The Constructivist Interpretation I: Nature and the Laboratory

How do we defend the contention that scientific enquiry ought to be viewed as con­structive rather than descriptive? And what exactly do we mean by this particular qualification? The first question can be answered rather simply. Even the briefest par­ticipation in the world of scientific investigation suggests that the language of truth and hypothesis testing (and with it, the descriptivist model of enquiry) is ill-equipped to deal with laboratory work. Where in the laboratory, for example, do we find the "nature" or "reality" so critical to the descriptivist interpretation? Most of the reality with which scientists deal is highly preconstructed, if not wholly artificial.

What, after all, is a laboratory? A local accumulation of instruments and devices within a working space composed of chairs and tables. Drawers full of minor utensils,

4 The Manufacture of Knowledge

shelves loaded with chemicals and glassware. Refrigerators and freezers stuffed with carefully labelled samples and source-materials: buffer solutions and finely ground alfalfa leaves, single cell proteins, blood samples from the assay rats and lysozymes. All of the source-materials have been specially grown and selectively bred. Most of the substances and chemicals are purified and have been obtained from the industry which serves science or from other laboratories. But whether bought or prepared by the scien­tists themselves, these substances are no less the product of human effort than the measurement devices or the papers on the desks. It would seem, then, that nature is not to be found in the laboratory, unless it is defined from the beginning as being the pro­duct of scientific work.

Nor do we find in the laboratory the quest for truth which is customarily ascribed to science. To be sure, the language of scientists contains innumerable references to what is or is not true. But their usage in no way differs from our own everyday use of the term in a variety of pragmatic and rhetoric functions which do not have much to do with the epistemological concept of truth. If there is a principle which seems to govern laboratory action, it is the scientists' concern with making things "work", which points to a principle of success rather than one of truth. Needless to say, to make things work—to produce results—is not identical with attempting their falsification. Nor is it the concern of the laboratory to produce results irrespective of potential criticism. Scientists guard against later attacks by anticipating and countering critical questions before publications. The scientists' vocabulary of how things work, of why they do or do not work, of steps to take to make them work, does not reflect some form of naive verificationism, but is in fact a discourse appropriate to the instrumental manufacture of knowledge in the workshop called a "lab". Success in making things work is a much more mundane pursuit than that of truth, and one which is constantly turned into credits in scientific everyday life via publication. Thus, it is success in making things work which is reinforced as a concrete and feasible goal of scientific action, and not the distant ideal of truth which is never quite attained.

But "truth" and "nature" are not the only casualties of the laboratory; the observer would find it equally difficult to locate those "theories" which are so often associated with science. Theories adopt a peculiarly "atheoretical" character in the laboratory. They hide behind partial interpretations of "what happens" and "what is the case", and disguise themselves as temporary answers to "how-to-make-sense-of-it" ques­tions. What makes laboratory theories so atheoretical is the lack of any divorce from instrumental manipulation. Instead, they confront us as discursively crystallised ex­perimental operations, and are in turn woven into the process of performing ex­perimentation.

In place of the familiar alienation between theory and practice,²⁰ we find an action/­cognition mesh to which the received notion of a theory can no longer be adequately applied. According to the scientists themselves, theories in research are more akin to policies than creeds.²¹ Such policies blend interpretation with strategic and tactical calculations, and are sustained by methodological "how-to-do-it" projections. Like the concern with making things work, policies are necessarily tied to an interest struc­ture. Pure theory, then, can be called an illusion the sciences have retained from philosophy.²²

The Scientist as a Practical Reasoner 5

1.3 The Constructsvist Interpretation II: The "Decision Ladenness" of Fact-Fabrication

The inadequacy of those concepts associated with the descriptive interpretation of scientific enquiry is not surprising, given the framework in which they developed. It is no less surprising that a shift in the framework of analysis to the actual process of research brings new conceptions into being. We have said that this process should be seen as constructive rather than descriptive. Let us be more specific. The thesis under consideration is that the products of science are contextually specific constructions which bear the mark of the situational contingency and interest structure of the process by which they are generated, and which cannot be adequately understood without an analysis of their construction. This means that what happens in the process of con­struction is not irrelevant to the products we obtain. It also means that the products of science have to be seen as highly internally structured through the process of produc­tion, independent of the question of their external structuring through some match or mismatch with reality.

How can we conceive of this internal structuring of scientific products? Scientific results, including empirical data, have been characterised as first and foremost the result of a process of fabrication. Processes of fabrication involve chains of decisions and negotiations through which their outcomes are derived. Phrased differently, they require that selections be made. Selections, in turn, can only be made on the basis of previous selections: they are based on translations into further selections.

Consider a scientist sitting at an electronic table calculator and running a regression programme on texture measurement data. The machine automatically selects a func­tion along which it plots the data. But in order to choose among the eight functions at its disposal, it needs a criterion. Such criteria are nothing more than second order selec­tions: they represent a choice among other potential criteria into which a first order selection can be translated. In our case, the programme actually offered a choice bet­ween two criteria, maximum R² and minimum maximum absolute residuum. The scientist had opted for a combination of the two.

He obtains an exponential function for his data, which he says he doesn't like. He reruns the programme, asking for a linear function, which he finds to be "not much worse" (than the exponential one). The idea, he says, is to get one type of equation, and eventually one size Beta coefficient for all runs of the problem, because it would be totally confusing to have different functions in every single case.

From observing the scientist, we might also conclude that the goal must have been to get a linear function. In order to reach a decision, the original task of the programme was to select a function translated into the selection between one of two forms of the statistical fit of the curves. In a stepwise procedure, the scientist added translations into other criteria, such as uniformity over comparable data and linearity. He eventually chose the latter because it offered greater ease of interpretation and presentation.

Gallon has recently illustrated how the relationship between supply and demand in regard to information can be seen as a symbolic operation of translation (Serres, 1974) which transforms one particular definition of a problem into another particular state­ment of a problem. For example, the problem of reducing urban smog may be translated into the problem of reducing the amount of led in petrol, or of transforming the affected area into a pedestrian zone. This implies that the solution of problem A re­quires the solution of problem B, into which A has been translated.²³ In the present case, this kind of translation is seen as an inherent feature of decision-making, or—to

6 The Manufacture of Knowledge

borrow an expression from Luhmann—of selectivity in general.²⁴ It allows us to see scientific products as internally constructed, not only with respect to the composite laboratory selections which give rise to the product, but also with respect to the transla­tions incorporated within those selections.

In other words, the scientific product can be seen as structured in terms of several orders or levels of selectivity. This complexity of scientific constructions with regard to the selections they incorporate is interesting in its own right because it does seem to sug­gest that scientific products are unlikely to be reproduced in the same way under dif­ferent circumstances. If a scientific product is characterised by several levels of selec­tion (or constellations of selections), it seems highly improbable that the process could be repeated, unless most of the selections are either fixed or made in a similar fashion. Given that scientists working on a problem are related through communication, competition and cooperation, and often share similar educations, instruments and in­terest structures, the latter situation is not really unusual.²⁵ But this translation of selections not only points to scientific products as complex constructions incorporating layers of selectivity, but also (as we shall see in Chapter 4) provides the threads with which laboratory selections and the products they compose are woven into the relevant contexts of research.

In order to reach some form of closure, selections are translated into other selec­tions. To break up that closure, the selections can be challenged on their own grounds. Selections can be called into question precisely because they are selections: that is, precisely because they involve the possibility of alternative selections. If scientific objects are selectively carved from reality, they can be deconstructed by challenging the selections they incorporate. If scientific facts are fabricated in the sense that they are derived from decisions, they can be defabricated by imposing alternative decisions. In scientific enquiry, the selectivity of the selections incorporated into previous scientific work is itself a topic for further scientific investigation. At the same time, the selections of previous work constitute a resource which enables scientific enquiry to proceed: they supply the tools, methods, and interpretations upon which a scientist may draw in the process of her own research.

The "artificial" character of the scientist's most important tool, the laboratory, lies in the fact that it is nothing more than a local accumulation of materialisations from previous selections. The selections of previous investigations also affect subsequent selections by modalising the conditions of further decision-making. Thus, the products of science are not only decision-impregnated, they are also decision-impregnating, in the sense that they point to new problems and predispose their solutions.

Briefly, then, a scientist's work consists of realising selectivity within a space con­stituted by previous selections, and one which is essentially overdetermined. In more economic terms, we could say that scientific work requires the re-investment of previous work in a cycle in which the selections generated by scientific work and their material equivalents are themselves the content and the capital of the work. What is reproduced in this cycle is selectivity per se. This form of auto-capitalisation in regard to selectivity appears as a precondition for the accumulation of scientific results. It can be multiplied by increasing the number of scientists and through increased financial resources. The conversion of scientific products into research money as described in re­cent economic models as discussed in Chapter 4 refers to this aspect. We can also say it refers to scientific productivity rather than to scientific production.

The Scientist as a Practical Reasoner 7

Поиск по сайту: