This is an important book. If Davis Baird is correct, and scientific instruments really are knowledge, then the implications for the philosophy of science are significant. In Thing Knowledge, Baird puts forth the intriguing thesis that standard philosophies of science, with their emphasis on theories and how they express knowledge, is fundamentally incomplete. He argues that we need to augment what he calls this text bias with an epistemology of instruments, for as he puts it, “These material products are constitutive of scientific knowledge in a manner different from theory, and, not simply ‘instrumental to’ theory” (1).

To make his case, Baird concentrates on three different kinds of instruments: models, instruments that exemplify working knowledge, and finally, measuring instruments. Models, such as orreries or Crick and Watson's model of a strand of DNA, provide representations of nature and show us, as it were, what we have found to be the case. Instruments exhibit working knowledge when they create phenomena. Baird sees them in a pragmatist's light as producing effective action: “A fundamental difference, however, is that with instruments, the action has been separated from human agency, and built into the reliable behavior of an artifact” (12). Finally, Baird considers instruments that produce measurements. The instruments combine the attributes of the first two kinds. Thus, Baird claims, “Measurement presupposes representation, for measuring something locates it in an ordered space of possible measurement outcomes. … At the same time, a measuring instrument has to do something and do it reliably” (12).

Baird's fundamental move is to argue that as scientific instruments become increasingly sophisticated, they contain more and more human knowledge. But, the knowledge is not evident. It is not transparent. The knowledge itself is not accessible. The instruments do what they do because of that knowledge, but the knowledge as propositional content is not available. Why the instrument does what it does becomes, as instruments evolve, less and less obvious and requires less skilled technicians. Thus, with an orrery we have a model of the solar system. It does not tell us that it is a model of the solar system, it just functions as one. If we put it in a space capsule and hurled it far beyond our galaxy where it was discovered by an alien race, it is not clear they would know what it was.

Baird's aim is to show that instruments have epistemic content that is different from, and yet similar to, the epistemic content of theories, and that understanding that is important to a full account of science. It is important for several reasons, but there are two he emphasizes. First, because of the thing knowledge of instruments, instruments can lead to further developments of theoretical knowledge. To illustrate this, he looks to the example of the indicator diagram that exhibits the pressure and volume inside a cylinder in a working engine. What actually went on inside the cylinder was not known at the time of the invention of the indicator diagram, but the indicator did its job of providing reliable information. Baird notes that the history of the indicator “shows how the stable empirical phenomenon produced by the instrument—irrespective of how this phenomenon was interpreted at the time—provided the necessary empirical foundations for fundamental theoretical speculation” (171). Second, as instruments become commercialized, the gift of free exchange of scientific ideas diminishes. As the gift economy of science disappears, the ethos of the free exchange of ideas is negatively impacted. Scientists are less eager to send an example of an instrument they have made if it can be patented and then sold. In this environment, we are not far from the patenting of ideas. Science as a community of cooperating and energetically competing individuals seems to be dissolving into the economy of a capitalistic market. Surely this cannot be a good thing in the long run.

What is nice about this book is that Baird makes us look seriously at instruments. Hacking tried to push us away from abstract theorizing about science to scientific experiments. But the move to experiments was not completely successful because philosophers still attempted to bring the old theoretical apparatus of theory confirmation, etc., to the problem of how science works and still talked about experiments, rather than looking at the material details of how they work. Here there is no idealized fantasizing about experiments. This is the next logical step in our journey toward the goal of understanding the evolving nature of scientific inquiry.

This work is worth taking seriously because the ideas Baird is trying to get across are deeply jarring. We are too used to thinking about knowledge from a propositional point of view. Baird has set us a real problem that old ways of thinking cannot readily handle. We need a new vocabulary to unpack fully these ideas, and we need more of the kind of analysis Baird offers.