Social Activity Theory

Paul Dowling

Culture Communication & Societies

Institute of Education

University of London

Four criteria for the evaluation of a theoretical apparatus

I am presenting here an apparatus or a protocol for the analysis of social action. It has been developed through a form of dialogue between the theoretical and empirical that I refer to as constructive description (Dowling, 1998, 1999a, 2000; see also Brown & Dowling, 1998)[1]. That which is constructed is, on the one hand, a description of one or more empirical sites (Dowling, 1996, 1998, 2001a, 2001b, Dowling & Brown, 1996, 2000, Dudley-Smith, 2000, Ensor, 1999, Sunnen, 2000) and, on the other, the apparatus that I am referring to here as social activity theory (Dowling, 1998). The use-value of the theory lies in its enabling of descriptions of diverse empirical sites in the same terms or in the same language of description (Bernstein, 1996). In a general sense, any theory will do for this purpose. However, to the extent that we are looking for coherence, reliability and delicacy in our descriptions, then these qualities must be facilitated by the theory. Furthermore, it is the theory that is going to enable us to make generalisations. It is therefore crucial that the constructions of the theory—its objects—bear directly on the theoretical region about which we intend to make generalisations; if we are interested in language, then a linguistic theory is appropriate. Such a theory is not appropriate, however, if we wish to make generalisations about the social. In this case, we must have recourse to a sociological theory and it is this kind of theory that I am presenting here. None of this entails, however, that a theory must reside wholly within a specific discipline (even supposing such entities to be meaningful). It is the objects ultimately constructed by the theory and not necessarily their antecedents in other work that must clearly be located in the relevant region of generalisations.

I have described the theoretical apparatus as having been generated out of a dialogue between the theoretical and empirical. This is an ongoing dialogue. The apparatus, on the one hand, and the descriptions that it produces of specific empirical sites are both in a continuous state of development. Revisiting already described sites enables the elaboration of their descriptions in terms of completeness and coherence. Exploring new settings generates new descriptions. In either case, there must be a certain openness of the theory to the possibility of its own development. The process might be describe as equilibration, in the Piagetian (1980, 1995; Dowling, 1998) sense: the revisiting of a site or the visiting of a new site in terms of the apparatus constitutes a disequilibrating move; analysis proceeds in terms of the restoration of equilibrium, via the Piagetian processes of assimilation/accommodation, specialisation of schemes at any given level of analysis, and reglobalisation at a higher level. This is a learning apparatus. However, a crucial distinction between a theoretical apparatus and an organic cognitive apparatus is that the activity of former but not that of the latter can be halted. It is legitimate, therefore, to think in terms of a current state of the theory. Following the analogy with Piaget’s epistemology, we might further expect a highly developed, mature theory to be characterised by a heightened level of inertia in respect of its own learning by comparison with a less developed (or juvenile) theory. Ultimately, this may lead to effective sclerosis. At that point, the theory is no longer learning but can only impose its structure on the descriptions of empirical sites. Arguably, this describes the state of school mathematics as a theoretical apparatus in relation to the domestic and other non-mathematical sites that it describes (see Dowling, 1994, 1995a, 1995b, 1996, 1998, 2001a). In each phase, however, the ‘current state’ of the theory stands to a degree apart from its empirical sites. Under these conditions, the power of the theory is maximised, in terms of its potential range of application, to the extent that its theoretical objects are not fixed in terms of level of analysis. The theory is not, in other words, a substantive theory about its referential field so much as a mode of working. It is for this reason that I have used the terms apparatus or protocol and, elsewhere (Dowling, 2000), machine.

I am now in a position to introduce four criteria whereby the use-value of a particular theory is to be assessed. Firstly, there should be a maximal level of internal explicitness in the theory in order to render as visible as possible the recontextualising effect of its description of the empirical site. This is the key feature that marks out theory from, shall we say, ideology. Secondly, there should be a maximal level of coherence or relational completeness in the theory. This is clearly a prerequisite of coherence of empirical description. I must emphasise that whilst we may be aiming at perfection in respect of these two criteria, there is no claim that this is ever achieved. There are inevitable lacunae and imprecisions and disjunctions within any theory and it is here that the silence of the Lacanian (1977) unconscious or, perhaps, Althusser’s (1972) ideology may be heard to speak. Shall we say that the test is that of the shibboleth; to what extent is it possible to apprentice an other into reliable deployment of the theory? The third criterion is that which I made more or less explicit above, that is, the objects constructed by the theory should relate directly to the field within which generalisations are to be made.

The fourth criterion derives from my proposed detachment of the theoretical apparatus from any fixed level of analysis and entails that the theory should be constituted as fractal in nature. This has two possibly unexpected but exciting implications. Firstly, the theory can take itself as its own empirical object. The fractal criterion, then, is satisfied only by the generation of an isomorphism between, firstly, general methodology—the interpretation of the nature of what the theory is or does, in my case, that which is referenced by the term constructive description—secondly, the form of the theory itself—social activity theory—and thirdly, the specific descriptions of empirical sites that the theory generates. Secondly, the theory is able to describe the relationships between the descriptions that it generates at different levels of analysis of a given site in formulaic terms. Specifically, given a description (d) at one level of analysis (signified by d1) a description (d2) of another level may be appended serially (d1+d2) or may contextualise (d2(d1)) or be contextualised (d1(d2)) by the first level. I refer to the first arrangement as a serial configuration of levels and the second and third as concentric configurations. In general, the description of a site comprising a complex structure of levels of analysis might look something like this: (dl+ dl+1)(dm(dn+ dn+1+ dn+2+…)+ (dp+…)(dq+ dq+1+…)+…). A feature of this system is that the additive operation (serialising) is commutative and associative and the multiplicative (centring) is associative but not commutative and distributes over serialising. This description is an example of the referring of one theory—the apparatus being introduced here—to another—in this case deriving from elementary mathematics. As I have indicated, the apparatus (and the more so its mathematising in this way) is appropriately construed as a mode of operation upon rather than a substantive theory about (in this case) the sociocultural. This being the case, the question of precisely what serialising or centring might constitute is—at least at this stage—empirical.

The objects of social activity theory

Action constituting positions

My starting point is an understanding of sociology as being concerned, primarily, with patterns of relations between positions. Essentially, these patterns comprise (exclusively) alliances and oppositions; this is the social. Alliances and oppositions are to be construed as established, maintained, and destabilised only in social action, the visible forms of which are cultural practices. Alliances and oppositions and, in particular, their associated actions may be described in terms of their relative strength of institutionalisation. Actions associated with the maintenance of relatively (ie within the context of any given deployment of the apparatus) strongly institutionalised alliances and oppositions are referred to as activities. Actions associated with the establishment or destabilising of relatively weakly institutionalised alliances and oppositions are referred to as strategies. Activities and strategies recruit, and thereby define or redefine, cultural practices as resources in the establishment, maintenance and destabilising of alliances and oppositions.

As will be apparent, the theory constitutes a strategic space. This being the case, the subject is placed at the centre of the apparatus and, in operational terms, the subject is always the subject or author of action—activity or strategy—at the relevant level of analysis. Authorship of necessity constructs, firstly, an audience and, secondly a content or practice. I propose that there are, logically, two modes of relation that can be established (textually) by an action. These are referred to as pedagogic and exchange modes and differ only in respect of the location of the principles of evaluation of the practice. Action in pedagogic mode locates these principles with the author; action in exchange mode locates them with the audience. In addition to mode of relation, action may distribute authority differently between author and audience. Thus pedagogic action clearly locates authority with the author. However, it may construct the audience (or part of the audience) as a potential author; this is apprenticeship mode. Alternatively, pedagogic action may distribute a reduced level of potential authorship to the audience such that the latter has limited access to the conditions of authorship; this is dependency mode. Finally, pedagogic action may distribute no potential authorship to the audience; this is objectification mode. In this last mode my analytic use of the term ‘audience’ is made visible; a fully objectified audience is, in everyday terms, no audience at all, but an absent or alienated audience that is recruited in addressing a present audience. This might apply, for example, in the case of my citation of antecedent publications where their authors are unlikely to be constituted as potential readers of my work.

Exchange action locates authority with the audience. In its weakest mode it attributes no authority to the author; this is compliant mode. Where some authority is distributed to the author we may refer to routine mode (routine sets some limits on the level or range of compliance). The distribution of maximal authority to the author is innovation mode. These categories are summarised in Table 1.



Mode of relation

Level of subjectivity attributed to position alienated from principles of evaluation




(authority with author)




(authority with audience)




Table 1

Distribution of authority in pedagogic and exchange modes of relation

Arranged in this way, the structure produced bears a superficial resemblance to Bernstein’s classification/framing schema (eg see Bernstein, 1996). This, however, is spurious. Specifically, framing, which derives from Goffman and concerns relations of control within interactions, whilst classification, which descends from Durkheim, is concerned with power relations between categories. As Bernstein puts it,

… briefly, control establishes legitimate communications, and power establishes legitimate relations between categories. Thus, power constructs relations between, and control relations within given forms of interaction. (Bernstein, 1996; p. 19)

As I have argued elsewhere (Dowling, 1999b), this attempt by Bernstein to unite Durkheim and Goffman introduces a fatal confusion in his work in respect of level of analysis. Since, as I have indicated above, independence with respect to level of analysis—the fractal criterion—is a crucial feature of my apparatus, the importing into it of Bernstein’s classification/framing schema would constitute a virus in my work as it does in his own. I avoid this by making no attempt to address more than one subjectivity at a time. I am not concerned with modes of interaction—although they may legitimately be constituted as outcomes of the empirical deployment of my apparatus—but rather with action—activity and strategy. As will become apparent below, I do make use of Bernstein’s category, classification, but framing is rejected.[2]

The nature of the authority that is distributed to author/audience (for example, whether it is constituted via the affiliation of the author to an acknowledged authority or via the identification of the audience to an acknowledged audience position, such as student or client) is an issue of cultural resources. I will discuss some of these alternatives later.

Action constituting practice

I constructing social activity theory, I am aspiring to the level of self-referentiality that characterises mathematical systems. It is precisely this feature of mathematical systems—I note that Deleuze and Guattari (1987) describe mathematics as a ‘monster slang’—such as the complex numbers or topology that constitutes their power to describe potentially all empirical sites. However, in doing so, they say nothing in particular about these sites unless the descriptions themselves have become implicated into their practices. The incorporation of arithmetic into commerce and other cultural practices would be an example of the latter. Indeed, it is often claimed that mathematics has its origins in such practices and only subsequently becomes detached in the constitution of an academic mathematical system (see, for example, Kline (1972) also, in a different vein, D’Ambrosio (1985)).[3] Elsewhere (Dowling, 1998), I have described the manner in which school mathematics and some of its proponents mythologise non-mathematical practices. The myths arise out of the casting of a mathematical gaze onto non-mathematical practices which are thereby redescribed or recontextualised. The myths propose that mathematics is about (myth of reference) or for (myth of participation) or embedded within (myth of emancipation) everyday non-mathematical practices thus in some way laying claim to an essential or original meaning to mathematics. This is an example of an ontologising strategy. Its import is the denial of the productivity or artifice of recontextualisation. Clearly, such action must counter the kind of claim self-referentiality that I want to make in relation to my own theoretical apparatus. My approach, then, is to disclaim all three myths in relation to this apparatus which is to be constituted as simply a protocol for the construction of descriptions of empirical sites. These descriptions are acknowledged to be recontextualisations and therefore not, in the strictest sense either about or for or embedded within the sites. The ontological comportment of my own action is clearly that of de-ontologising. As with the above description of the relationship between mathematics and commerce, there is no denial of the origins of my apparatus within engagements with empirical sites. Indeed, I have explicitly embraced this relationship in my introduction above. However, the application of generalising strategies have constituted at least a degree of self-referentiality, the extent of which is a measure of the power of the apparatus.

Ultimately, it is my claim that there are, in principle, no possible empirical sites which are a priori beyond the description of the apparatus; provided that subjectivity can be attributed at some point. The issue would be regarding the usefulness of applying this particular theory in any given case; it is not immediately obvious, for example, that a description of the physical universe as a social system would be of particular interest to physicists.[4]

The intention of social activity theory is clearly the construction of a theoretical apparatus, albeit a de-ontologising construction. In addition to mathematics as it has been described here, we might interpret the cyborgs proposed by Haraway (1991) and other postmodern constructions as de-ontologising and, in Haraway’s case, pragmatic constructions. We might interpret the construction of other apparatuses, such as those generated by physics or many of the more substantive theories in the social sciences such as those proposed by Piaget or Durkheim, as ontologising constructions. By contrast, the action of critique proposed by authors such as Freire (1970) seem to be characterised by the opposite vector. They do not operate via the construction of alternative systems of power so much as the dismantling of existing ones via localised pedagogies of conscientisation. These actions are ontologising insofar as they lay claim to an effectivity within a region of practice that is beyond their own discourse. In effect, they propagate the myth of emancipation (Dowling, 1998). Alternatively, the deconstructive action of, for example, Lacan (1977) involve a perpetual dismantling of their own constructions—Lacan’s oeuvre is, it would seem, a paragon of self-conscious incoherence—are clearly de-ontologising. At this point, I am able to summarise the practice constituting activity/strategy thus far introduced in Table 2.

It will be noted that I have tagged the ontologising and de-ontologising comportments ‘modernist structuralism’ and ‘postmodernist poststructuralism’, respectively and that this constitutes my own theoretical apparatus as well as the ‘monster slang’ of mathematics as ‘design-oriented postmodernist poststructuralism.[5]


Ontological comportment









Modernist structuralism




Postmodernist poststructuralism

Table 2

Action/strategy constituting practice by ontological comportment and vector

I have mentioned earlier that I incorporate, mutadis mutandis, Bernstein’s category, classification, in my apparatus. Specifically, classification action constitutes practices that vary in terms of their distinction from other practices; they mark out a territory.[6] Where classification action is strong, the territory is clearly demarcated; where it is weak, the territory is less clearly demarcated. I have fount it to be productive to make an analytic distinction between the content of practice that is constituted and its expression or, in Saussurian terms, between signified and signifier. Thus classification action can be conceived as operating independently on content and expression, giving rise to four domains of practice-constituting action. The principal domain is the esoteric domain. This is the most strongly demarcating region. The domain which is most weakly demarcating is the public domain. In both of these cases action operates on content and expression in the same way. Action that tends to classify content strongly but expression weakly constitutes the expressive domain; action that tends to classify content weakly but expression strongly constitutes the descriptive domain. These domains of action are represented in Table 3.

The esoteric domain is clearly the region of action within which stakes are highest, so to speak. In particular, the authoring of esoteric domain action constitutes—and here I shall say (re)produces to index the dialectic of production and reproduction that is characteristic of at least my design-oriented postmodern poststructuralism—what the practices that stand as generative resources for the further (re)production of a given alliance/opposition. Apprenticing pedagogic action must, at some point, provide access to this domain. The descriptive domain is the region of action within which a given alliance/opposition visibly recontextualises the practices of another alliance/opposition; the content refers to this recontextualised alliance/opposition, but the expression is more or less consistent with the language of the recontextualising alliance/opposition. This is the region in which my own theoretical apparatus reports its descriptions of empirical sites. Action in the public domain also recontextualises another alliance/opposition. However, here, it does so invisibly. From the perspective of the recontextualised alliance/opposition, action in this domain must look simply like error. The expressive domain is the region of action in which the content of a given alliance/opposition is recoded in the expression of the public domain. An illustration of action within this domain would be the attempt by a player (or walkthrough writer or, indeed, a compiler) of a computer game to recode the language in which the game is written—the machine code of its esoteric domain—in terms of the public domain moves permitted to avatars and so forth (see Sunnen, 2000).


Expression (signifiers)

Content (Signifieds)




Esoteric domain

Descriptive domain


Expressive domain

Public domain

Table 3

Domains of action

In presenting my apparatus, I am deploying action that tends or at least attempts to render the principles of my theory available within discourse. I am also deploying action which illustrates the application of the theory by producing specific examples of its descriptive artifice, on the one hand, or which detach it from such specifics so as to extend its power, on the other. These latter two modes of action respectively delimit and extend the power of the apparatus; I shall use the terms specialising and generalising, respectively, to refer to them. Other alliances/oppositions tend to be (re)produced via action that tends to privilege context dependency over the kind of context independency that I am attempting to establish here. Most commonly, context dependency is associated with manual as opposed to intellectual forms of practice, although we would expect most if not all empirical instances of alliances/oppositions to deploy both modes of action. I refer to action tending to make the principles of an alliance/opposition available within discourse—to render it relatively context-independent—as actions of high discursive saturation (DS+) and actions that tend to present context-dependent exemplars as actions of low discursive saturation (DS-). I refer to DS- power delimiting and extending as localising and articulating respectively.

Where a practice-constituting action is interpreted as activity, that is to say, where the level of institutionalisation is relatively high, it contributes to a repertoire of practice that we can understand as a competence. This is because the claim to a high level of institutionalisation is tantamount to a claim that the practice is shared by multiple subjects. On the other hand, where the action is a strategy, the weak claim to institutionalisation establishes its contribution to a repertoire as performance. There are two modes of repertoire at each level of institutionalisation. DS+ competence is referred to as discourse, whilst DS- competence is skill. Performances are presented as more idiosyncratic, so that a DS+ performance is an ideolect and a DS- performance is a trick. These categories are summarised in Table 4.










High (DS+)








Low (DS)








Table 4

Action of discursive saturation, power and repertoire

Action articulating positions and practice

Finally, in terms of action, I want to consider the modes of authority action as they relate to the level of institutionalisation (activity/strategy) and the openness or closure of the category to which membership is being claimed by the author. Firstly, action that tends to close the category within the context of a high level of institutionalisation of the alliance/opposition bears a resemblance to Weber’s (1964) traditional mode of authority and I shall use this term for this action. Here, authority is vested in the person of the authorship within the context of a codified discourse or skill. The opening of the category of authorship within this context aligns, in effect, with Weber’s bureaucratic authority. Here, authority is vested in the category occupied by the author, persons being interchangeable. Where the category of authorship is closed in the context of a weakly institutionalised alliance/opposition and authority is vested in the person of the author in the context of an ideolect or trick that is specific to that person, we have something that looks like Weber’s charismatic authority. This exhausts Weber’s ideal types. However, the space established by my product of two binary variables—category of author and institutionalisation—includes a fourth mode of authority action. Here, authority is vested in an open category of author, but the institutionalisation of performances is weak. Relabelling the institutionalisation variable may make this situation a little clearer. Weak institutionalisation entails an openness in relation to the practice—there is, in other words, comparatively weak constraint on what might be said/done, this is why I refer to performance rather than competence. Strong institutionalisation, on the other hand, entails comparative closure of the practice, which is why I refer to competence rather than performance. The new mode of authority action is characterised by an openness of both the category of authorship and performance. This is the kind of authority claim that might be made by a client in response to a compliant exchange action. In their response, of course, the client is now constituting a pedagogic action because they are claiming authority. I have used the term client to refer to this mode of authority action. The four modes are summarised in Table 5.


Category of author








Vested in the person




Persons interchangeable

Table 5

Modes of authority action

Before moving on to consider resources, I need to establish, in part retrospectively, a mode of reading the schema and, in particular, the tables that I have presented above. Firstly, I have defined the orientation of my theoretical apparatus as strategic in nature. This marks it out from, for example, the kind of apparatus constructed by Bernstein which establishes categories of state—a pedagogic situation, for example, may be described within his scheme as strongly classified and framed. I have also introduced a fractal criterion which establishes that the theory may legitimately become its own empirical site. Thus the constitution of cells in the tables should not be regarded as locations in the conventional sense. Rather, the theoretical and empirical constitution of action as, for example, more charismatic or more traditional is to be regarded as a strategic move in the constitution of alliance (charismatic with charismatic) and opposition (charismatic versus traditional).

Secondly, I have already established, again via the fractal criterion, the centering and serialising modes of relating empirical sites to each other. Now the conventional approach to the use of ideal types would be to describe specific empirical instances as particular combinations of the established types. I am happy to refer to my theoretical objects as ideal types—given that they are not simply locations—it is unlikely to be helpful to describe a given empirical site as exhibiting purely exchange action, for example. However, the fractal criterion enables us to go further than the simple combination of ideal types. Thus the centring of sites enables us to combine types by moving up and down in terms of levels of analysis, whilst serialising allows us to partition a site (or to append additional sites) at a given level of analysis.

I shall offer an example. It is often claimed that the move from traditional to digital text media entails a transformation from linear to non-linear modes (see, for example, Cockerton & Schimell (1997) and, from a different perspective, Luke (1997)). Clearly, the traditional novel, for example, may deploy diverse ordering resources such as page and chapter numbers, adherence to traditional narrative structures, anaphoric binding at sentence, chapter and above levels of analysis, the recruitment of linear time and so forth. The recruitment of such resources may be interpreted as pedagogic action because they tend to assert a linearity in respect of the manner in which the text is to be read; in other words, they tend to locate the authority over the reading with the authorship. On the other hand, many websites display few of these ordering resources, tending to privilege resources of navigation including, in particular, hypertextual hotlinks. In this respect, we may refer to exchange action in passing authority over at least the ordering of the reading over to the audience. However, the generalisations that enable the classifying of traditional media as one mode—linear—and digital media as another—non-linear are here based on stereotypical images of these media; the traditional novel, on the one hand, a heavily hypertextual webpage, on the other, for example.

If, however, we consider another category of traditional media text we may get a rather different picture. Consider The Compact Nelson Japanese-English Character Dictionary.[7] We can consider, firstly, the level of analysis that focuses on the macro-structure of the book. As with the novel, the book recruits ordering resources from the natural numbers. These are quite diverse, but the principal ordering is achieved by resolving each character—they are called kanji—into two components, the ‘radical’ and the ‘residual’. The first principle of classification is in terms of the radicals of which there are 214. These are arranged in order of number of strokes that are conventionally taken to write them by hand. Thus the radical (kuchi, mouth) is the first 3-stroke radical and comes after (mata, again) which is the last 2-stroke radical. In the left and right hand margins of each page are printed all of the radicals having the relevant number of strokes, for example, on the pages relating to there are thirty-four radicals listed in the margins with marked as the current one. Supposing we wish to look up the kanji, . We must first recognise as the radical (a small and elongated can be seen on the lefthand side of the kanji). We then count the residual number of strokes, in this case five. We can locate (MI/aji, taste) under kanji having radical and five additional strokes; it happens to be the second head kanji in this category.

It is of interest, here, that the ordering resources constitute a skill in respect of a prior apprenticeship into writing kanji which enables the counting of strokes. To the uninitiated, might seem to be composed of three strokes and of four. However, Japanese children are trained to write the upper across line of and its right-to-left downward line as a single stroke and similarly the upper across line and righthand downward line of . Additionally, the audience is constituted as competent in respect of recognising the radical—comparatively straightforward here but not always so. An alternative listing is provided that lists each of the head kanji in the dictionary under any radical that might plausibly be interpreted as the appropriate one. Thus 味 is listed under the radical (ki, tree) with four residual strokes as well as under with five.

As I have indicated, the traditional novel recruits ordering resources to pedagogic action which privileges linear reading. However, whilst The Compact Nelson Japanese-English Character Dictionary also recruits ordering resources which structure its construction, here they function not to privilege linear reading, but to facilitate navigation by the audience. In other words, we should interpret the action as exchange and not pedagogic.

If we now drop down a level of analysis and look at the specific entry that we have just located, we find the following on page 113:


MI taste, flavour, dash;

Touch, tinge; counter for

foods and drinks. aji(wau) taste;

appreciate, experience. aji, aji(wai) taste,

flavor, aroma; zest; experience; tinge. aji

na clever, witty, smart.

The rather complicated, fraction-like object following the kanji relates to another navigational system. Following this, however, the entry is offering a short list of definitive meanings for the kanji, first in the ON pronunciation (MI—conventionally capitalised) and secondly in the kun (aji—conventionally in lower case) pronunciation. The ordering of ON and kun readings is sustained throughout the dictionary, facilitating navigation and the definitions are available for selection without, in this case, any clear privileging of one meaning over another. However, the entry at this level is characterised by pedagogic action because, by providing a definition—albeit constituted as a range of possible synonyms—the author is establishing authority over the audience that here is constituted as dependent because, unlike some other dictionaries, there is no inclusion of etymology or other action that might apprentice the audience into the principles of interpretation.

This example stands as an illustration of the centring of one level of analysis in relation to another; the exchange action of the navigational strategies contextualises the pedagogic action of the entry. Clearly, we might move to a different entry which we would expect to follow the pattern of this one. The result, however, would be to centre two serialised pedagogic actions within an exchange action. This might be represented, symbolically, as E(P1+P2). I have clearly simplified the construction by ignoring the exchange action within the entry (eg the navigational fraction), but these could clearly be incorporated into the description.


The examples of the novel kanji dictionary introduced the category of resource into the apparatus. Both the novel and the dictionary incorporated ordering resources, including resources recontextualised from the mathematical system of natural numbers. However, although similar resources were recruited, it became clear that the mode of action in each case was different: pedagogic action privileging linear reading, in the case of the novel; exchange action facilitating non-linear navigation, in the case of the dictionary. Resources, in other words—even resources as highly structured as the natural numbers—do not determine the mode of action to which they may be recruited. Furthermore, nor are given modes of action constrained in their selection of resources. Exchange action relating to dictionaries in modern languages other than Japanese and Chinese deploy alphabetical ordering rather than numerical ordering in facilitating navigation. Thus we can posit, analytically, the existence of a reservoir of resources. Action selects from this reservoir but, in doing so, recontextualises the resources selected so that they align with the mode of action. This is neither bricolage nor science, in Lévi-Strauss’ (1972) pure terms. The reservoir comprises the resources that (notionally) lay to hand. However, their recruitment is always transformative so that within the context of an empirical action, they are rendered as purpose-built.

Of course, the reservoir is not being understood as a substantive reservoir of resources awaiting selection. That which might be recognised as a resource is always already contextualised by an action. In recruiting a resource, the subject of an action metaphorically casts a gaze onto other actions which are thereby recognised in terms of the gazing alliance/opposition and recontextualised. The advantage of thinking of resources in this way is that it enables us to describe the relationship between alliances/oppositions, which is to say, the manner in which one regards another. This is most clearly productive in respect of the formation of the public domain. Elsewhere (Dowling, 1996, 1998), for example, I have described the recontextualisation of domestic settings in and by school mathematical action.

In Dowling (1998) I have also described the recruitment of the semiotic resource, signifying mode in localising action. Here, I have argued that the recruitment of a conventionally representational image or icon can be interpreted as privileging a reading which entails a strong visual code of presence. In other words, the image is constituted so as physically to position to audience in relation to that which is signified. This would enable a difference to be signified between two images of police officers measuring car skid marks on the road where the audience is differentially located in relation to the scene. To the extent that this semiotic code is characterised by a strong modality—a strong claim to the reality of the event signified—these images may be recruited to localising action. I have also considered the ways in which the representational image may be scaled in terms of the level of disruption of the visual code of presence. In this schema, a representational photograph may be interpreted as minimally disrupting the code (and so potentially maximally localising), a drawing (which does not incorporate exaggeration or other tropic devices) as somewhat more disrupting of the code and a cartoon (employing tropic devices) as maximally disrupting of the code and so comparatively weakly localising.

Again, the form of the resources does not determine the nature of the action that recruits it. For example, I found that some apparently representational photographs in one of the school textbooks that I was analysing were recruited in such a way as to signify, not their content—a duck—but the relationship between different instances of the same photograph under enlargement. Thus the icon was not strictly being deployed through the privileging of the code of presence. I also noted that cartoons may signify other than their apparently representational content, for example, they may connote juvenile or humorous literature, again not privileging the code of presence in their reading. These examples illustrate the comparative arbitrariness of resources in relation to action and also insert a caveat to any ill-considered attempt to move too fast from qualitative semiotic analysis to quantitative content analysis; in the former approach, analysis must be elaborated and rendered properly context-specific.

The final issue that I want to consider here is the recruitment of audience voices. In Dowling (1998, see also Sunnen, 2000) I introduced the notion of a displaced author. This is a form of positioning action in which the authorial voice affiliates to the addressed audience voice either tacitly or explicitly. It might occur, for example, where a school textbook incorporates jokes etc that operate at the expense of the apparently pedagogically privileged content; a mathematics textbook representing school mathematics as tedious or impossibly difficult, for example. A similar situation occurs in a current advertising campaign for Skoda automobiles in which the authorial voice appears to recognise the poor reputation of the company. Under such circumstances, the author can be interpreted as siding with the audience and, in consequence, shifting a pedagogic text in the direction of exchange relations. Clearly, affiliation might operate in the opposite direction where, for example, the authorial voice affiliates to a high status objectified audience voice (again, some of my citations—such as that of Max Weber—might be appropriately interpreted in this way. Similarly, objectified audience voices may be recruited by action that identifies them with the addressed audience. One of the images of police officers referred to above, for example, locates the audience squatting at the shoulder of one of the officers who is also squatting down with the tape measure. I have interpreted this as constituting an identification of the addressed audience voice with the police as a possible career trajectory; this interpretation is supported by other features of the text.


My intention in this paper has been to introduce the principles and elements of social activity theory as a theoretical apparatus. I have interpreted the apparatus as a protocol for the description of empirical sites. The apparatus and potential empirical sites are to be understood as an equilibrating system, which is to say, it is able to develop or ‘learn’ in its repeated engagement with the empirical. The general methodology of this equilibration has been described as ‘constructive description. I have introduced four criteria which the theory must attempt to satisfy. These are: i) the criterion of explicitness; ii) the criterion of coherence; iii) that the theory must be located in the region in which general statements are to be made—in this case the social; iv) the theory must be maximally fractal in nature. This last criterion entails that: a) the theory can take itself as its own empirical object, thus generating a potentially high degree of coherence between the levels of general methodology, the conceptual form of the theory and the specificities of the empirical descriptions that it generates; b) the theory is able to describe the relationships between different levels of analysis in respect of a given empirical setting (or arrangement of settings) in symbolic terms as a combination of centring and serialising arrangements.

I have then introduced the main conceptual apparatus of the apparatus which is conceived as strategic in nature. At the most fundamental level, the social is construed as a terrain in which action is exhaustively defined as attempts to form alliances and oppositions. Like the theory itself, it is useful to consider the ‘state’ of a given opposition/alliance at a given point, although it is crucial to understand that this ‘state’ is in the permanent (equilibrating) state of (re)production via action, the latter being understood as either activity—where the alliance/opposition is interpreted as being at a comparatively high state of institutionalisation—or strategy—where institutionalisation is comparatively low. The rest of the paper has been concerned with the detailing of the apparatus in terms of action constituting positions, action constituting practice, action articulating position and practice, and the more arbitrary category of resources.

In summary, my claim is that this apparatus is sufficiently detached from any specific empirical site that it is potentially of considerable power in respect of its range of potential sites. I am also claiming that the apparatus is of sufficient level of complexity and coherence to render its empirical deployment of potentially great interest in providing new ways of thinking about cultural practices.

I want, finally, to make explicit a crucial feature of my approach that is quite central to but only implicit in the discussion to this point. I have established that subjectivity is central to the apparatus and that the apparatus itself exhibits a fractal-like invariance with respect to level of analysis. Consider, then, a specific description of an empirical site—say the Japanese dictionary. Here, I have, in effect, described the relationships that are (to be) established between a textual author and textual audience—these relations constitute a specific combination of pedagogic and exchange modes. However, this description is the artifice of a recontextualising gaze that has been cast onto the dictionary text from the esoteric domain of social activity theory; I have generated a descriptive domain action and this has extended the power of my language in respect of its public domain sites. In this respect, my theoretical apparatus has learned; what it has learned about, so to speak, is itself. This is to say, in general terms, it has ‘learned in respect of its potential to satisfy more fully the criteria that I have laid out for its evaluation. Because the description of the dictionary is a recontextualisation and because we can have no measure of the extent of the impact of this recontextualisation, what the theory has not learned about is the dictionary. It will doubtless be unsurprising to some that the struggle to establish a dualism of theory and theorised has resulted in no more than a radical monism. Michel Foucault famously (or perhaps I only imagined it) described his archaeologies as fictions; insofar as my approach is embraced, then it itself embraces the conclusion that the only narratives that it can ever relate are autobiographies. だから?



Althusser, L., (1971). Lenin and Philosophy, London: New Left Books

Bernstein, B. (1996). Pedagogy, Symbolic Control and Identity. London: Taylor & Francis.

Bernstein, B. (1999). “Vertical and Horizontal Discourse: an essay”. British Journal of Sociology of Education. 20. 2: pp.157-174.

Brown, A.J. & Dowling, P.C. (1998). Doing Research/Reading Research: A Mode of Interrogation for Education. London: Falmer Press.

Cockerton, T. & Shimell, R. (1997). “Evaluation of a Hypermedia Document as a Learning Tool.” Journal of Computer Assisted Learning 13:133-144.

D’Ambrosio, U. (1985). “Ethnomathematics and its Place in the History and Pedagogy of Mathematics.” For the Learning of Mathematics 5:pp. 44-8.

Deleuze, G. & Guattari, F. (1987). A Thousand Plateaus: capitalism and Schizophrenia. London: The Athlone Press.

Dowling, P.C & Brown, A.J. (2000). “A Grand Day Out: towards a mode of interrogation of non-school pedagogic sites”. The Curriculum Journal. 11, 2. pp. 247-271.

Dowling, P.C. (1994). “Discursive Saturation and School Mathematics Texts: a strand from a language of description.” in Mathematics, Education and Philosophy: an international perspective. London: Falmer.

Dowling, P.C. (1995a). “A Language for the Sociological Description of Pedagogic Texts with Particular Reference to the Secondary School Mathematics Scheme SMP 11-16.” Collected Original Resources in Education 19.

Dowling, P.C. (1995b). “Discipline and Mathematise: the myth of relevance in education.” Perspectives in Education 16:pp. 209-226.

Dowling, P.C. (1996). “A Sociological Analysis of School Mathematics Texts.” Educational Studies in Mathematics 31:pp. 389-415.

Dowling, P.C. (1998). The Sociology of Mathematics Education: Mathematical Myths/Pedagogic Texts. London: Falmer.

Dowling, P.C. (1999a). Interrogating Education: Texts, Social Activity and Constructive Description. Plenary presentation at Centro de Investigação em Educação da Faculdade de Ciências da Universidade de Lisboa, April 1999.

Dowling, P.C. (1999b) ‘Basil Bernstein in Frame: ‘Oh dear, is this a structuralist analysis’. Presented to the School of Education, Kings College, University of London. 10 December 1999.

Dowling, P.C. (2000). ‘A Manifesto for Design and the Charismatic Intellectual’. Presented at Education for Social Democracies: changing forms and sites. Institution of Education, University of London. 3-6 July 2000.

Dowling, P.C. (2001a). ‘School Mathematics in late modernity: Beyond myths and fragmentation’ invited contribution to Mathematics Education Research Group of Australia collection, Socio-Cultural Research on Mathematics Education: An International Perspective. Atweh, B. et al. (Eds). Mahwah: Lawrence Erlbaum.

Dowling, P.C. (2001b). ‘Reading School Mathematics Texts’. In Gates, P. (Ed). Issues in Mathematics Teaching. London: Routledge-Falmer.

Dudley-Smith, R. (2000). A social activity analysis of the Institute of Fiscal Studies ‘virtual economy’ internet site. Unpub. MA dissertation, Institute of Education, University of London.

Ensor, P. (1999). A study of the recontextualising of pedagogic practices from a South African university preservice mathematics teacher education course by seven beginning secondary mathematics teachers. Collected Original Resources in Education 24, 3. Also available as a PhD thesis, University of London.

Foucault, M. (1972). The Archaeology of Knowledge. London: Tavistock.

Friere, P. (1970). Pedagogy of the Oppressed. New York: Continuum.

Haraway, D. J. (1991). A Cyborg Manifesto: science, technology, and socialist-feminism in the late twentieth century. In D. J. Haraway (Ed.), Simians, Cyborgs and Women: the reinvention of nature. London: Free Association Books.

Kline, M. (1972). Mathematics in Western Culture. London: Pelican.

Lacan, J. (1977). Écrits: a selection. New York: W.W. Norton..

Lévi-Strauss, C. (1972). The Savage Mind (La Pensée Sauvage). London: Weidenfeld & Nicholson.

Luke, C. (1997). “Technological Literacy.”

Piaget, J. (1980). “The Psychogenesis of Knowledge and its Epistemological Significance.” in Language and Learning: the debate between Jean Piaget and Noam Chomsky. London: RKP.

Piaget, J., (1995). Sociological Studies, London: RKP

Strauss, C. (1972). The Savage Mind (La Pensée Sauvage). London: Weidenfeld & Nicholson.

Sunnen, P. (2000). “Making Sense of Video Games: a textual analysis of Tombraider II.  Unpub. MA dissertation, Institute of Education, University of London.

Weber, M. (1964). The Theory of Social and Economic Organization. New York: The Free Press.

[1] The terminology introduced in the present paper is not everywhere consistent with the earlier publications cited. This is because what is being introduced here is the latest attempt to maximise the coherence of the apparatus. Most significantly, the category that is here referred to as action was earlier referred to as strategy whilst activity was reserved for a notional structural level of the theoretical apparatus; this level is captured here through the use of the expression alliance/opposition. I have avoided citing papers published before 1994 in order to minimise confusion.

[2] My use of classification also differs from Bernstein’s through my rejection of the concepts of boundary and insulation as misleading (see Dowling, 1999b).

[3] This point is of interest in relation to Foucault’s (1972) attribution to mathematics of the singular status of being the only science to have crossed simultaneously and at its inauguration the thresholds of positivity, epistemologisation, scientificity and fomalisation because the condition of possibility of its positivity was the establishment of a formalised discursive practice. Foucault periodises history nicely (although without actually giving us a date) and denies us the use of ‘mathematics’ to identify proto-mathematical antecedents.

[4] Arguably, the ontologising strategies of physics (motivated, possibly, by the exigencies of funding not to mention the prevalence of the discourse of ‘discovery’ as a metanarrative (pax Lyotard) have denied it the possibility of the development of a generalised apparatus of its own, so that it must borrow from mathematics or the rather clackety apparatus of the ‘scientific method’. It might be amusing to reconceptualise, for example, gravitational and other force relations as pedagogic and, perhaps, ionisation as facilitating exchange relations.

[5] I borrow the term ‘design’ from my colleague Gunther Kress, although he may well want to situate his own work within another cell of this table.

[6] I am not necessarily referring to territorialisation, here; classification may be ontologising or, as in the constitution of the specialism which is my own apparatus, de-ontologising.

[7] Nelson, A.N. (1999). The Compact Nelson Japanese-English Character Dictionar. Abr. J.H. Haig. Rutland: Charles E. Tuttle Inc.