I am contemplating a piece of a wall. It is a piece of the Berlin Wall, in fact (I have a certificate that attests to its authenticity). More about the wall a little later. However, in the year of the tenth anniversary of its destruction, I am reminded of John F. Kennedy’s famously grammatical faux pas in front of the wall and am inclined to pun. It may not be quite the case that the proudest thing a man or woman can say is ich bin ein bernsteinian, but in the same sense that Michel Foucault (I’ve forgotten where) declared that we are all marxists now, all of us in the field of educational studies at the close of the twentieth century inhabit a world that has been profoundly influenced by the thinking of Basil Bernstein. We are all, in some respects, Bernsteinians now.

This being the case, an exposition or even an exegesis of Bernstein’s work seems inappropriate. Rather, I shall explore several responses to it, including my own. In doing so, I shall of course need to have recourse to his own writing and here I shall, in the main, restrict myself to his most recent book, Pedagogy, Symbolic Control and Identity: theory, research, critique (1996) as this volume may be taken to represent Bernstein’s most mature thinking in book form.

In my preparation, therefore, I generated a list of authors whose texts have exhibited some explicit relationship to Bernstein’s work. Contemplating the list, it seemed to me that they might be organised within a two dimensional space. I shall use this organisation to structure my discourse.

The first dimension is scaled according to the strength of voice that is given to Bernstein by the author. Where the voice is strong, the author has clearly been profoundly influenced by Bernstein’s work in terms of the integrity of the latter. Where the voice is weak, terms may have been recruited, but without allowing Bernstein’s own conceptualisations to impose upon their use and development. The second dimension is scaled according to the extent to which the author reproduces (weakly or strongly) Bernstein’s conceptual structures or, alternatively, works with them and other theoretical antecedents in the production of new frameworks or, to use Bernstein’s own term, new languages of description. We might conveniently refer to this dimension as message. Voice and message are, of course, technical terms in Bernstein’s own work, but here they are being used differently. Configuring these dimensions as a cartesian product, we obtain the space that is represented in Figure 1.

I am conceiving of the scaling of voice and message as operating at a level analogous to the interval level of measurement rather than the ordinal form that is suggested by Figure 1. [1] Clearly, voice may be more or less strong or weak and message may be more or less reproductive or productive. The configuration of Figure 1 represents simply a clustering of convenience. In discussing these clusters I follow a Z-shaped course, beginning in the upper right-hand cell so that I can finish with my own response in the lower left.

The vulgarisers

Nearly twenty years ago, now, I was breakfasting with a friend at her home when her young son dipped his hand into the pot of jam that had been placed in the centre of the table.

"Why are you putting your fingers in the jam?" my friend asked, "Wouldn’t it be better to use a knife or a spoon?" [2]

"Why don’t you just tell him to get his hand out of the pot?" I enquired.

"Because Basil Bernstein has told us that we should speak to our children in elaborated code."

Well, there is a sense in which her control strategy exhibits a degree of context independence that would be characteristic of an elaborated orientation. Certainly more so than, for example, ‘Get your hand out of there’. It might be argued, though, that ‘wouldn’t it be better’ tacitly indexes a more or less locally normalised practice rather than constitutes an appeal to a generalised rationality. Clearly, however, the moralising, ‘we should’, in this context, is a misreading of Bernstein. Bernstein’s voice is clearly very weakly realised in this parent’s text. Furthermore, as a fairly recent graduate of a teacher training programme, she appeared to be attempting to implement—to reproduce—a message rather than to construct something new of her own.
Such vulgarising is certainly not excluded from the academy. Only a few years ago a prominent educationalist (who must remain anonymous) announced in a plenary address to a large American conference that Basil Bernstein was one of a small number of academics who had been responsible for the poor performance of many UK schools because he had declared the working classes to be ineducable. It may be that a title such as ‘Schools Cannot Compensate for Society’ (Bernstein, 1971) might alone be interpreted in this way. It does seem evident, however, that this speaker had failed to read the paper with any attempt at understanding; he had failed to give it voice in his own text. As with the elaborated mother, this speaker was not in any sense incorporating Bernstein’s work—even in this misinterpreted form—into his own production. He was merely reproducing what might appropriately be described as a myth.

Vulgarising is, of course, an inevitable consequence of notoriety. It is an interesting and important factor in the circulation of rumour and myth and may have crucial consequences for the career of the author whose work is vulgarised. Because of the lack of any serious engagement with the vulgarised work, it may be very difficult for the originating author to engage with the vulgariser other than by dismissal and this, from their point of view, may be less than ideal. For my purposes, however, vulgarising is the least interesting mode of response because it entails no internal or external development of the work vulgarised. It is sufficient to say (and it is sufficiently well known) that Bernstein’s work has, indeed, been widely vulgarised in the three decades of his prominence in the field. I shall move directly on to the second category of response.

The disciples

I am here referring to authors whose responses to Bernstein’s work constitute applications, explorations and developments of that work and that adopt, to a greater or lesser extent, its central organising principles. Many, but by no means all of the authors that I would categorise here have worked directly with Bernstein, often under his supervision. I will mention, here, just two of the many available authors in this category. These have been chosen in order to introduce two of Bernstein’s key concepts, classification and framing, both of which will be key objects of discussion under my final category, ‘the heretics’.

Classification and framing refer, respectively, to the structural and interactional levels of analysis of the social. For the time being, it is sufficient to describe classification as a measure of the extent to which categories—for example, curriculum subjects—are structurally distinct. Framing, then, is a measure of the degree to which the transmitter of a message maintains control over the communicative context.

Harry Daniels (1988, 1995) was interested in school students’ recognition of and ability to realise specialised school discourses across school contexts exhibiting variations in classification and framing. On the basis of observation and interview data, Daniels first described each of four ‘moderate learning difficulty’ schools in terms of the strength of classification and framing characterising each of a number of aspects of school relationships. These aspects included the relationship between the school and that which was outside school (including parents, employers etc), relationships between subjects, between teachers, and in the classroom, between teachers and students in terms of the instructional and regulative discourses. The result was a complex description of each school. However, a cline could be described between the school exhibiting weakest to that exhibiting the strongest classification (C) and framing (F).

Moving between mathematics and art classrooms in each school, Daniels showed ten pictures individually to a total of eighty children who were asked:

We are in a (Maths/Art) lesson. Your teacher is teaching you about (Maths/Art). What would your teacher like to hear you say about this picture in this lesson?

Daniels, 1995; p. 525)

The pairs of statements for each child were then shown to two teachers, one from a strong C/F school and another from the weakest C/F school. The teachers were asked:

1. Can you tell the difference between these two statements?
2. If you can, which one do you think was made in which context?

(ibid)

Sixteen children from the original sample were subsequently asked the same questions as the teachers with reference to statements made by three other children (one judged to be good, one poor and one intermediate at producing distinguishable statements).

In summary, Daniels found that the ability of the children to realise the specialised discourses was positively correlated with the strength of classification and framing. All but one of the sixteen children in the second phase of the study judged the statements in the same way as the teachers despite having been selected on the basis of their differential performances in the first phase. It would seem, then, that recognition precedes realisation. [3]

This study raises, for me, at least one interesting question. The display of realisation rules by children was evaluated by the teachers. One of these teachers was from a high C/F school, the other from the lowest C/F school. In the case of the first teacher, we might presume that the exemplifying and explicit evaluation of such displays in these terms would be a routine feature of their pedagogic practice. In the case of the second teacher, however, this is unlikely to be the case as such actions are themselves strategies that would tend to produce strong C/F. Indeed, viewed from the perspective of the low C/F pedagogic practice, a request to evaluate students’ displays in these terms might be interpreted as bizarre or, at least, inappropriate. This is because the research design itself is clearly construed from within the perspective of high C/F; it is not reasonable to ask which statement relates to which context under circumstances in which the contexts are not clearly distinguished.

My question, then, is what findings would the research have produced had it exhibited a more balanced design. For example, by avoiding specifying the subject to the students (‘we are in a maths/art lesson’) and by asking teachers to evaluate the responses from the perspective of an undifferentiated pedagogic context—‘what does this response tell you about the student’? Bernstein’s language has clearly opened up the space for both the original research and for my response and provides a means for the systematic interrogation of the nature and productivity of open and closed pedagogies. [4]

The second ‘author’ that I am considering under this category relates to work carried out at the University of Lisbon under the leadership of Ana Morais (Morais et al 1992, 1994, 1999). I shall refer in particular to a comparative content analysis of science curricula relating to the curricular reforms of 1975 and 1991. In this study, the authors coded each sentence of the curriculum documents in terms of their indication of very strong, strong or weak framing. They provide illustrations for each coding thus:

F++ […] which permits one to call the attention of the student to the multiple types of inanimate bodies, of living things and energy manifestations which are part of the environment and, consequently, to sensitize him/her to the relations between the physical and biological world […]
F+ The investigation of the environment […] has the objective of propitiating to the student situations in which he/she is asked to be interested by the environment which surrounds him/her […]
F- To have attitudes of personal initiative, responsibility and decisions […]

(Morais et al, 1999; p. 43)

In general terms, the authors find that there has been a shift towards stronger framing between 1975 and 1991. They suggest that this seems to be counter-intuitive. However,

… if we bear in mind that a given official pedagogic discourse tends to reproduce the dominant principles of society, we can understand the direction of change. In fact, the period which corresponds to the 1975 reform is a period which, in socio-political terms, is characterized, in Portugal, by a period of great change—a turning, as a consequence of the 1974 revolution, from a dictatorial to a democratic society. In 1991, the young democracy of 1975 had changed to a more stabilized democratic system where there is higher equilibrium between the several socio-political forces. The change may also translate the disenchantment with the innovative pedagogic principles of the 1960s and 1970s.

(ibid; p. 51)

Now the first sentence in this statement addresses ‘society’ at a very high level of analysis, that is, at a level at which ‘dominant principles’ can usefully be articulated. Here, the dominant principles are couched in terms of ‘socio-political’ modes: dictatorship; young democracy; stabilised democracy. The implication seems to be that the transition between these three modes is characterised by a shift, firstly, from stronger to weaker framing (dictatorship to young democracy) and secondly from weaker to stronger framing (young democracy to stabilised democracy). Thus official pedagogic discourse—in this case, science curricula—is construed as a mirror of the socio-political mode in terms of framing. This interpretation is, on the face of it, challenged by the publication of The Plowden Report in 1963 (Central Advisory Council, for Education, 1963) at a time when the United Kingdom might reasonably be described as a ‘stabilised democracy’.

Nevertheless, the strong statement by Morais et al points to a potentially fertile region for research. Is the modality of socio-political forms adequately given in the strength of framing realised in official documentation? An initial response to this question coming from the direction of Bernstein’s own work might suggest that the socio-political form is not exclusively concerned with control, but also with power. This being the case, a measure of the realisation of classification in official documentation might appropriately be undertaken alongside the interest in framing.

The final sentence in the above extract suggests the possibility of a degree of upward influence between what Bernstein describes as the ‘pedagogic recontextualising field’ and the ‘official recontextualising field’ in the reflection within official documentation of a disenchantment on the part of teachers, teacher educators and so forth.
I shall return to a further consideration of classification and framing in the section on the heretics. Before that, I want to give brief attention to the exploiters category.

The exploiters

Authors in this category give weak voice to Bernstein’s work which they recruit and exploit in the production of their own message. I am not, in any sense, evaluating this position negatively. However, insofar as this paper is primarily concerned with Bernstein’s voice, a lengthy discussion of work in this category seems inappropriate. This is an important area, however, because Bernstein’s own responses to this work have often tended to be forceful which itself illustrates a feature of the activity of the academic field. In terms of my own language, these authors cast a gaze on Bernstein’s work from the perspective of the esoteric domains of their own respective discourses subordinating the former to the principles of the latter. There is a degree of inevitability about this recontextualising. However, where the recontextualising discourse is sufficiently distinctive with respect to the recontextualised discourse, the result is likely to be constituted as a ‘misrecognition’ or even a misrepresentation by disciples of the recontextualised discourse. To the extent that the recontextualiser’s message hegemonises the field, the situation is clearly disastrous for these disciples. Bernstein and his disciples have clearly recognised this and have, on occasion, made strenuous counter-hegemonic efforts to reclaim the message. In this section I shall simply illustrate three of these responses.

The first is a response by William Tyler (1988) to Ronald King’s (1976, 1981) attempt to put Bernstein’s early theorising of the school to the empirical test. King notes:

Bernstein’s theories have been generated independently of original research, and so are explanations of a presumed reality, which has been shown here to be, in part, empirically false. Since the theories do not adequately explain the actual changes that have occurred in schools doubts must be expressed about Bernstein’s basically adaptive view of educational change, as a response to presumed changes in homologous, superordinate external structures, including the division of labour and the basis of social solidarity. It also suggests that the real existence of the mediators of change, the educational knowledge codes, is at best doubtful, and that their application to the sociology of the school is hypostatic.

(King, 1981; pp 261-3)

This is a very strong challenge that appears to strike at the heart of Bernstein’s message. Tyler, however, constitutes Bernstein’s theory in a form which, he claims, makes it inaccessible to the kind of testing that King applied:

Because Bernstein’s structuralist interpretation of school organisation is so differently conceived from other theories of the school, it does not lend itself easily to conventional empirical testing. Not only does it reconstitute the elements of a theory of school organisation, it also generates its own methodological principles which make any ‘objective’ empirical test to some degree self validating. Since the instrumentation of empirical research is an aspect of coding practices, an appropriate choice of a method could produce a very distorting result. The main danger with such structuralist theories therefore is that they are not testable by the usual empiricist methods which deal by definition with ‘surface’ appearances or phenomena.

(Tyler, 1988; pp. 159-60)

In particular, Tyler accuses King of falling into this trap:

By taking the patterns of correlation as evidence of the existence of the codes, he appears to have missed the whole point of a structuralist approach, that it is not the size of the correlation that matters but its position within a patterned field of such relations.

(ibid. p. 162)

Tyler claims that by choosing the more appropriate ‘canonical correlation analysis’, his own work was ‘highly confirmatory’ of Bernstein’s position. The reclaiming job having been performed by Tyler, Bernstein himself needs only to record it in his own reflection:

I did not personally carry out empirical research on the model, but R. King collected data on a range of schools and then tried to interpret his results in terms of the model. He concluded that there was only very weak evidence for the relations expected (King, 1976, 1981). However, Tyler (1988) criticizes severely King’s statistical treatment of the data.

(Bernstein, 1996; p. 100)

The contrast between ‘tried to interpret’ and ‘criticizes severely’ carries the force of the dismissal.

It is of passing interest that Bernstein neither carried out himself nor supervised the empirical counter to King’s work (Tyler, 1984). However, where an exploiter engages him on theoretical grounds, Bernstein’s own involvement is far more likely. His response to Edwards’ (1987) recontextualising of ‘elaborated code’ and questioning of the empirical basis for the work, for example, was published as a paper in the same journal as Edwards’ original paper and is reprinted in Bernstein’s book (1994, 1996). In his response Bernstein clearly positions Edwards as what I have termed an exploiter rather than as a disciple, which is to say as an illegitimate rather than a legitimate critic, thus:

I shall propose that Edwards’s concept of an elaborated code may suit his narrative but has little place in mine.

(Bernstein, 1996; p. 157)

Edwards’ interpretation of Bernstein’s concept is corrected by the latter’s survey of that of his own work which was available to Edwards.

[Edwards] states ‘an essential defining feature of elaborated codes as Bernstein himself presents them is that meanings are transmitted in ways which give access to the grounds for accepting them and which are therefore open to being challenged. It remains an unusual classroom in which pupils find opportunities for disturbing a body of knowledge.’ Edwards gives as his reference for this and other statements Bernstein (1973, 1977, 1982, 1985, 1986). Now there is no mention of any statement about elaborated codes as Edwards defines them in Bernstein 1982, nor in the expanded versions of Bernstein 1987 (Bernstein 1990). However, in Bernstein 1986 (p. 182) ‘We are arguing that elaborated orientations, and even more elaborated codes are the media for thinking the "unthinkable", the "impossible" because the meanings they give rise to go beyond local space, time, context and embed and relate the latter to a transcendental space, time and context. A potential of such meanings is disorder, incoherence, a new order, a new coherence.’ Thus ‘disturbance’ is a potential of such meanings, not a necessary realization. The meanings are the media, the resource, the condition for the unthinkable, for disturbance. Edwards conveniently or otherwise excludes this sentence from the above paragraph (p. 183): ‘Through its distributive rules the pedagogic device is both the control on the "unthinkable" and the control on those who may think about it. …’

(ibid. p. 158)

These distributing strategies (this is Edwards, this is me) and positioning strategies (‘conveniently or otherwise excludes’) effectively install Edwards in the correct quadrant of my diagram.

Similar strategies are employed in respect of the Edwards’ other major misrepresentation:

Edwards not only disconnects the core concepts but he also perpetuates the myth of the abstract theorist whose work both lacks an empirical base and lacks the power to describe that base adequately.
I might be guilty of some immodesty if I claim that it is unlikely, in the British context of research, that a theory has had such a close interaction with empirical research. What is staggering is either the ignoring or unawareness of commentators, even when (as Edwards) they cite papers wholly devoted to giving an account of research (Bernstein, 1987).

(ibid. p. 157)

An even more forceful strategy is represented in the reprinting in Bernstein’s book of Edwards statement of mea culpa (originally published immediately following Bernstein’s response in the Oxford Educational Review (Edwards, 1994). Edwards’ apology concludes:

I am also persuaded to reconsider whether I have underused if not misused Bernstein’s codes as a uniquely powerful heuristic in the study of pedagogic discourse.

(Edwards, 1994, p. 184; Edwards in Bernstein, 1996, p. 168)

Having thus recanted, Edwards is reinstated as a disciple.

The third exploiter is perhaps particularly dangerous insofar as the authors, Richard Harker and Stephen May (1993) produce a comparison between Bernstein and Bourdieu, coming out very much in favour of the latter. Again, Bernstein published his response in the journal in which the original appeared (Bernstein, 1995) and reprinted it (with a minor alteration to the title and one less appendix) in his book (Bernstein, 1996). Again, I shall simply illustrate Bernstein’s deployment of strategies which begins with an effective positioning of Harker and May as dilettanti within the academic field:

Misrecognition takes a few lines but its exposure takes many.

(ibid., p. 182)

Later in the paper Bernstein introduces a quotation from the forward of a book edited by Bourdieu. The forward was written by Chamboredom and is reproduced by Bernstein in French and without translation. Bernstein’s and his readers’ constructed bilingualism is clearly contrasted with the constructed monolingualism of Harker and May who can quote Bourdieu only in translation.

In the (final) appendix to the paper/chapter, Bernstein lists what he describes as ‘the most puzzling features of the representation of the code thesis’ (Bernstein, 1996; p. 199) again casting Harker and May as dilettanti he notes that:

Clearly these puzzling features must be distinguished from serious, critical engagement without which there can be no development of a theory or of the intellectual field.
There appear to have been three powerful motivations at work in the positioning of the thesis: religious/moral, discursive and epistemological, although sometimes it is difficult to separate the three.

(ibid.; p. 199)

Bernstein proceeds to gloss each of these motivations together with what he describes as ‘an optional field strategy’—‘time warping’—and finishes his book thus:

Now if we relate religious, epistemological and discursive positioning possibilities with temporal possibilities (time warping), we have a truly formidable set of combinations for generating and particularizing practices. We can render this slightly more formally. Religious, epistemological and discursive positioning are category relations, and so spatial and therefore subject to classificatory principles, whereas time warping is temporal, and so subject to framing. Thus we can begin to see how classification and framings regulate positioning and how modalities are a realization of field-constructed motivations.
' Oh dear, is this a structuralist analysis?

(ibid.; pp. 200-1)

I shall refer again to this extract in the next section of this paper. Here, however, it is sufficient to say that Bernstein has succeeded in constituting the Harker and May paper as data in respect of his own message. In the absence of an apology from the authors there can, of course, be no rehabilitation.

The heretics

Heretics, of course, must speak their own message. I shall, therefore, confine myself to my own response to Bernstein’s work. For many years now I have benefited and continue to benefit enormously from serious engagement with Bernstein’s work, both published and, in the past, in terms of personal interactions. My own concern, however, is not the reproduction of this work, but to move from critical interaction with it and with the work of other key theoretical antecedents in the production of my own theoretical and empirical structures. I want here to give an indication of some of the crucial ways that key developments in my own work are related to Bernstein’s ideas.

There is no point in beating about the bush, so I shall begin with the two fundamental concepts of classification and framing. Specifically, I make extensive use of a concept of classification, but rarely refer to framing. Why is this?
The origins of these concepts contribute to the specialising of Bernstein’s own heresy. Classification has its roots in the work of Émile Durkheim and framing in the work of Erving Goffman. In Bernstein’s work, they carry, respectively, the principles of power and control, which is to say:

… 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)

Within Bernstein’s work, the concepts are associated with opposing sets of terms as illustrated in Table 1.

(ibid.; p. 107)

(ibid.; p. 27)

Consider an example which is based on Mark Warschauer’s (1999) observation of an English non-fiction writing course at an American university. Warschauer’s interest was in the ways in which the teacher’s and students’ practices changed as the medium changed between face-to-face (f2f) and computer mediated communication (cmc). In particular, he found that the teacher tended to operate in a didactic lecturing mode in the f2f situation. CMC was described as more ‘democratic’, which is to say, the teacher intervened far less and with more open questions than in the f2f mode and it was the students rather than the teacher who apparently directed the discussions. Now in terms of the above definition, it would seem that the change in practice between the two modes constitutes a weakening of frame. Classification, however, has remained constant insofar as there has been no change in respect of the degree to which this class is to be distinguished from other classes. However, consider this statement by Bernstein:

In the case of invisible pedagogic practice it is as if the pupil is the author of the practice and even the authority, whereas in the case of visible practices it clearly is the teacher who is author and authority. Further, classification would be strong in the case of visible forms but weak in the case of invisible forms.

(Bernstein, 1996; p. 12)

Now the cmc mode looks very much like an invisible pedagogy which here is described as exhibiting weak classification with no reference being made to framing. Elsewhere in Bernstein’s book there is a virtual exclusion of the category ‘framing’ in favour of classification; I noted only a single instance of it in chapter 3, for example.

The source of the confusion, for me, resides in the fact that, as Bernstein himself notes (p. 19), power and control and so classification and framing operated at different levels of analysis. A crucial feature of power relations, for Bernstein, is the construction of boundaries or insulation, thus:

The distinction I will make here is crucial and fundamental to the whole analysis. In this formulation, power and control are analytically distinguished and operate at different levels of analysis. Empirically, we shall find that they are embedded in each other. Power relations, in this perspective, create boundaries, legitimize boundaries, reproduce boundaries, between different categories of groups, gender, class, race, different categories of discourse, different catefories of agents. Thus, power always operates to produce dislocations, to produce punctuations in social space.


[…]


But I want to argue that the crucial space which creates the specializations of the category—in this case the discourse—is not internal to that discourse but is the space between that discourse and another. In other words, A can only be A if it can effectively insulate itself from B. In this sense, there is no A if there is no relationship between A and something else. The meaning of A is only understandable in relation to other categories in the set; in fact, to all the categories in the set. In other words, it is the insulation between the categories of discourse which maintains the principles of their social division of labour. In other words, it is silence which carries the message of power; it is the full stop between one category of discourse and another; it is the dislocation in the potential flow of discourse which is crucial to the specialization of any category.

(Ibid.; pp. 19-20)

This now finally brings me back to the Berlin wall. Now the wall is certainly implicated in the establishing of distinct political regimes. Implicated, but not imbricated. The substance of the wall is suitable for its purpose solely by virtue of its sharing of a predicate with that which it keeps apart. Specifically, the wall and the people on either side of it are mutually impervious. That is to say, in the constitution of a ‘division of labour’—the differentiating of political regimes—the function of the wall is to assert a sameness, not a difference. The same, incidentally, is true in respect of insulation. The plastic material surrounding domestic electrical cable shares the predicate of ‘electrical conductance’ with bodies that it separates.[5] Again, the introduction of the boundary constitutes an assertion of sameness. To take a symbolic example, a full stop—written or spoken—again asserts that the same kind of grammatical object is (or is potentially, in the case of a termination) to be found on either side. [6]

To state the situation in terms of fundamental principles, a boundary is of necessity a moment in the precise region of a system in which it is constituted as a boundary. Classroom walls, then, create punctuations of space not curricular subjects. How do we move from a strongly classified physical space to a strongly classified curriculum? Not simply by labelling the doors—such labels are merely addresses and addresses are like boundaries insofar as they assert participation in the same system. A strongly classified curriculum is not in any sense predicated upon a strongly classified physical space, although the former may well recruit the latter in sustaining its classification, just as an existing political system recruited the Berlin wall in sustaining its classification.
Rather, the strongly classified curriculum is achieved by strategies that—at any given level of analysis—specialise the various contents. Specialising always takes place within; the between is always established in terms of intertextuality. Minimally, this may be established in terms of negativity: in mathematics we use symbols that are not used in geography, and so forth. Walls are, of course, no barriers to intertexuality.

Bernstein is correct only to a very limited extent in claiming that what is classified may be realised in different ways, specifically, in different interactional modes. The cmc classroom is plausibly one in which very strong classification is realised. Suppose, for example, that the teacher is completely silent, or ‘lurks’. S/he can, nevertheless, review every contribution made by the student which might then be graded according to highly specialised principles and pass lists subsequently published. The problem for the student, of course, would be gaining access to the principles; arguably, this is not a pedagogic situation, because there is no transmission. The teacher may transform the situation into a pedagogic one by employing either weakly or strongly framed strategies; open questioning or lecturing, say. However, open questioning can remain open only insofar as the principles to be transmitted are weakly specialised or, rather, only in respect of those regions or aspects of the discourse that are weakly specialised. The panopticon (Foucault, 1977) might be construed as the archetype of weak framing.[7] However, as with the teacher-lurker in the cmc environment, this can work only where the prisoner already possesses the principles of evaluation of their behaviour and that would not be a pedagogic situation because it would entail no transmission.

Essentially the situation is as follows. Where that which is classified is the privileged content, or that which is to be transmitted, in a pedagogic situation, then the strength of framing of interactions must coincide with the strength of classification. Only where that which is classified is decoupled from this privileged content can classification and framing vary independently. An example of the latter would be, ‘you can do anything you like so long as you do it in this room’. Strong classification/weak framing, yes, but only because they do not refer to each other.

The problem can be resolved once we recognise that it can be traced to Bernstein’s original decoupling of space and time. Such decoupling is, of course, characteristic of various strands of structuralism and has been challenged in each of them; Derrida (1978) in respect of Saussure; Baudrillard (1993) in respect of Marx; Lacan (1977) in respect of Freud; Bourdieu (1977) in respect of Lévi-Strauss; and so forth. Bernstein and Piaget—the great educational structuralists—have remained substantially immune, to date (although see Dowling, 1998, in respect of Piaget). Essentially, a space-time decoupling can be sustained only to the extent that we ignore a shuffling between levels of analysis or (perhaps and) we keep our distance from the empirical.

A consequence of the resolution of the problem is that of the four concepts, power, control, classification and framing, three are redundant. I propose to retain classification and, in the construction of my own language, dispense with the other three. Thus, I propose that activities—say school mathematics—be construed as strategic spaces whereby subjects are positioned and practices distributed. In particular, specialising strategies constitute practices which are strongly classified with respect to those of other activities. Pedagogic action must entail the transmission of these specialised practices. In order to achieve this, the transmitter must constitute a discourse that is accessible to the acquirer. This in turn is achieved when the transmitter—as a subject of the activity in question, mathematics—casts a gaze beyond mathematics and recontextualises non-mathematical practice so that it conforms to the principles of specialised mathematical practice. Recontextualising is achieved by localising strategies, thus, the purchase of a loaf of bread in a supermarket becomes a local instance of specialised arithmetic. [8]

The effect of these localising strategies is to constitute a region of school mathematical practice that is weakly classified with respect to the non-mathematical. This is the public domain which contrasts with the esoteric domain that comprises practices that are strongly classified with respect to the non-mathematical. [9]

My concept of ‘recontextualisation’ is also a heresy in respect of Bernstein’s work. Bernstein’s use of the term refers to the creation of imaginary discourses from real discourses according to the ‘recontextualising principle’ that is ‘pedagogic discourse’ via the embedding of an ‘instructional discourse’ in a ‘regulative discourse’. Instructional discourse refers to specialised skills and regulative discourse to a moral order. By way of examples, Bernstein offers the recontextualising of carpentry—a ‘real discourse’—as ‘woodwork’—an ‘imaginary discourse’ and the recontextualising of university physics as school physics. My concern is that Bernstein’s theorising is constituting an unnecessary priority and not a little confusion. In order to demonstrate this, I shall need to work towards Bernstein’s higher level concept, the pedagogic device. I will begin with an extract from Class, Codes and Control volume 4.

It is of course obvious that all pedagogic discourse creates a moral regulation of the social relations of transmission/acquisition, that is, rules of order, relation, and identity; and that such a moral order is prior to, and a condition for, the transmission of competences. This moral order is in turn subject to a recontextualizing principle, and thus this order is a signifier for something other than itself.

(Bernstein, 1991; p. 184)

The mode of expression has been modified in the revised version of this chapter (see Bernstein, 1996, c. 3)—it is no longer quite so universally obvious, apparently. The theoretical formulation is retained, however, and the regulative (moral) discourse remains ‘the dominant discourse’ (Bernstein, 1996; p. 46) vis a vis the discourse concerning the transmission of competences (the instructional discourse). However, in the 1996 version, the distinction between regulative and instructional discourse is analytic or, perhaps, ideological:

In my opinion, there is only one discourse, not two, because the secret voice of [the pedagogic] device is to disguise the fact that there is only one. Most researchers are continually studying the two, or thinking as if there are two: as if education is about values on the one hand, and about competence on the other. In my view there are not two discourses, there is only one.

(Bernstein, 1996; p. 46)

There is only one, yet one of them is dominant. A little further on:

… pedagogic discourse is a recontextualizing principle. Pedagogic discourse is constructed by a recontextualizing principle which selectively appropriates, relocates, refocuses and relates other discourses to constitute its own order. In this sense, pedagogic discourse can never be identified with any of the discourses it has recontextualized.
We can now say that pedagogic discourse is generated by a recontextualizing discourse […]. The recontextualizing principle creates recontextualizing fields, it creates agents with recontextualizing functions. These recontextualising functions then become the means whereby a specific pedagogic discourse is created.

(ibid.; pp. 47-8)

The apparent confusion here is, so far as I can determine, the result of a failure by Bernstein to use key terms consistently and to constitute neologisms when and only when they are needed.

Bernstein is clearer when providing an example.

… the authors of textbooks in physics are rarely physicists who are practising in the field of the production of physics; they are working in the field of recontextualization.
As physics is appropriated by the recontextualizing agents, the results cannot formally be derived from the logic of that discourse. Irrespective of the intrinsic logic which constitutes the specialized discourse and activities called physics, the recontextualizing agents will select from the totality of practices which is called physics in the field of production of physics. There is selection in how physics is to be related to other subjects, and in its sequencing and pacing (pacing is the rate of expected acquisition). But these sections cannot be derived from the logic of the discourse of physics or its various activities in the field of the production of discourse.

(ibid.; pp. 48-9)

Bernstein may well be correct in his claim that school physics textbook authors are not generally productive phycisists. However, this rather misses the point. The authors of university physics textbooks generally are productive physicists, yet there are many important differences between university textbooks and, say, research papers (see Myers, 1992, also Dowling, 1998). The downplaying of the relevance of the logic of the discourse of physics in its recontextualised form is also open to some challenge. In an associated field, for example, the development of the ‘modern’ school mathematics in the nineteen fifties and sixties was heavily influenced by the Bourbakiist principle of mother structures (see Dowling, 1989, Moon, 1986). It is also questionable whether the ‘field of production’ is the only or even the dominant object of the gazes of recontextualising agents. The nature of integral calculus in advanced level mathematics certainly attests to this.

Essentially, Bernstein is making empirical claims and providing quasi-empirical illustrations in order to bolster his theoretical apparatus. The productivity of the more esoteric regions of this apparatus is difficult to imagine. In an earlier formulation, the instrumental and expressive orders constituted schemes through which the school might be and indeed was analysed as I have indicated in the comments relating to King.[10] Now ‘there is only one discourse’.

Jean Lave and Etienne Wenger agree that school physics is very different from university physics. They suggest that the decoupling may be even greater than Bernstein seems to imply:

… in most schools there is a group of students engaged over a substantial period of time in learning physics. What community of practice is in the process of reproduction? Possibly the students participate only in the reproduction of the high school itself.

(Lave & Wenger, 1991, p. 99)

In my conception the school as a site is to be conceived as a moment of a sociocultural system (Baudrillard, not Parsons). In terms of interaction, all such sites are characterised by a specific form of articulation of two modes of social activity. Pedagogic activity entails a transmitter, and acquirer, and a privileged content in respect of which the principles of evaluation of texts or performances resides with the transmitter. In an exchange activity the principles of evaluation are located with the acquirer. Strongly classified content is transmitted under pedagogic relations; weakly classified content is transmitted under exchange relations. The distribution of these relations constitutes differentiation within the school.

I have now made tacit reference to all three dimensions of Bernstein’s pedagogic device: distribution, recontextualisation, and evaluation. Bernstein’s pedagogic device—presumably, the ‘something other than itself’ for which pedagogic discourse is a signifier—is a heretical recontextualising of Chomsky’s language acquisition device with Bernstein explicitly adopting a Hallidayan rather than a Chomskian methodology. Bernstein argues:

Both the language device and the pedagogic device become sites for appropriation, conflict and control. At the same time, there is a crucial difference between the two devices. In the case of the pedagogic device, but not in the case of the language device, it is possible to have an outcome, a form of communication which can subvert the fundamental rules of the device.

(Bernstein, 1996, p. 42)

But, of course, these devices cannot become sites for any such thing. They are not, in fact, sites at all because they are not, ultimately, empirically operationalisable. The pedagogic device is a very high level theoretical object and we must descend through multiple layers of theory before we ever get to something that we might validly refer to an empirical text. The pedagogic device is a part of Bernstein’s ‘internal language’:

Briefly, a language of description is a translation device whereby one language is transformed into another. We can distinguish between internal and external languages of description. The internal language of description refers to the syntax whereby a conceptual language is created. The external language of description refers to the syntax whereby the internal language can describe something other than itself.

(Bernstein, 1996; pp. 135-6)

Bernstein’s own work has a strong tendency to reside in the internal. It is left to his disciples to generate external languages. By and large, these tend to be very thin, commonly making reference to boundaries and insulation and so forth which, as I have argued above, entail serious theoretical problems.

My formulation of my own general methodology—which I refer to as ‘constructive description’ is illustrated in Figure 2. I shall not give any extended discussion of this diagram here (but see Dowling 1998, 1999). It is important to note, however, that the principal theoretical and empirical achievement—such as it is—of my work takes place in the area signified by the box corresponding to Bernstein’s ‘external language’. That is, at the point of interface between the theoretical and the empirical. Bernstein seems to want to produce a theoretical system that is a model of what might metaphorically be described as the consciousness of society. My own project is far less ambitious. I am simply trying to manufacture a machine that will help me to organise what I see. In order to move between levels of analysis—say between the analysis of a conversation and the analysis of school practices (move 1) and the analysis of state policy (move 2), and so forth, I simply reapply the same conceptual framework, generating indicators that are appropriate to the new level. The method has, in this respect, a fractal quality. Recontextualisation in my language, then, is far more generalisable than it would appear to be in Bernstein’s. I define it as the subordination of the practices of one activity to the principles of another. It is precisely the empirical analysis of the productivity of recontextualisation that enables the constructive description of the recontextualising activity.

The categories, classification and framing also exhibit this fractal quality. Their disadvantage lies, as I have argued, in the fact that they do not themselves occupy the same level of analysis. A good deal of Bernstein’s theory (and here he is certainly not alone) is fixed in terms of its referent level. It may be that this is associated with his preference, following Halliday, for network analysis. This is an approach that fixes levels of analysis in relation to each other as one moves between levels of the network. I would describe a network as an analysis that has been terminated at a stage prior to the full development of theory. Bernstein has similar reservations about ideal types:

Classically the ideal type is constructed by assembling in a model a number of features abstracted from a phenomenon in such a way as to provide a means of identifying the presence or absence of the phenomenon, and a means of identifying the 'workings' of the phenomenon from an analysis of the assembly of its features. Ideal types constructed in this way cannot generate other than themselves. They are not constructed by a principle which generates sets of relations of which any one form may be only one of the forms the principle may regulate.

(ibid.; p. 126-7)

My feeling is that it is inappropriate to crystallise a method in this way. I explicitly describe my own approach as employing ideal types. However, my approach is to make a group of ideal types conceptually coherent to the point that they participate in the same theoretical system. Pedagogic and exchange relations constitute a case in point: they are defined in relation to each other in terms of the variable, ‘location of the principles of evaluation’. The application of an empirically driven network analysis does not encourage theorising to this level; the development of a theoretically driven network does not encourage empirical operationalisation.

Bernstein’s networks are commonly theoretically driven, hence his resistance to the accusation of having produced ideal types. The empirical is not absent in his theory building, but appears, shall we say, hazily. His description of vertical and horizontal discourses is a case in point. The network (Figure 3) is beautifully clear in terms of its oppositions: vertical/horizontal, between/within, strong/weak, explicit/tacit. The difficulty arises when we try to assign empirical instances to locations in the network. There are two modes of vertical discourse:

A vertical discourse takes the form of a coherent, explicit, systematically principled structure, hierarchically organized, or it takes the form of a series of specialized languages with specialized modes of interrogation and specialized criteria for the production of texts.

(ibid.; p. 171)

(ibid.; p. 142)

Notes