Lawrence Solum's recent elegant posting on burden of proof helps to clarify a related problem in understanding the relationships among scientific disciplines. Our model for thinking about burden of proof is, of course, the criminal trial and the rules of procedure which have grown up around it. But scientific research also involves extensive burden-of-proof considerations, for claims to discovery are subject to a similar kind of public scrutiny, complete with versions of cross-examination and rebuttal. Indeed, the historian Barbara Shapiro argues that scientific notions of fact-establishment are modeled on the legal procedures of the early modern period.
Here, I want to look at the way that distributing the burden of proof in scientific research acts to organize the relationship among specialties.
When scientists claim to make discoveries, their work is subject to review and test. While a discovery is still provisional, those who want to build upon it must rehearse the evidence and arguments for it when they write papers for the journals. As the discovery becomes more widely accepted, the amount of justification required for relying on it diminishes. After a while, referees start taking it for granted, and-- eventually-- reject papers that do not include the discovery (at least tacitly) or that argue against it. At that point, when acceptance has become obligatory, the discovery is established. Newer work which does not recognize it will be rejected as inadequate. In the extreme case, a scientist's persistent refusal to accept a discovery can result, not in the rejection of the discovery, but in the rejection of the scientist as aberrant in judgment.
Clearly, this is a kind of burden of proof; scientists who wish to reject an established claim are obliged to explain why their rejection should be accepted.
Situations in which this burden of proof crosses specialty boundaries are especially interesting. When a discovery is first claimed, the audience that reviews and accepts the claim is typically made up of experts in the line of research from which the claim emanates. These experts are most familiar with the phenomenon and the difficulties of studying it, and are most likely to notice errors of technique, omissions of pertinent data, and so on. They are also most familiar with the scientists making the claims, and are hence best placed to assess their reliability and probity.
Scientists rarely are equipped to review and test discovery claims in other specialties. Instead, they are likely to rely on the judgment of specialists in the relevant area, and on the reputation of the claiming scientists, their host organizations, the journals they publish in, and so on. Typically, discoveries can be accepted in this second- or third-hand way without changing anything that happens in other specialties. Sometimes however, a development in one area has implications for work in another. Often, the discovery is helpful; for example, scientists in one specialty may develop techniques which can be used as the basis of measurement procedures in another specialty. So, for example, studies of interferometry in physics underlie the design of instruments in radio astronomy.
As new ideas from outside become established in a specialty, they become obligatory there. To use the new techniques, instruments, and models is to accept the externally established ideas which underlie them. Very often, nobody has a reason to contest these obligations, and no problems arise.
Sometimes, however, things are not so simple. Sometimes, an external discovery does not fit well in the system of relationships which make up a line of research. Perhaps some of its implications are at odds with conventional assumptions. Darwin's natural selection hypothesis, for example, seemed to conflict with the assumption (held by some scientists) that the earth was relatively young. This objection plagued evolutionary theory for decades, until the great age of the earth was demonstrated by new techniques early in the twentieth century.
Such problems with new discovery claims are likely to arise in a line of work when it has multiple competing theories or perspectives. The new discovery will typically fit well with some of these, and poorly with others. Poor fit may result in controversies that last a long time if each of the competing views is anchored in the supporting system of institutions in different ways. I suspect, for example, that some of the long-standing disputes between geneticists and paleontologists have continued because paleontologists are usually based in departments of geology rather than life science departments, and have so been less subject to the force of biological reason.
Much of the debates which surround models, methods, and concepts can thus be analyzed as negotiations over who shall bear the burden of proof or (conversely) what points must be included. Bruno Latour refers to this necessary inclusions as "obligatory points of passage".
The relations among many specialties then, can be seen in part as a network of more-or-less settled chains of obligation, each link consisting of a burden-of-proof relationship. And the contentious relations among some specialties can also be seen as disputes about unsettled burdens of proof.
I'll draw some of the implications of this way of looking at it another post.
Comments