In the previous chapter we have discussed the sociolinguistic codes, both as Bernstein has explained them and also in the ways I have suggested they might be reinterpreted or extended, but without either trying to tie the different issues together or suggesting what kinds of research would be appropriate to investigate the empirical utility of my suggestions. This chapter seeks to address these issues, and in so doing to say something more about the kinds of operationalizing of Bernstein's ideas that have been current so far. I shall discuss these matters with the Jamaican context in mind.
Perhaps it would be best to start with the one piece of work that has used Jamaican material to examine Bernstein's ideas, Craig's thesis and its reanalysis that was mentioned in the second chapter. In introducing his reanalysis, Craig tells us that his aim is to "look at social-class language differences in such a way as to clarify the differential in habitual purposes, referential content and communication styles that can be perceived" (1982, p. 4). On the face of it, this focus might be close to the questions of possible differences in dispositions to different sorts of speech act that I have suggested in the previous chapter, or again to the recent Bernstein work on types of categorization.
But unfortunately, when one examines exactly how Craig proposes to analyze his data, these concerns give way to the much more narrow, and as it were grammatical considerations typical of so much of the SRU work itself. In these terms, Craig has displayed differences in "referential content and communication styles" among his groups of children - urban children "showed a greater interest in concrete nominal references related to fiction and the mass media than their rural counterparts" (ibid., p. 18); when one allows for the very different linguistic resources for marking tense, "there are no significant social-class differences between the children in references to time by means of tense aspect marking" (p. 37); "the two sets of urban girls tended, in different ways, to reference possession with about equal frequency and somewhat more than their male counterparts do" (p. 34); and so on. There are very many such findings, just as there are in the various SRU monographs. But what is important for my purposes is the fact that these types of difference in language use are not directly connected to the sorts of question I suggested ought to be looked at in the previous chapter. Counts of particular words, or types of structure, do not necessarily reveal anything of interest about those kinds of language use, about the varying dispositions to engage in cognitively different sorts of speech act. And as I suggested there, one would think that the cognitive acts were more important in school success or failure than at least some of the grammatical patternings.
Craig's way of categorizing distinct sentence types in fact makes it virtually impossible for him to look at the different cognitive speech acts - his list of distinct types includes questions, onomatopoeic "sentences" to convey aeroplane noises, negative sentences, and "NP-VP affirmative sentences which need not be specially noted" (p. 10). One might well wonder at a category of negative sentences when commands or questions could easily be negative, but the more significant point is that a simple NP-VP sentence such as John is sick can, as Craig well knows, be performing various functions in speech. It might simply be a description, but it might be functioning as part of an explanation or excuse. Craig's analytical framework thus makes it impossible for him to search for differences in a child's disposition to offer reasons with any real hope of success. It remains in the predicament Adlam described of supposing that each significantly different speech act will come explicitly marked as such, though ironically enough it has always been Craig's main point that creole speakers can and do perform various cognitively complex operations (causal inference, justification, etc.) using modes of language that do not explicitly mark these acts. His method of analysis here, however, belies this insight.
The problem seems to be that in his eagerness to insist that creole, or working class speakers do do these things, even if they do them differently, he does not examine the data to see whether there might be styles, not of linguistic expression of these acts, but of the distribution of the acts themselves.
Before leaving Craig's work there are two other points that are worth mentioning. One was noted in chapter 2 - Craig's very crude SES classification of his subjects. I would not criticize him for this - the difficulties of doing individual research are such that one must often settle for crudity - but it is worth noting how far it is from what a Bernsteinian theory would require for its testing. Even if one decided not to test the higher reaches of Bernstein's account, in which the division of labour and the powers that maintain it connect with interrelationships in the home (this strand would surely in practice call for a sufficiently laborious investigation on its own without adding specifically sociolinguistic investigations of the speech in such homes) one does need to examine kinds of home interactions and foci of attention.1
This might well be rather more difficult and involved in a Jamaican context where children are brought up in a very wide range of home environments, with or without one or both physical parents, and typically with larger numbers of siblings and other children than in Bernstein's England. One imagines that the range of possible family role structures which might result in different coding orientations in Bernsteinian terms could be quite considerable and would cut across class boundaries drawn in normal economic or status ways. On the other hand, the broad contrast between home bequeathed elaborated codes and the rest might be somewhat easier to discern, given that the school system is so strongly differentiated.
The second point relates to the kind of data Craig used. He recorded children speaking spontaneously among themselves. There are of course good reasons for so doing if you want to get some idea of what children are capable of, since it is well known that in general their linguistic activity is inhibited by adults, particularly comparatively unknown or authoritative ones (this is a widely documented fact, one of the points Labov has used against much of Bernstein's own work). But here it reveals again Craig's major concern to demonstrate that creole speakers do not necessarily suffer from any conceptual or cognitive deficiencies. Bernstein need never have thought otherwise, but as he has often insisted, what the school system uses to differentiate people is not spontaneous speech among peers but precisely the kind of linguistic and cognitive activity that children offer to authoritative adults. From that perspective, it matters little whether, as in Holland's research, some children are perfectly capable of understanding and using abstract categorization schemata if they do not habitually use such ways of categorizing the objects or situations presented to them in school environments, in routine school work or in examinations.
This point undermines one of Jackson's supposedly apodictic criticisms:
a man cannot be constrained by his own mere tendencies.... if in a stretch of speech I use the word 'car' 150 times, and the word 'automobile' only once, that proves that I know the word 'automobile', not that I do not know it; and that I am not constrained to use the word 'car' not that I am so constrained (1974, pp. 70-1).No doubt Bernstein's talk of constraint can be faulted for connotations of helpless determination, though Bernstein has been keenly aware of the need for a way of conceptualizing human action that steers some plausible via media between a rigid fatalism and self-indulgent fantasies of extreme freedom. But Jackson's more restrained suggestion that different frequencies in the usage of some word can show different motivations which should be tied to probabilistic features of the environment is unnecessarily restrictive. There could be an internalized preference structure (indeed this might well seem to be simply the "inner" correlative of Jackson's preferred environmental features) which played a significant part in producing observed frequency distributions, which themselves are one major basis for the school's treatment of the speaker.
In suggesting that one needs to look at children interacting with adults, especially adults within the school system, I am raising the question of the role of Bernstein's codes, or my glosses thereon, in educational success or failure. In this area I am sanguine enough to suspect that work in Jamaica might bear out his contentions somewhat more clearly perhaps than similar work in the United Kingdom.
As we have noted, one of the main functions of the total educational system for Bernstein is to differentially distribute elaborated codes. We have already noted the kind of criticism that stops at the gruff ultra-restricted code of typical teachers trying to stay in control of their classes, instead of probing beyond to the kind of language or cognitive activity required by written work in schools. In discussing that criticism I also noted that since Bernstein himself characterizes what happens in the lower streams of many secondary schools as an instance of our having given up trying to educate children he would not expect the typical demands of such schooling, spoken or written, to exemplify elaborated coding. Perhaps we can say more generally that if one job of a school system is to differentially distribute something, one way it can do this is never to put it on offer for some participants. In post-Plowden England, with lots of bright and airy primary schools with falling enrolments and a fairly committed and extensively trained teaching force, it would perhaps be difficult to show that some children were never put within reach of educational capital, though on Bernstein's account of the ideology of the "invisible pedagogy" it is no longer thrust upon them. Instead one would have to point as Bernstein did himself to what typically happens to a sizeable proportion of secondary students in the upper reaches of that segment of the system. But Jamaica is not post-Plowden anywhere. Its primary schools are typically overcrowded, rigidly organized, dominated by the Common Entrance Examination which fails approximately 75% of its entrants (which are by no means all those children of the relevant age group), and staffed by ill-trained and disgruntled teachers (for much more extensive documentation of this sector, see UNESCO, 1983, pp. 67-81). There are, of course, also the private "prep" schools (and a number of public primary schools that have managed to maintain higher educational standards), which with the extensive market in extra tuition help to ensure that parents who can pay will stand a much better chance of seeing their children pass the Common Entrance. I do not know of published data for Jamaica, but informal observation suggests that one would also find here the discrepancies in pedagogy and classroom interactions that have been found elsewhere between the schools for the rich and the schools for the poor (for Chile, see Filp, Cardemil, Donoso, Torres, Diéguez, & Schiefelbein, 1981; for New York State, Anyon, 1981): rote learning for the poor, "discovery", activity and understanding-oriented work for the privileged.
Such tendencies exist everywhere, as Anyon's data suggest for the USA, but they are likely to be more extreme in the context of Jamaica than in the more extensively researched "first" world. Another pertinent finding in the USA which is likely to be confirmed here is the tendency for teachers to help to create a downward spiral of inability for children who appear not to be achieving as well as is expected. Thus Hansen and Pearson, 1983, report several studies showing that reading teachers expect less inferential work from students they see as poor readers. Adding that to Wells' findings mentioned in the previous chapter, we can suggest a process whereby a home interest in literacy translates fairly smoothly into a disposition in the child to perform more frequently and more explicitly the sorts of reasoning tasks that are the stock in trade of tests of general ability and which can surely help in passing the more specialized tests that schools use.
It must be admitted that such evidence as there is suggests that very little of any pupil's spoken work will involve complex cognitive or logical operations, so Bernstein's occasional talk of elaborated codes giving their users access to "meta-levels" of control or cognitive activity might suggest erroneously that this is often an explicit part of the lesson. As Edwards says,
it is also common for the relevant display of knowledge to be very inexplicit, the pupil needing only to touch on the required facts to launch the teacher into filling-out an answer which was itself highly allusive (1980, p. 239).And Young sums up a survey of studies of classroom interaction by saying that "the crucial fact about teacher talk ... is that teachers do most of the higher level cognitive work and they do it implicitly" (1983, p. 11). But once again, one point is that success in schools is not just a matter of listening to teachers, notoriously you have to regurgitate it for exams; another is that the quality of the implicit work one is exposed to can differ quite dramatically. In a school system as differentiated on economic lines as Jamaica's, such differences will often mean that many schools will virtually never offer their pupils exposure to elaborated codes, even in the typical implicit, unexplained way.
This is not a particularly easy charge to substantiate since I have done nothing to fill in what kinds of speech act or cognitive act one should look for within the broad areas of explanation, justification and "critical thinking". And there are difficulties in judging the quality of many such cognitive activities. While some inferences from evidence to a conclusion are patently absurd, they are not usually the ones that people make; within the range of inferences people do make it is not always clear which are good, bad or indifferent. An additional problem is that one does not want to look for subject-specific types of knowledge or cognitive functioning so much as for general intellectual skills; but then it is not easy to say what level of these skills should count as suggestive of an elaborated code orientation. To take an example from an area I have begun to explore in the Jamaican education system: everyday deductive reasoning is not very well grasped by ordinary people, especially those aspects relating to deductively invalid arguments. As far as I know, there are no experiments focussing on performance differences between people who have independently been allocated to Bernstein-type codes, but it is more than likely that even clearly elaborated code persons would do rather badly at invalid arguments - professors of logic have been known to get one of the popular test items wrong, so the rest of us can have little hope of consistent success. So there is a currently unresolved question of what quality of performance should be used to differentiate possible code differences. What one can say, however, is that there are few differences between typical Jamaican school teachers and typical secondary school children on these reasoning tests.2
Given the usually comparatively poor performance of these secondary students in public examinations, one might suggest that this sort of reasoning ability has little to do with examination success (which is fairly plausible) or that the pupils are not likely to pick up much from their teachers that they don't already have, or both - the teachers themselves tend not to have been particularly successful in such examinations either. Referring back to Young's summary of the evidence in general, one can say that implicitness is going to reproduce only what is already there; if the teachers are not noticeably different from their pupils, then any differences are going to be due largely to non-school factors. In this respect, at least, schools in Jamaica are then not so much differentially distributing a property they have as serving to filter out those who do not already have it.
The difficulty I have already alluded to of establishing a standard of competence in this area can be seen when it is noted that the performance in deductive logic is not very different, at least among teachers and higher stream students from that of similarly aged students in the USA (Ennis and Paulus, 1965; a few comparisons are given in Brandon, 1984). What might be somewhat more disturbing are scores in the more diffuse area of critical thinking - something more central to my rendition of Bernstein's codes than pure deductive logic. Being more diffuse, I cannot establish how far the differences that have been found are due to differences in the approach adopted towards the scoring instructions, but if they are not a spurious creation of my lack of familiarity with the instrument, they are seriously worrying educationally and point to a standard of ordinary critical thinking significantly below that of North American high school students.3 If they are on the right lines they suggest again and much more strongly that the ambiance of typical Jamaican schools is not going to encourage the kinds of critical thinking skill required for effective engagement in the activities typical of my gloss on elaborated codes. I have said many times that using my glosses on the codes would not be an easy task. In the previous paragraphs I have been focussing on the different sorts of speech or cognitive acts that I suggest are an important part of the contrast between codes and which have often come close to the surface in Bernstein's own talk of access to principles and explanations. Looking for such activities involves not only observing the bare activity but also grading it somehow, and this is not so clear-cut in many fields as it might appear in areas like mathematics.
The difficulty in operationalizing the other important contrast, implicit and explicit, arises from a similar awkwardness in defining a cut-off point. The contrast is one of more or less, and in practice there is a fairly small range over which language use varies. Bernstein's collaborators have acknowledged the difficulties with their own counts of exophoric versus anaphoric reference in practice, and similar problems would be sure to arise if one turned towards more diffuse but arguably more important modes of implicitness such as were discussed above under the umbrella of ellipsis. Perhaps an even greater difficulty arises here in that it is likely that a person's awareness of elided structure, in the case of ellipsis as described above, is going to be revealed, not so much in what she says in the immediate context (not, that is, by her explicitly filling in the missing elements) but by a larger scale strategy of argument or exposition. If it is difficult to tell what speech act a declarative sentence is being used to perform, it must be more difficult to delineate the overall assumptions of an entire discourse. But those are the directions my suggestions point towards.
URL: http://www.uwichill.edu.bb/bnccde/epb/msc4.html
HTML prepared June 2nd, 2000.
© E.P. Brandon. Last revision: 9th June, 2000.