ABSTRACT
In late summer 1995 the French government started to detonate a series of atomic explosions in the mid-Pacific. Clearly there were a number of reasons for these events, not least the assertion of nuclear independence by the French authorities. But the reasons surrounding the technology of nuclear devices are also instructive. For all the computer-modelling of whether and how nuclear devices will operate, scientists still depend greatly on tacit knowledge in predicting whether they will work. Designers of nuclear weapons know full well that computer predictions are only approximations of the reality that will ensue when a detonation takes place (McKenzie and Spinardi 1995). They know, for example, that contingent factors such as ambient temperatures, the ages of the weapons and irregularities in the production of weapons, cannot be accommodated in their computer-based modelling. These factors have be taken into account by judgements and skills accumulated over long periods and in different places. To an increasing extent such tacit knowledge can be built into the more abstract and quantitative modelling of the processes and relations involved. But such qualitative and unformulated judgements still have to be made. Note, incidentally, that tacit knowledge cannot be equated with ‘lay’ understandings. A scientific prediction as to how the causal powers of nature combine with one another also depends on unformulated tacit judgements. Everyone uses tacit knowledge all the time, but some forms of this knowledge are consistently neglected and allowed to decay while others, such as those of the designers of nuclear weapons, remain relied upon for the development of scientific authority.
