of the chapter in question into a more penetrating light. The task of these sources is to enforce a deepening of the analysis so as to identify the long lineages or historical 'red threads'  of category formation. These classical theories were chosen on the basis of their known general characteristics, but in the course of the investigation, each one of them turned out to be a well of surprises. In Chapter 1,  I use the theory of C. G. Jung. In Chapter 2,  three classical lineages are examined: the semiotic and epistemological lineage from C. Peirce to K. Popper; the lineage from the symbolic interactionism of  G. H. Mead to modern interactionist developmental psychology; and the lineage of cultural-historical psychology from Vygotsky to Leont'ev.  In Chapter 3,  the work of G. Bateson is used. And in Chapter 4,  the theories or J. Dewey, M. Wertheimer, and F. Bartlett are examined.  

Thirdly, to develop the argument further, I take up and analyze the ideas of the cultural-historical theory of activity  in its modern form. This is the line of thought I try to continue and develop further. For that purpose, it is necessary to explicate the relevant insights produced within or close to this school of thought. In Chapter 2,  I discuss especially the analyses presented by A. N. Leont'ev and E. V. Il'enkov, but also those of V. P. Zinchenko, L. A. Radzikhovskii, and D. B El'konin. In Chapter 3,  I continue employing the work of  L. S. Vygotsky, A. N. Leont'ev and  their students, but related western works  by M. Wartofsky, R. Harré & al., I. Prigogine, M. Cole, S. Scribner, K. Holzkamp,  and others are also drawn upon. In Chapter 4,  especially the work of E. V. Il'enkov and V. V. Davydov on concept formation and dialectics is discussed, as well as the related ideas of M. Bakhtin on the dialogical nature of thought. And in Chapter 5,  the methodological ideas of L. S. Vygotsky, S. Scribner, and M. Cole are considered, along with the more specific suggestions of G. Altshuller and B. Fichtner. In general, this third step is not carried out in a dogmatic manner. Often in this stage of the analysis I take up theoretical insights that have not originated within the confines of any strictly delimited school - or have originated within schools of their own. Usually those insights are, however, based on philosophical and methodological assumptions which are substantively very much akin to those that have inspired the the cultural-historical school founded by Vygotsky, Leont'ev and Luria.

In all the three steps, I approach and use theory-historical data much in the same manner as Habermas approaches his data. The theories considered are taken as live discussion partners. While criticizing and often plainly rejecting them, I try to incorporate some of their wisdom into my further argumentation. Criticism for criticism's sake would not make much sense.   

The second type  of my data consists of general  historical accounts  of the development of human learning and expansion. Such data is mainly used in Chapter 2, in the sections concerning the evolution of activity and the cultural-historical evolution of human learning.

The section on the evolution of activity is a condensed systematic reconstruction based on the evolutionary and anthropogenetic data presented in works of Keiler, Leakey, Lewontin, Reynolds, and Schurig.This section does not intend to display an extensive variety of data because the subtle disagreements and variations in the interpretation of anthropogenesis  are not relevant for my argument. My conclusions rest on fairly generally accepted main features of the anthropogenesis. The end part of that section is based on the analysis of human societal production provided by Marx in Grundrisse.

The large section on the cultural evolution of human learning is divided into three sub-sections. The first one is a systematic reconstruction of the historical development of learning within schooling. In this sub-section, I rely on data on the development of literacy and schooling,  presented by researchers like Fichtner, Ong, Scribner & Cole, and others. The second sub-section is a reconstruction of the development of learning within work, this time restricted to the era of capitalism. This section begins with the data provided by Marx in Capital,  then goes on to discuss the effects of Taylorist rationalization, countering Braverman's linear dequalification thesis with a case provided by Hirschhorn. Finally the third sub-section discusses the development of learning within science and art. Studies by Zilsel, Lefèvre, Malinowski, Bronowski, Vygotsky, and Wartofsky are used as material in the reconstruction.

All these three sub-sections, as well as the section on the evolution of activity, bear the character of historically informed sketches, limited in scope and coverageThey are not object-historical investigations in themselves. They are sketches in the sense of working out preliminary basis for hypothetic categories.  Object-historical material is used much in the same way as the Theory of Logical Types  was used by Bateson & al., namely as a heuristic gateway (or a shortcut, or perhaps a crutch) for reaching the formulation of a hypothetic novel category. That is why secondary object-historical sources, used almost in an anecdotic fashion, are considered sufficient in this book. On the other hand, the gateway is here grounded in and preceded by the larger theory-historical discussion.  

The third type  of my data consists of accounts of specific historical cases  in the development of human learning and expansion. These cases serve as test material to which I apply the the categories formulated. At the same time, the analyses of the cases produce findings which enable me to develop the categories further.  There are two types of main cases and additional subsidiary cases.

The two types of main cases are (a) literary cases and (b) cases from the history of science. Two cases of both types are analyzed. In Chapter 3, I analyze the literary cases of The Adventures of Huckleberry Finn  by Mark Twain and Seven Brothers  by Aleksis Kivi. In Chapter 4, I analyze Mendeleev's discovery of the periodic law of elements, described and documented by B. F. Kedrov,  and the discovery of nuclear fission  which led to the construction of the atom bomb, as described and documented by R. Jungk. All the four cases are examples of expansive developmental transitions.

The reason for using literary fiction as data on developmental transitions is the following. Expansive developmental transitions are relatively long in duration. They are complex collective dramas where both the context and the actors are profoundly changed. Such processes are difficult to document, especially if one wants to catch the psychological aspects of the process. Classic developmental novels are often excellent reconstructions of such processes, "viewing the individual in movement, in constant development, as a necessary condition of his existence" (Bratus 1986, 95). Their validity and 'truthfulness' may of course be questioned. Surely they are not simple descriptions or direct recordings of events that have 'really happened'. But they have become classic for the very reason of expressing and reflecting, and indeed breeding and promulgating, something essential and concretely general in the expansive processes emerging in and typical to a certain culture and certain historical period. 

The use of accounts of important scientific discoveries, on the other hand, is justified by the increasing societal impact of such expansive processes. Also there exist some relatively well documented cases, such as the two I am using. In the case account on Mendeleev's discovery, Kedrov has had an exceptionally complete archive material at his disposal. Mendeleev had the habit of writing down even the small events and thoughts that occurred to him, and he also stored all these written documents with great care. In the case account on the discovery of the nuclear fission and on the subsequent construction of the atom bomb, Jungk had the opportunity of not only going through extensive written materials, including private correspondences, but also of interviewing personally an impressive number of the central personalities directly involved in this historical process.

Beside these four main cases, a few subsidiary cases are taken up and analyzed more superficially. These include Hirschhorn's account of the accident in the nuclear power plant on the Three Mile Island (Chapter 2) and Grünewald's account on Children's Campaign for Nuclear Disarmament (Chapter 3), as well as some other minor cases, presented mainly for the purpose  of illustration and concretization of the argument.

It may be asked why I have not used a single comprehensive report of my own concrete research as data. The answer is that the expansive developmental research methodology outlined in this book, especially in Chapter 5, requires a complex and extensive report to be understood. I found it impossible to incorporate such a report without either making the book unbearably voluminous or severely mutilating  the concrete research report.  This may be due to the fact that I am still too close to and too involved in the concrete pojects I could in principle have used as sources of data. In the text, I have also refrained from referring to any other publications of my own. My previous publications pertaining to the themes of this book are listed in a separate bibliography at the end of the book. 


In the presentation of a theory, i.e., in the outcome of theoretical research, the emergence of the categories may look simple, as if they had appeared from the 'pure thought' of the author. This kind of presentation is deceptive. It only reveals that the author himself is not conscious of the path he has gone. The better this path of processing categories out of data is brought into the open, the greater is the possibility that the reader may become involved in the theory as an active discussion partner and contributor to its further development. The theory becomes a processual entity and an instrument of its own development.

On the other hand, if the path or the process of derivation and critical analysis becomes the sole central focus, the outcome itself may get lost. When nothing seems to get fixed into clearcut categories, the reader has little to cling to in his own efforts of reconstruction, application and critique. Theory becomes a stream in which the reader tries to hold his head above the surface without quite knowing where he is floating to.

In the paper by Bateson & al., the new category (double bind) is presented immediately after the discussion of the use of Logical Types in communication. The category is first provisionally defined with the help of a series of six necessary ingredients. Then the effects of a double bind are characterized in general terms. After that, the category is concretized by embedding it into the context of the family situation, and further concretized by presenting illustrations from clinical data. The procedure is rather deductive and straightforward.

Strangely enough, unlike in so many deductive theories, the whole argumentation does not look like a finished and frozen structure. To the contrary, it evokes (and has indeed evoked) a host of questions, counter-arguments, application ideas, etc. How is this possible?

 I think that the reason is twofold. Firstly, about half way in the middle of the paper, the authors specify their database in an important way: "The theoretical possibility of double bind situations stimulated us to look for such communication sequences in the schizophrenic patient and in his family situation. Toward this end we have studied the written and verbal reports of psychotherapists who have treated such patients intensively; we have studied tape recordings of psychotherapeutic interviews, both of our own patients and others; we have interviewed and taped parents of schizophrenics; we have had two mothers and one father participate in intensive psychotherapy; and we have interviewed and taped parents and patients seen conjointly." (Bateson 1972, 212.) It seems obvious that this data has actually not only been used after  the category was found and formulated theoretically, as if for verifying and concretizing it only - although this impression is built into the deductive structure of the paper. Clearly the kinds of object-historical and actual-empirical data characterized above have played an important role in the very finding and formulation of the category. This conclusion is further supported by a footnote where the authors refer to one of the most famous first-hand object-historical sources on schizophrenia, namely Perceval's Narrative  from 1830-1832.  My argument is that Bateson & al. succeeded so well in hitting the core of their research object, or in finding something like its germ cell, not only because they had become acquainted with the philosophical Theory of Logical Types (as the paper implies) but because they actually had done and were doing very demanding object-historical and actual-empirical analysis of their object. The Theory of Logical Types probably functioned more like a springboard, a novel analogy needed for the breakthrough to take place.

The second reason for the liveliness of the theory of Bateson & al. is simply its incomplete and open-ended nature. Unlike the classical deductive theory, the paper stops short before even starting to deduce sub-categories from the central category of the double bind. The paper gives barely enough concretization by clinical illustrations to set off the reader's own thought experiments. This has been a source of much frustration and much creative effort.

 If Bateson & al. develop their category with one piercing sting, the method employed by Habermas is more like spinning and weaving a complicated conceptual texture or web.  The entire texture is extremely demanding for the reader because of the multitude of excursions and  sidetracks. But on the whole, the chain of argumentation is logical.

 Habermas's starting point is an explicit shift from the paradigm of consciousness to the paradigm of language  as speech. The goal-directed actions of different individuals are socially coordinated, and language is the means of coordinating them. The fundamental category of communicative action  is established on this basis: it is a coordinating action aimed at "reaching understanding in the sense of a cooperative process of interpretation" (Habermas 1984, 101). From this basis, the category of communicative competence  is derived. This in turn implies a general category of rationality  as achieving mutual understanding in communication that is free from coercion.  The category of communicative action is used to analyze "whether and in what sense the modernization of a society can be described from the standpoint of cultural and societal rationalization" (Habermas 1984, 6). The categories of modernity  and rationalization  areanalyzed with the help of the categories of lifeworld  and system  which together form Habermas's two-level concept of society. Modernity is analyzed as rationalization and colonization of the lifeworld, or as the decoupling of lifeworld and system.

All these categories are worked out and elaborated through the theory-historical data provided  by  the classical sociological theories of Weber, Lukacs, Adorno, Mead, Durkheim, Parsons, and Marx.

 This chain of categories - coordination - language - communicative action - communicative competence - rationality - modernity - rationalization - life world - system - is not linear or deductive in any simple sense. The links of the chain, i.e., the chapters and sections of the book, are in themselves relatively independent cycles of argumentation and analysis.  Still the chain is a logical whole. It follows a complex and bouncy logic of interconnections and mutual transitions which is not very clearly explicated by the author. The reader has to reconstruct the logic for himself with great efforts. This is obviously the intention of the author. The ideal reader dwells in the book, moves back and forth, discovers new connections and  ideas by diving into the texture time and again. Of course the problem is that there may not be very many such ideal readers. Many a reader will drown in the conceptual stream, never reaching the point of constructing his own vessels for sailing.

In the present book, too, the central chapters are relatively independent cycles of analysis and category construction. Each one of Chapters 2, 3, and 4 follows roughly the same logic. At first, the problem is presented by introducing certain antinomies or conceptual troubles within cognitive psychology. Secondly, the problem is elaborated using theory-historical data. Thirdly, the new categories are provisionally characterized, defined and modelled. Fourthly, the new categories are tested and further elaborated using general object-historical accounts or specific object-historical cases as data. Fifthly, some implications are discussed and an intermediate balance is drawn as a preparation for the next round of category construction.  The  sequence may be partially repeated and the order of some steps may be changed, but this is the general logic of the argumentation.

In Chapter 2, the task is to find the initial abstraction, the germ-cell category that can mediate between learning and expansion. The analysis proceeds through the five steps named above in the following manner. (1) The problem is presented as the 'learning paradox' of Bereiter and as the problem of the evolution of learning as posed by Klix. (2) The problem is elaborated using  the theory-historical data from three lineages which have taken the system of man-in-society or individual-in-context  as their basic unit of  analysis. (3.1) The general category of activity  is defined and modelled. (4) Three historical lines of the cultural evolution of human learning are interpreted with the model of activity. (3.2) The germ-cell category of learning activity,  or learning by expanding,  is defined and modelled as the outcome of the preceding step. (5) Two sets of implications are discussed, namely those concerning the subject of learning activity and those concerning the emergence of learning activity in the ontogenesis.

In Chapter 3, the task is to find the mechanism of transition from learning to expansion, from everyday individual actions to novel collective activity. (1) The problem is presented as the dilemma of learning vs. development and as the dilemma of individual vs. societal development. (2) First Bateson's work, then more recent activity-theoretical and related works are employed as theory-historical data to elaborate the problem. (3.1) The category of the zone of proximal development  is defined as the solution to the problem. (4)  Two historical case accounts of expansive transition (classic developmental novels) are analyzed with the help of the category of the zone of proximal development. (3.2) The analyses yield a more detailed picture of the phases or steps within the zone of proximal development - the stepwise structure is modelled. (5) Instructional implications of the category are discussed.

In Chapter 4, the task is to find the central instruments needed for the mastery of expansive transitions,  or zones of proximal development.  (1.1) The problem is presented in the form of three dichotomies in cognitive theories of thinking. (2.1) The ideas of Dewey, Wertheimer and Bartlett are analyzed as theory-historical data to elaborate the problem. (1.2) The dilemma of advanced cognitive theories of concepts is taken up as an extension of the initial problem. (2.2) Activity-theoretical ideas of concepts are analyzed as theory-historical data to elaborate the problem further. (3.1) Three basic types of secondary instruments of expansive transitions are defined: springboards, models and microcosms.  (4) Two historical case accounts of expansive transition (scientific discoveries)  are analyzed and the secondary instruments employed in the cases are identified. (2.3) Theories of dialectical and dialogical thinking are analyzed as further theory-historical data. (3.2) A provisional definition of dialectics  as the tertiary instrument of expansion is suggested. (5) Implications for concrete research methodology are pointed out.

My way of processing categories out of data in these three chapters has certain affinities both with Bateson & al. and with Habermas. I try to share with Bateson & al. the way of defining the novel categories found in a relatively unambiguous and systematic manner. This entails a certain risk of rigidity. On the other hand, I share with Habermas the aspiration to proceed through a chain of cyclic analyses of theory-historical data where theories are treated as live discussion partners. This entails a certain risk of drowning the reader in theories. In the worst event, these risks reinforce each other. In the best event, they balance and neutralize each other.

There are further two specific features of presenting and processing data in this book. The first one is the extensive use of quotations from the theoretical sources discussed and analyzed. The second one is the almost equally extensive use of graphic models.

All theories have a dual character. They are simultaneously fixed conceptual structures and living processes of continuous concept formation. The continuous development of the theory is possible only from within it, through its immanent contradictions and gaps. The more polished and closed the appearance of the theory, the harder it is for the reader to enter the immanent process of its critical elaboration. Glazman (1972, 204) points out that scientists may more or less consciously construct 'windows' in their theories. These windows are gaps, inconsistencies or ambivalent formulations which invite the reader to engage in immanent polemics with the author.

In this book, I use quotations as windows into the innermost movement and dynamics of my theory construction. In theoretical research, the difference between displaying original quotations and only the author's own interpretations of the given theoretical sources is much the same as the difference between displaying original interview protocols and only questionnaire data in actual-empirical research. In other words, the quotations serve in theory what in empirical anthropology would be called 'thick description' (Geertz 1973).

An original quotation, when it is not mishandled and mutilated so as to be totally subordinated to the single-minded purpose of the author, represents a voice  and a language  of a researcher other than the author. It represents a dynamism of its own, never perfectly in line with the author's intentions. It allows for a variety of interpretations and associations, not only the ones the author employs in his line of reasoning. The intentional use of multiple voices, multiple languages, is called heteroglossia. 

"Heteroglossia (...) is another's speech in another's language,  serving to express authorial intentions but in a refracted way. Such speech constitutes a special type of double-voiced discourse.  It serves two speakers at the same time and expresses simultaneously two different intentions: the direct intention of the character who is speaking, and the refracted intention of the author. In such discourse there are two voices, two meanings and two expressions. And all the while these two voices are dialogically interrelated, they - as it were - know about each other (...); it is as if they actually hold a conversation with each other." (Bakhtin 1982, 324.)

 For example in this quotation, Mikhail Bakhtin is speaking about heteroglossia in the novel, not in scientific theorizing. I am using his voice to express, in a refracted form, my  intentions and arguments about heteroglossia in theoretical research. But his voice does not yield to my purposes without simultaneously producing what Bakhtin (1982, 325) calls 'dialogized ambiguity'.

 Quotations are not  primarily used for illustrative purposes in this book. To the contrary, quotations function here like pieces of a puzzle or a mosaic. The overarching theme and conceptual pattern of this book emerge  through the quotations. The dialectical derivation of categories demands that the research becomes "sunk into the material in hand", "following the course that such material takes" (Hegel 1966, 112). The aim is that "by this process the whole as such, surveying its entire content, itself emerges out of the wealth wherein its process of reflection seemed to be lost" (Hegel 1966, 113).

 My extensive use of graphic models serves a twofold purpose. For the first thing, it aims at making the central categories found transparent and compact. This the representation  function of the models. But I use the graphic models in series of successive variations, not just as singular representations. The series of successive variations serve the instrumental  or processual  function of the models. With the help of such variations, I try to demonstrate how the models can depict movement and change. The reader is invited to formulate and test his own variations. 


A theory is a potential instrument for dealing with practice. Within theories of man and society, such as those discussed in this book, different intended practice-relations are embedded. The practice-relation built into traditional theories is that of speaking to academic empirical researchers  who shall verify and concretize the theoretical categories.  In such traditional theories, the societal practice remains a distant testing ground, used mainly (a) as source of ex post facto  data or of data abstracted via experimental designs (see  Maschewsky 1977), and (b) as object of benevolent recommendations based on the findings gained in research.

There are at least two more radical and direct ways of building the practice-relation into the theory. One alternative is to speak directly to professional practitioners  in the field the theory is concerned with, that is, to prompt them to act as experimenters in their practical contexts.  Another alternative is to speak to social movements  concerned with the problems the theory is trying to illuminate. The classical example is of course the theoretical work of Karl Marx and Friedrich Engels.

The paper by Bateson & al. quite clearly speaks to professional practitioners in the field of psychotherapy. "The understanding of the double bind and its communicative aspects may lead to innovations in therapeutic technique. (...) double bind situations occur consistently in psychotherapy. At times these are inadvertent in the sense that the therapist is imposing a double bind situation similar to that in the patient's history, or the patient is imposing a double bind situation on the therapist. At other times therapists seem to impose double binds, either deliberately or intuitively, which force the patient to respond differently than he has in the past. (...) Many of the uniquely appropriate therapeutic gambits arranged by therapists seem to be intuitive. We share the goal of most psychotherapists who strive toward the day when such strokes of genius will be well enough understood to be systematic and commonplace." (Bateson & al. 1972, 225-227.)

The practice-relation built into Habermas's work is more ambiguous. Habermas emphasizes that he has written his book for researchers, "for those who have professional interest in the foundations of social theory" (Habermas 1984, xlii). On the other hand, he points out that new kinds of conflicts and social movements have developed in advanced Western societies during the last years. "They do not flare up in areas of material reproduction; they are not channeled through parties and associations; and they are not allayed by compensations that conform to the system. Rather, these new conflicts arise in areas of cultural reproduction, of social integration and of socialization; they are carried out in subinstitutional, or at least extraparliamentary, forms of protest (...). It is not primarily a question of compensations that the social-welfare state can provide, but of protecting and restoring endangered ways of life or of establishing reformed ways of life." (Habermas 1981, Vol. 2,  576.) Here, toward the end of his book, Habermas is increasingly speaking to the 'new social movements'. He mentions such phenomena as the ecology and antinuclear movements, the limits-to-growth debate, the peace movement, the women's movement, experiments with communal and rural living, liberation movements of various minority groups, conflicts over regional and cultural autonomy, protests against 'big government', religious fundamentalism and the proliferation of religious sects, the multifarious 'psychoscene,' the proliferation of support groups, and the like. Most of these are purely defensive, only some (like feminism) have offensive features grounded in modernity. Habermas summarizes his message to such movements: "Restricting the growth of monetary-administrative complexity is by no means synonymous with surrendering modern forms of life. In structurally differentiated lifeworlds a potential for reason is marked out that cannot be conceptualized as a heightening of system complexity." (Habermas 1984, xlii.) The perspective offered in this message is vague optimism, promising some free room for the movements with their emancipatory and defensive communicative actions in the enclaves of the modern rationalized society.

In the present book, I am speaking to both researchers and practitioners,  whether the latter be professional or blue collar, or engaged in activities entirely other than wage labor. The methodology of expansive research sketched in Chapter 5 is necessarily a joint venture. The researcher (or rather, the team of researchers) has the task of pushing the cycle of expansive transition forward and introducing instruments or components for new instruments into it. The practitioners have the task of facing and solving the contradictions of their activity system as they are identified and aggravated along the voyage through the zone of proximal development. In this process, the practitioners tendentially become subjects - or rather a collective subject - of their evolving new activity system, thus also subjects of analysis and intervention.

In other words, the methodology proposed in Chapter 5 is not only a methodology of research but also a methodology of practical societal transformation. This means that I am also speaking to social movements. But social movements are not empiristically taken as something given. Rather, they are conceived of as something potentially emerging,  something in the process of becoming, within any real societal activity system.

Here I disagree with Habermas who seems to see hope only outside the system of production and administration. I contend that such a stance indicates a lack of intimate knowledge about the inner contradictions and emancipatory dynamics within the world of wage labor, be it in production or administration. In the heart of modern production and administration, also the hidden powers of qualitative change are greatest.  Retreat into the safe world of academic discourse is today almost a guarantee of distorted observation. The naive optimism of Bateson & al., prophesying 'innovations' in professional therapeutic work, has a deeper historical truth in it than the wordy roundabouts of Habermas. 


 The problems motivating this inquiry are (1) the increasingly recognizable futility of learning in its standard reactive forms, and (2) the elusive and uncontrollable nature of expansive processes where human beings transcend the contexts given to them. The hypothesis guiding the further course of my study is that learning and expansion are becoming integrated, forming a historically new type of activity.

 Thus, the present study falls into the category of general developmental and educational theory. For reasons that will become clear in Chapter 2, I see the central fields of application of this theory in the life practices of adults and adolescents, especially in the interrelations of work and learning.

 The method used in this study is dialectical derivation and construction of categories. Each substantive chapter is a relatively independent cycle of analysis and construction, following roughly the same logical sequence. (1) The problem is presented by introducing certain antinomies or conceptual troubles within cognitive psychology. (2) The problem is elaborated using theory-historical data. (3) The new categories are provisionally characterized, defined and modelled. (4) The new categories are tested and further elaborated using general object-historical accounts or specific object-historical cases as data. (5) Some implications are discussed and an intermediate balance is drawn as a preparation for the next round of category construction.

 The outcomes of the study are condensed into a series of graphic models. Since these models are instruments of thought and practice, they are best understood by following their creation and by applying them in activity.