Thursday, June 04, 2015

Reading :: Learning by Expanding, 2ed


Yrjo Engestrom's 1987 book Learning by Expanding is a hugely influential book. Published the same year that he earned his PhD, LBE lucidly described activity theory, applied it to adult education, and expanded it to address a more globalized, networked world. The book was picked up fairly quickly in human-computer interaction research, but also in my field of writing studies, and for the same reason: we were searching for a contextualized, developmental account of cognition that was more socioculturally oriented than information-processing cognitive science, but more structured and cognitively oriented than anthropology.

LBE has been available for some time on the UCSD website, both in HTML and PDF. I inadequately reviewed that version back in 2003. But in late December 2014, Cambridge University Press put out the second edition, making it far easier for scholars to obtain and cite this influential book.

Although this book is the second edition, as Engestrom explains in the Preface, the original text is only lightly revised for style. The major addition is a new introductory chapter that discusses how the theory of expansive learning has been taken up since its publication.

So let's spend some time with that introduction: "Learning by Expanding: Origins, Applications, and Challenges." Here, Engestrom recalls three reasons for writing the book.

The first reason was "in order to formulate a strong alternative to the dominant Cartesian views of cognition and learning that depicted the human mind as if it were a computer, isolated from the cultural context." It was part of the same "groundswell" as everyday cognition, situated action, and cognition in practice (p.xiii). (And that is what made it especially attractive to those of us in writing studies in the early 1990s.)

The second reason, he says, was methodological. Engestrom had "grown up as an activist of the radical student movement [and] was convinced that research needs to be actively involved in making the world better" (p.xiii). So, rather than developing an observational methodology, he wanted to develop an interventionist one. Activity theory, he says, was developed in the USSR in interventionist terms, so it was a strong fit.

The third reason was that AT was a
potent framework for understanding and changing the world. In the Soviet Union, activity theory had a sixty-year history of original insights, groundbreaking research, and severe oppression. In the West, Vygotsky's work was found and promoted from the 1960s on in escalating steps by well-known North American scholars, such as Jerome Bruner (1962) and Michael Cole and Sylvia Scribner (1978). Activity theory, the most important heir and extension of Vygotsky's legacy, was primarily discovered by radical European scholars and students in the 1970s and 1980s mainly through the works of Leont'ev (1978). In the late 1970s the works of Vassily Davydov (1977, 1990) made a strong impression on me, and I was fortunate enough to persuade him to visit Finland in the early 1980s. The first international conference on activity theory was organized in West Berlin in 1986. Learning by Expanding is a fruit of that movement. (p.xiv). 
I've quoted the above at length because it's part of the AT origin story that Engestrom set forth in the original LBE.  He recapitulates that origin story in the next section. According to Engestrom, AT has gone through three generations, which I'll designate as 1GAT, 2GAT, and 3GAT:

1GAT: Centered around Vygotsky's insight into mediation, illustrated in Vygotsky's triangular model of subject, object, and mediating artifact. Inserting a cultural artifact into the subject-object relationship "was revolutionary in that the basic unit of analysis now overcame the split between the Cartesian individual and the untouchable societal structure. ... Objects became cultural entities and the object orientedness of action became the key to understanding human psyche" (p.xiv). Yet the focus was limited to the individual.

2GAT: That limitation was "overcome by the second generation, led and inspired by Leont'ev's work" (p.xiv). Leont'ev "showed how historically evolving division of labor had brought about the crucial differentiation between an individual action and a collective activity" (pp.xiv-xv). Yet, Engestrom says, "the relationship between object-oriented production and communicative exchange beween people remained somewhat unclear in Leont'ev's work" and Vygotsky's triangle diagram was never graphically expanded in this work (p.xv). Beyond that, in the Soviet Union, activity theorists mostly restricted their studies to "play and learning among children" and "Contradictions of activity remained an extremely touchy issue" (p.xv). In the 1970s, he recounts, radical researchers in the West picked up the concept and applied it to other domains, including work, and Ilyenkov's writings on contradictions became incorporated. Yet 2GAT was insensitive to cultural diversity, as Cole pointed out (p.xv).

3GAT: Here, Engestrom drops the human actors in his sentences; you can mentally insert "Engestrom" as the actor in these sentences, since he defined this agenda in LBE.
The third generation of activity theory is developing conceptual tools to understand networks of interacting activity systems, dialogue, and multiple perspectives and voices. In this mode of research, the basic model is expanded to include minimally two interacting activity systems. This move toward networks of activities, while still in an embryonic form, was anticipated in the original text of Learning by Expanding" (p.xv)
He adds that in 3GAT, intersecting activity systems have "partially shared and often fragmented objects" (p.xv) and studying them involves studying subjectivity, embodiment, emotion, and identity (pp.xv-xvi). And
there is a risk that activity theory is split into the study of activity systems, organizations, and history, on the one hand, and subjects, actions, and situations, on the other hand. This is exactly the kind of split the founders of activity theory set out to overcome. To bridge and integrate the two directions, serious theoretical and empirical efforts are needed. (p.xvi)
He sets out such efforts, seeing developmental work research as an agenda of application. Here's his summary of LBE's themes:
The central ideas of this book may be condensed into the following five claims: (1) The object-oriented and artifact-mediated collective activity system is the prime unit of analysis in cultural-historical studies of human conduct; (2) historically evolving inner contradictions are the chief sources of movement, change, and development in activity systems; (3) expansive learning is a historically new type of learning, which emerges as practitioners struggle through developmental transformations in their activity systems, moving across collective zones of proximal development; (4) the dialectical method of ascending from the abstract to the concrete is the key for mastering cycles of expansive learning; and (5) an interventionist research methodology that aims at pushing forward, mediating, recording, and analyzing cycles of expansive learning in activity systems is needed. (p.xvi)
Here, Engestrom discusses how each of these claims has been explored in the time since LBE was published (primarily citing his work and the work of his students). I won't treat these as exhaustively as the above, but I'll pick out some things that are interesting for my own current line of inquiry.

One is that "The theory of expansive learning is a process or phase theory of learning. In other words, it proposes an ideal-typical sequence of learning actions that together make an expansive learning cycle"—but, Engestrom adds, this sequence is not universal or strongly prescriptive (p.xviii).

Another is that "The theory of expansive learning is based on the dialectics of ascending from the abstract to the concrete" (p.xx; cf. Ilyenkov). In the expansive cycle, individual subjects begin by questioning an accepted practice. Through a cycle or spiral of learning, they expand into a "collective movement or institution." This learning spiral involves constructing and resolving "successively evolving contradictions in the activity system" (p.xx). Engestrom cites Davydov's ideal-typical sequence of learning in the classroom:
According to Davydov (2008), an ideal-typical sequence of learning activity consists of the following six learning actions: (1) transforming the conditions of the task in order to reveal the universal relationship of the object under study; (2) modeling the identified relationship in a material, graphic, or literal form; (3) transforming the model of the relationship in order to study its properties in their "pure guise"; (4) constructing a system of particular tasks that are resolved by a general mode; (5) monitoring the performance of the preceding actions; and (6) evaluating the assimilation of the general mode that results from resolving the given learning task. (p.xx-xxi)
Engestrom describes what he sees as a parallel expansive learning cycle:

  1. Questioning
  2. Analyzing
  3. Modeling
  4. Examining the model
  5. Implementing the model
  6. Reflecting on the process and "consolidating its outcomes into a new stable form of practice" (p.xxi)
Large-scale cycles, he says, may contain smaller-scale ones (p.xxii). (I of course thought of Boyd's OODA loop here, although it is not learning so much as adjusting tactically and strategically.)

Under "The Vertical and the Horizontal in Learning and Development," Engestrom discusses the fact that (3G)AT's activity systems are "multi-voiced" and the "horizontal" aspect of the activity is thus dialogical, involving a "discursive search for shared meanings" (p.xxiv). Engestrom cites Bakhtin via Wertsch here. Citing Ritva Engestrom's 1995 argument that Bakhtin's ideas on language parallel Leont'ev's levels of activity, he adds, "One might say that activity theory, and developmental work research as its application, incorporated dialogue and discourse into their foundational repertoires in the 1990s" (p.xxiv). Certainly that's about the time that we in writing studies started putting together AT-genre syntheses. And most likely it could not have been done earlier, in the Soviet Union. 

Engestrom addresses some common critiques of 3GAT. For instance, he acknowledges Bakhurst's critique that AT's triangular models have minimal predictive power and seem developed to address only certain kinds of (well structured) activities; he says that the counterexamples that Bakhurst proffers are not "relatively durable collective activities" (p.xxviii). Bakhurst also critiques the 3GAT understanding of contradictions and says that Ilyenkov would be "dismayed" by this application; but Engestrom argues that Bakhurst doesn't really understand what Ilyenkov means by contradictions (p.xxiv; contra Engestrom, Bakhurst does understand what Ilyenkov is trying to say, he just doesn't find it to be plausible).

Finally, at the end of the introductory chapter, Engestrom addresses the future of the theory. 3GAT faces challenges. For instance, "Activity systems are increasingly interconnected and the combinations of activities keep changing rapidly"; thus Engestrom and collaborators have developed the concept of negotiated knotworking, "involving new tools and rules for rapidly pulsing renegotiation of the object" (p.xxxvi). Another is "runaway objects" or "hyperobjects," "objects so massively distributed in time and space as to transcend localization, such as climate change or pandemics" (p.xxxvi; see also Engestrom 2008). A third concept is "'wildfire activities' and mycorrhizaelike forms of organizing" (p.xxxvi), which are "hybrid and poorly bounded" (p.xxxvii). A fourth is that of polycentric social movements (p.xxxvii). Engestrom suggests that all of these must be investigated next.

So—the above is a summary of just the new introduction to the book. Now we get to the book itself.

In Chapter 1, "The Futility of Learning," Engestrom critiques influential books on learning that are situated within cognitive psychology. Such books focused on helping learners to "cope with the tasks given to them"—and focused on quantitative improvements such as speed and scale (p.2, his emphasis). Instead, Engestrom argues, we should explore expansive learning, "which transcends the context given" (p.4). He argues:
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. (p.7, his emphasis)
This argument is theoretical, but based on data:

  1. "theories and theoretical propositions pertaining to human learning and expansion." (p.11)
    1. "an identification and characterization of the most advanced state of theorizing within the currently dominant paradigm." (p.11)
    2. "I examine and employ certain classical theories which put the problem of the chapter in question into a more penetrating light." (p.12)
    3. "I take up and analyze the ideas of the cultural-historical theory of activity in its modern form." (p.12)
  2. "general historical accounts of the development of human learning and expansion." (p.13)
  3. "accounts of specific historical cases in the development of human learning and expansion" (p.14), including 
    1. literary cases and 
    2. cases from the history of science
Engestrom sets his intention to draw from many different perspectives, respecting the heteroglossia (p.20) while striving to develop a unified theory.


In Chapter 2, "The Emergence of Learning Activity as a Historical Form of Human Learning," Engestrom uses Carl Bereiter's "learning paradox" to illustrate the difference between cognitivism and the cultural-historical approach to psychology. How can a structure generate another structure more complex than itself? How can one learn complexity without an existing structure that embodies that complexity? (pp.25-26). Bereiter briefly considers Vygotsky's answer, in which complexity is situated within the culture rather than the child, but dismisses it because Vygotsky doesn't solve the concept of "internalization" to his satisfaction (p.26). But, Engestrom argues, Bereiter is looking for an explanation of how complex structures are (re)created within the head of the individual; learning is reduced to internalization (p.27). "Is the evolution of learning essentially a story of progressively enlarged capacity for internal individual processing of information?" (p.28). Drawing on Zinchenko, Engestrom argues that to understand learning, we must be able to conceive it within "concrete historical activities with specifiable subjects, objects, and instruments, within specifiable contexts" (p.30).

This argument leads him to discuss 19th-century breakthroughs in philosophy (Hegel), biology (Darwin), and social sciences (Marx):

Two fundamental features are evident in these breakthroughs. First, they meant that organism and environment, man and society, were no more seen as separate entities but as integral systems within which retroactive causality and internal dynamic transitions prevail. Second, these breakthroughs meant that organism and environment, man and society, could no more be understood as stable, unchanging entities but only as something characterized by qualitative transformations requiring a historical perspective. (p.31)
Along these lines, Engestrom set his intention to develop
a viable root model of human activity with the following initial delimitations. First, activity must be pictured  in its simplest, genetically original structural form, as the smallest unit that still preserves the essential unity and quality behind any complex activity. 
Second,  activity must be analyzable in its dynamics and transformations, in its evolution and historical change. No static or eternal models will do. 
Third, activity must be analyzable as a contextual or ecological phenomenon. The models will have to concentrate on systemic relations between the individual and the outside world. 
Fourth, specifically human activity must be analyzable as culturally mediated phenomenon. No dyadic organism-environment models will suffice. This requirement stems already from Hegel's insistence on the culturally mediated, triadic  or triangular structure of human activity. (pp.32-33)

He finds prerequisites in three traditions, all of which put mediation at the forefront: the study of signs (Pierce); intersubjectivity (Mead); and CHAT (Vygotsky) (p.33). Rather than get bogged down in this lengthy discussion, let's fast forward to CHAT, where Engestrom has this interesting critique of Vygotsky:
For me, Wartofsky's secondary artifacts and Vygotsky's psychological tools are essentially the same thing. Vygotsky's intellectualist bias (see Leontiev & Luria 1968, 354-355) led to a somewhat one-sided emphasis on signs and word meanings. The broader category of psychological tools, as well as the exciting relations between technical and psychological tools were not elaborated concretely by Vygotsky. Ironically, the activity-oriented approach in Soviet psychology after Vygotsky tried to get rid of Vygotsky's intellectualism by neglecting the problem of signs and psychological tools in general: "if the polemic with concrete works of Vygotsky on the problem of the sign was necessary and natural, the removal of this problematic - in principle - led only to a substantial 'narrowing' of the theory of activity" (Davydov & Radzikhovskii 1985, 60). In the recent revival of Vygotskian studies, signs may again be treated too much 'on their own', separated from the spectrum of psychological tools and their relations with primary tools. This danger seems to lure even in outstanding analysis, such as that of Wertsch's (1985b) on Vygotsky's concept of semiotic mediation. (p.50)

Engestrom is here referencing the split between Vygotsky and Leontiev when the latter left for Kharkov, the one that Kozulin alleges created later theoretical and methodological problems for 2GAT. According to both of these linked references, the problem was not that Vygotsky emphasized signs too much and Leontiev emphasized them too little; the problem was that they disagreed on what root a Marxist psychology should have. Do we start with signs, as Vygotsky did, focusing on the semiotic characteristics that distinguish us from apes? Or do we start with labor, as Leontiev did, hewing closely to the Engelsian account in which labor made man? Leontiev made the politically smart choice, and won the Lenin Prize. Vygotsky made what he would consider the methodologically smart choice, and faded into obscurity as his books and articles failed to be reprinted. (Later his reputation was revived by Leontiev and Luria.)


The Leontievan account of labor emphasizes tool use, and as Engestrom describes the levels of activity, he bluntly states that "Tools are crystallized operations" (p.54). In fact, tools are one of three categories of mediators in human activity—the others are rules and division of labor—that are missing for animals (pp.59-61). For animal activity, an individual member of the species interacts with others in the population within the natural environment; this configuration is represented by a flat-topped triangle (p.60). For animals evolving into humans, these points are mediated by emerging tool making; emerging collective traditions, rituals, and rules; and emerging division of labor; the population is now a community; and the environment is now not only natural but also artificial (p.61). (This evolutionary account isn't sourced to Leontiev in this text, but it's essentially his account, which is in turn based on Engels'.) After this transitional phase, we get the familiar triangular diagram of activity, with area labels for production, exchange, distribution, and consumption; Engestrom sources these area labels to Marx's Grundisse (p.63).

In fact, it's striking how Engestrom has not only illustrated Leontiev's structure of activity but also developed it using other Marxist components. In addition to drawing the area labels from Grundisse, he draws from Marx and Ilyenkov to supplement the account of contradictions:
The basic internal contradiction of human activity is its dual existence  as the total societal production and  as one specific production among many. This means that any specific production must at the same be independent of and subordinated  to  the total societal production (see Damerow, Furth, Heidtmann & Lefèvre 1980, 240-241). Within the structure of any specific productive activity, the contradiction is renewed as the clash between individual actions and the total activity system. This fundamental contradiction acquires a different historical form in each socio-economic formation. 
The fundamental contradiction arises out of the division of labor. (p.66)
He argues that  "In pre-capitalist socio-economic formations, the basic contradiction, the subordination of individual producers to the total system of production, took the form of immediately visible personal suppression by force,  be it exercised by slave-owners or feudal lords"; but "In capitalism, the contradiction acquires the general form of commodity.  Commodity is an object that possesses value  (i.e., exchange value),  not only and not primarily use value. The value of the commodity is basically determined by the average necessary amount of social labour needed for its production" (p.67). "In capitalism, all things, activities and relations become saturated by the dual nature of commodity - they become commodified. The relation between individual actions and collective activity, between specific productions and the total production, is transformed accordingly" (p.68). And thus "The essential contradiction is the mutual exclusion and simultaneous mutual dependency of use value and exchange value in each commodity. This double nature and inner unrest is characteristic to all the corners of the triangular structure of activity. It penetrates the subject and community corners because labour force itself is a special kind of commodity" (p.69). He cites Leontiev here—Leontiev was of course safe in pointing out the contradictions in capitalism!



Engestrom goes on to discuss four levels of contradictions:
The primary contradiction of activities in capitalist socio-economic formations lives as the inner conflict between exchange value and use value within each corner of the triangle of activity. 
The secondary contradictions are those appearing between the corners. The stiff hierarchical division of labor lagging behind and preventing the possibilities opened by advanced instruments is a typical example. 
The tertiary contradiction appears when representatives of culture (e.g., teachers) introduce the object and motive of a culturally more advanced form of the central activity into the dominant form of the central activity. For example, the primary school pupil goes to school in order to play with his mates (the dominant motive), but the parents and the teacher try to make him study seriously (the culturally more advanced motive). The culturally more advanced object and motive may also be actively sought by the subjects of the central activity themselves. 
The quaternary contradictions require that we take into consideration the essential 'neighbour activities' linked with the central activity which is the original object of our study. (pp.70-71).


Some pages later, Engestrom illustrates the primary contradiction of modern work activity—in capitalism—which, as stated above, is the conflict between use-value and exchange-value at every corner of the activity system (p.90). And he adds:
The increasingly societal nature of work processes, their internal complexity and interconnectedness as well as their massive volumes in capital and capacity, are making it evident that, at least in periods of acute disturbance or intensive change, no one actually quite masters the work activity  as a whole, though the control and planning of the whole is formally in the hands of the management. This creates something that might be called 'grey zones' (Projekt Automation und Qualifikation 1981), areas of vacuum or 'no man's land', where initiative and determined action from practically any level of the corporate hierarchy may have unexpected effects. (p.90)
Later, he argues:
The object of learning activity is the societal productive practice, or the social life-world, in its full diversity and complexity. The productive practice, or the central activity, exists in its presently dominant form as well as in its historically more advanced and earlier, already surpassed forms. Learning activity makes the interaction of these forms, i.e., the historical development of activity systems, its object. 
This object appears to the subject first in the form of discrete tasks, problems and actions. As Michael Cole (1983, 51) notes, "discovery of the goals is essential to true activity". Learning activity (a) analyzes and connects these discrete elements with their systemic activity contexts, (b) transforms  them into contradictions demanding creative solution., and (c) expands and generalizes them into a qualitatively new activity structure within societal productive practice. (p.99)

This brings us to Chapter 3, "The Zone of Proximal Development as the Basic Category of Expansive Research." Here, Engestrom discusses Bateson's double bind (p.113), reinterpreting it within 3GAT:
The distinctive feature of human activity is that it is continuous creation of new instruments that in turn complicate and change qualitatively the very structure of the activity itself.  It is essential that human activity cannot be reduced to the upper sub-triangle of Figure 2.6 alone. Human activity is not only individual production. It is simultaneously and inseparably also social exchange and societal distribution. In other words, human activity always takes place within a community governed by a certain division of labor and by certain rules. (p.114)

 After reviewing work by Leontiev, Harre, Bateson, Raiethel, Wartofsky, and Judin, Engestrom develops an elaborate table describing the hierarchical structure of activity in terms of the points of the 3GAT triangle (p.122). And he returns to the question of the double bind:
A double bind may now be reformulated as a social, societally essential dilemma which cannot be resolved through separate individual actions alone  - but in which joint co-operative actions can push a historically new form of activity  into emergence. (p.131)
He describes Vygotsky's Zone of Proximal Development (pp.134-139), then illustrates it with The Adventures of Huckleberry Finn (pp.139-153) and Seven Brothers (pp.153-167). I found these illustrations less satisfying than the actual cases he uses elsewhere.

In Chapter 4, "The Instruments of Expansion," Engestrom argues that "a new type of 'learning by expanding' is emerging in the current phase of human history. This implies that the transitions mentioned above are becoming potential objects of conscious or intuitive mastery" (p.169). We'll skip past his account of dichotomies in the literature and look at how he characterizes instruments:
In the realm of secondary artifacts, similar types may be tentatively distinguished. Firstly, the continuously changing flow of information, consisting of  specific opinions, news, descriptions, advertisements, etc. may be identified as the rapidly renewed  objects of consumption on the secondary level. 
Secondly, the relatively stable and general representations with which we filter and modulate our daily information flow may be identified as the slowly renewed objects of consumption on the secondary level. 
Thirdly, both above-mentioned types may be turned into objects or raw materials of production, to be molded and transformed into something new. 
Fourthly, sign systems such as gestures, spoken and written language, or mathematical and musical notation may be identified as typical, continuously available instruments of production on the secondary level. (p.185)
and
Expansive thinking requires that relatively stable objects  of consumption and production are transformed into instruments  of production. Cycles of expansion, or zones of proximal development, activate the "complementarity of representational and instrumental aspects" (Otte 1980, 64) of such conceptual objects. The representational concept, as a static and uncritically accepted frame, must be transformed into an instrumental concept, critically reflected, molded and applied, and back to a new representational frame. (p.185)
Engestrom draws on Ilyenkov's account of dialogical logic, filtered through Davydov (pp.190-191). In dialectical logic, the concrete is the interconnected whole, while internal contradictions are at the core of the interconnections (p.193). Again relying on Davydov's interpretation of the source material, Engestrom recounts the method of ascending from the abstract germ cell to the concrete generalization (p.195).

After discussing models, Engestrom argues that there are "four historical types of activity": craft activity, rational activity, humanized activity, and a newly emergent "collectively and expansively mastered activity" (p.224). That is, he provides a brief historical account of transitions between activity types. But how do we understand these transitions? Engestrom nominates dialectics, arguing that
The dialectical method is a method of grasping the essence of the object by reproducing theoretically the logic of its development, of its historical 'becoming'. The dialectical method is thus a historical method. But it is also a unity of the historical and the logical.  The history of the object is purified of its arbitrary details, it is elevated to the level of logical succession from which the details in their full richness may again be derived, now 'with the meaning they have thereby acquired'.
...
 In my analysis, dialectics is the logic of expansion. And expansion is essentially a social and practical process, having to do with collectives of people reconstructing their material practice. (p.242)
But (and?) for understanding sociality, he turns to Bakhtin's dialogism:
The new sociality envisioned here is one of heteroglossia and polyphony, orchestrated and organized around a common object. Borrowing from cognitive science, one could perhaps speak of parallel distributed processing systems. An evolving activity system socially based on such parallel distributed modules could be conceived of as a local or global paradigmatic network  of groups and individuals sharing a common object/motive and common instruments. 
But how would such a social structure differ from the classical idea of a community of scholars, or from an invisible college of related research groups? We get advice from Bakhtin: "the novel must represent all the social and ideological voices of its era, that is, all the era's languages that have any claim of being significant; the novel must be a microcosm of heteroglossia" (Bakhtin 1982, 411). Applied in expansive learning and research, this means that all the conflicting and complementary voices of the various groups and strata in the activity system under scrutiny shall be involved and utilized.  As Bakhtin shows, this definitely includes the voices and non-academic genres of the common people. Thus, instead of the classical argumentation within the single academic speech type, we get clashing fireworks of different speech types and languages. (p.247)
Engestrom doesn't go into how these "clashing fireworks" resolve with the developmentally oriented, Marxist dialectics in the previous section.

In Ch.5, "Toward an Expansive Methodology," Engestrom discusses his cycle of expansive transition.
It traces the generation of socio-culturally new activity systems by collectives of concrete human beings. Here, individually manifested doubt, hesitation and disturbance is the starting point. The direction is from the individual to the societal.  However, the individual point of departure is itself understandable only as a cultural-historical product. (p.252)
Engestrom says that one must delineate the activity system by first gaining preliminary phenomenological insight (p.253): "expansive research is not dealing with activities 'in general' but with real activities realised by identifiable persons in identifiable locations. Delineation is this very act of identifying the personal and geographical locus and limits of the activity. The reason for putting delineation after phenomenology is obvious. Often the locus and limits of activity can be properly defined only after a relatively extensive  'dwelling' in it" (p.254).

Once that is done, the second step is to rigorously analyze the activity system through " (a) the object-historical analysis, (b) the theory-historical analysis, and (c) the actual-empirical analysis" (p.254).

The third steps is to encourage the activity's participants to formulate "qualitatively new models as  genuine keys for resolving the double bind" (p.256). Then
The new instruments can only be implemented in selected strategic tasks. Such tasks represent the points of probable breakthroughs into the qualitatively more advanced form of practice. In carrying out these tasks with the help of the new instruments, the participants of the activity system face intense conflicts between the old and the given new ways of doing and thinking - the tertiary contradiction. (p.261)
Finally, one reports the results.

Engestrom ends the chapter by recentering us on the mission of expansive developmental research:
Expansive developmental research aims at making cycles of expansive transition  collectively mastered  journeys through zones of proximal development. In other words, it aims at furnishing people with tertiary and secondary instruments necessary for the mastery of qualitative transformations of their activity systems. (p.263)
That's it—a book with just five chapters, each of which is lengthy and dense. If you are interested in activity theory, particularly Engestromian activity theory—if you are working in HCI, CSCW, writing studies, or other areas that use activity theory—if you read articles in which people present triangles and call them "activities"—obviously you should (still) read this book.