Search

In the original exchange with Michael Lynch two things were at stake: first, we disagreed on the best way to analyse science as a social activity, and second, we disagreed on how to exploit the legacy ofWittgenstein's work in the course of such an analysis (see Pickering, 1992). Lynch opts for an ethnomethodological approach, which is descriptive but not explanatory in any causal way; I opt for a form of the sociology of knowledge in which the aspiration to provide causal explanations plays a central role. Lynch sees Wittgenstein's work as constituting a challenge to the sociologist, while I see it as a resource that can be used to deepen and further the sociological enterprise. Although we have a running disagreement about the interpretation of particular arguments in the Philosophical Investigations we agree, in general terms, that Wittgenstein's work is capable of being read in either way. The question is: which is the most fruitful? Predictably, we define fruitfulness differently. The result is that Lynch sees Wittgenstein as an embryonic ethnomethodologist; I see him as a proto-sociologist. I readily concede that the big battalions are on Lynch's side. Most philosophers selectively emphasize the points in Wittgenstein that Lynch deems decisive and share his lack of concern for the points I find most interesting. Wittgenstein was, after all, the patron saint of the 'ordinary language philosophy' or 'linguistic analysis' that dominated Oxford philosophy in the 1950s and 60s. These philosophers said they were doing neither metaphysics nor empirical science, but merely studying the 'logical geography' of concepts. Neither constructive theory building nor causal explanation was seen as necessary just 'analysis' and 'clarification'. Though fashions change quickly in philosophy, something of this spirit lives on in current Oxford readings of Wittgenstein (for example, those of Baker & Hacker, 1984). It is intriguing to find these readings enthusiastically absorbed into ethnomethodology. It has to be admitted that the two fit together very well. But for those of us who cut our philosophical teeth on Oxford-style linguistic analysis, and decisively rejected it, it is sad to see its narrow and self-imposed limitations reproduced and relabelled as a form of radicalism. This is why, in the original exchange, I spoke of a Right-wing

A case study in the history of transonic aerodynamics, circa 1950, is used as a basis for reflecting on the character of the distinction between pure and applied mathematics, including the mathematics used by engineers. The case study is set against an historical background of disciplinary confrontation led by such eminent representatives of mathematics and aerodynamics as Garrett Birkhoff and Theodore von Kármán. The successful attempt to construct an adequate account of the aerodynamics of the transonic realm highlighted some sharp differences in the procedures and preferences of mathematical practitioners operating in different fields. The existence and general character of these differences is already widely acknowledged but the task of exploring them from a sociological standpoint still requires much work. The nature of the disciplinary distinctions between different areas of mathematics is examined using Barnes’ theory of (idealized) natural-and social-kind terms (so-called N-and S-predicates). Although the ultimate status of the disciplinary boundaries turns out, uncontroversially, to be ‘conventional’, the attempt to make out and exhibit the conventionality in detail proves to be a non-trivial exercise. It transpires that a thorough study of the issues turns on deep questions about the nature of mathematical rigor and a process that might be called ‘the exploitation of contingency’. These points are illustrated in detail by reference to the technical work of the area (in which one of the authors was an active participant).

The dialogue between natural scientists and sociologists has never been extensive. As everyone in our field knows, it is currently in danger of breaking down almost entirely. In the rancorous atmosphere created by recent attacks on the sociology of knowledge it is a pleasure to see a different attitude find expression, one which may be critical but which is also generous and civilized. Refreshingly, David Mermin begins his Comment on our book, Scientific Knowledge,1 with a clear acknowledgement that we do not wish to deny the role played in science by human interaction with the material world. Every sane person knows this is central to scientific activity, just as it is to, say, mining, manufacturing or making music. It is true that there has been a small number of highly publicized lapses by a few sociologists who fail, or refuse, to distinguish between reality and reality-as-described, or reality-astheoretically-understood. In this mistake the errant sociologists are not alone. Many philosophers likewise run together 'facts-as-states-of-affairsin-the-world' and 'facts-as-descriptions'. Oddly enough, philosophers who speak in this way often present themselves as 'realists' - and are allowed to get away with it. The authors of Scientific Knowledge are not among the offenders, and have always maintained a materialist (or, in a genuine and robust sense, realist) position. In the present climate, however, critics cannot always be relied on to draw such basic distinctions between radically different positions. Fortunately Mermin has our stance clearly in view in this respect and does not blur our position with some of the quite different approaches on offer. For us it is knowledge that is 'socially constructed', not reality. Mermin also realizes that there is a deep difference between the scientist's own involvement with his or her craft and the approach which properly informs the sociology of science. Doing physics, chemistry or biology is different from analyzing the cognitive processes involved in these practices. Mermin conveys this fundamental difference of perspective by an interesting analogy. He likens it to a change in coordinate systems.

The sociology of scientific knowledge is an empirical discipline, but occasionally it can be fruitful to reflect on its methodological basis. Critics have sometimes claimed that it is committed to a form of `idealism' — that is, to discounting or playing down the input of the material world. This arises because sociologists often sum up their conclusions by saying that `knowledge is a social institution', or that `concepts are institutions'. If we think of social institutions according to the self-referential or performative model outlined by Barry Barnes, this may at first seem to reinforce and justify the charge of idealism. The main argument of this Comment is to show that while an `idealist' account of institutions is correct, the conclusion alleged by the critics does not follow. A secondary purpose is to compare Barnes' account of institutions with recent work by John Searle, and to show the significance of their different underlying assumptions about the nature of meaning.

Robert Nola claims that 'ordinary human inference' refutes the strong programme in the sociology of knowledge.' Since there are lots of such inferences, the result is a plethora of counter-examples. These reveal the sociology of knowledge, in the form defended by the present writer, to be an untenable construction based on 'fantastic' (125) behaviourist assumptions, denying any role to cognitive processes and postulating 'unmediated' stimulus-response connections between social context and individual belief (108, 123-24). A symptom of these failings, says Nola, is that the strong programme (which he abbreviates to SP) is 'entirely impotent when it comes to the reasons we can have for adopting the views we do' (122). Giving and responding to reasons is at the heart of ordinary human inferring, and that is partly why it constitutes a standing refutation. For the sake of brevity, I shall not attend to all of the shortcomings that I find in Nola's paper, but I hope silence will not be taken to imply consent. To give just one example, Nola builds on the work of

Oral history materials from the South Wales Miners' Library are used to examine the communal understandings of, and collective responses to, the scourge of Miners' Lung (pneumoconiosis) in the 1920s and 1930s. Lay epidemiology in mining communities attributed an aetiological rôle to coaldust at a time when many experts believed miners' pulmonary disease to be bronchitic, or to be silica-induced. In their efforts to secure compensation claims for their members, union officials instrumentally used scientific expertise in a variety of forms: they contributed to epidemiological evidence; they lobbied for more government-funded research; they `bought' experts; they duped expert witnesses; and they made sophisticated instrumental appeals to the supposed independence of favourable expert judgements. Eventually, miners' situated `local knowledge' became scientific orthodoxy, a success story which may be associated with the class-conscious miners' `bump of irreverence' about expert knowledge, and with the divided character of the expert core-set, sections of which were receptive to miners' `local knowledge' claims.

In October 1967, the Science Studies Unit at Edinburgh University began teaching courses on social aspects of science and technology to undergraduates in Science Faculty. From the start, these courses drew material from the disciplines of philosophy, history and sociology, selected in such a way as to introduce science students most effectively 'to ways of analysing the role of the scientist (both within the scientific community and in society at large), and to some of the wider implications of scientific thought and practice'.' The 'stream' which draws on philosophical material has always been available to students as a separate 'half-course', and has been taught essentially in the form described below since 1970. It is intended that the course will occupy between one-sixth and one-quarter of the student's time for two terms (i.e. twenty weeks). No prior knowledge is assumed. Every student prepares four essays during the course: these are read and discussed in fortnightly tutorials. The outline below contains details of the literature which is expounded and discussed in the lectures. Most of the course deals with a closely-knit set of issues connected with T. S. Kuhn's The Structure of Scientific Revolutions (Chicago and London: The University of Chicago Press, Second Edition, 1970). This is the 'set book' for the course, and students are advised to purchase it. Most of the readings are articles and papers, rather than books. Where a book title is mentioned below without reference to specific chapters, it is usually within

Eric Livingston's book, The Ethnomethodological Foundations of Mathematics, addresses two of the most intriguing and difficult questions in the philosophy of mathematics. First, what is the nature of mathematical objects, for example, the circles and lines mentioned in geometrical theorems? Second, what is the nature of the compulsion that attaches to mathematical reasoning? The idealized and perfect circles whose properties are revealed in Euclid are a far cry from the poor, crooked things that a mathematics teacher draws on a blackboard. Again, the proofs of theorems seems to embody a peculiarly potent form of persuasion. How can we possess such superlative insights into such superlative things? Livingston says that the answer to these long-standing questions will be found by attending to certain of the local contingencies that surround particular episodes of mathematical reasoning and theoremproving. His aim is to get us to appreciate the moment-by-moment 'work' of mathematical reasoning at the 'mathematical work-site' (6). The claim running throughout his book is that by attending to these details we will come to understand how the upshot of this 'work' can be seen as possessing the transcendent qualities just described. Livingston says his concern is with the 'living foundations of mathematics' (x), in contrast to the 'classical' study of foundations. Classical studies, such as those of Russell, aim to provide a definition of mathematical concepts in terms of even more basic logical concepts. Their aim is to justify mathematics by deriving its operations, such as addition, from formalized logical operations of a kind that are even more primitive. However, all the fundamental processes of reasoning that underlie our

In 1840 J.S. Mill surveyed the intellectual landscape around him and reported that it was dominated by two great figures: Jeremy Bentham and Samuel Taylor Coleridge.' Between them, he said, they defined the terms of reflective thought, embodying the choices that had to be made and the dilemmas that had to be resolved. Cold calculation or warm sentiment? Individualism or community? Utilitarianism or Romanticism? If Mill was right then we should expect to find these two currents of thought flowing through all the cultural products of that time. This seems plausible if we are dealing with philosophy or economic theory or political rhetoric, but what about science? While there may seem to be a natural affinity between science and utilitarianism, the link with the romantic tradition is less easy to grasp. The temptation is to say that the idealist metaphysicians, and nature-philosophers who belonged to this tradition, were not contributing to science so much as perverting it. They might be using it, or extrapolating it, but they were not doing it. In short, we draw a line of demarcation between men like Coleridge and those like his friend Sir Humphry Davy who were, we are in no doubt, really scientists. Trevor Levere's fascinating book, Poetry Realized in Nature: Samuel Taylor Coleridge and Early Nineteenth-Century Science, is one of a number of recent works that are helping to challenge our