The Brilliant Hallucination of Untested and Unmeasured Theory

The Opinion piece by Michael Riordan (Physics Today, August 2003, page 50 ) and its follow-up discussion in the December 2003 issue contained some strong statements about the nature of scientific knowledge. Many of those statements touch on the cornerstones of philosophy of science, and therefore merit a lengthy debate. However, the one I find the most interesting is that theorists who generate empirically indemonstrable models are doing metaphysics rather than regular physics. I think this view and its underlying assumptions are dangerous in that they may communicate an inaccurate image of science.

The overall view conveyed by Riordan’s opinion and some of the letters is that scientific knowledge is somehow superior to other kinds of knowledge because it holds a deep connection to reality. The comments also convey the idea that only empirically corroborated theories are purely scientific because only through experiments can we discover scientific truths. The disregard of the importance of metaphysics in scientific discovery is disconcerting.

Every scientific enterprise is deeply rooted in a metaphysical context. That we subscribe to a particular “ism”—pragmatism, relativism, positivism, reductionism, and so forth—is in itself a clear choice of a philosophical framework. We cannot say, a priori, that a particular framework is better than the others, but we may, through pure rhetoric, persuade some of our peers that one is more appealing than the rest. In some way, choosing a specific philosophic approach is like choosing a religious belief: a personal and subjective decision.

Embracing the philosophical content of theories is not equivalent to relaxing our criteria of what is and what is not scientific. It is simply acknowledging that science has a great structural complexity shaped by many theoretical and discursive levels that go well beyond Galileo’s “Book of Nature.”

In my opinion, the interplay between science and metaphysics is crucial in the evolution of scientific theories. Often, the metaphysical nature of a theory is what dictates the path a research program will follow. ¹ In Galileo’s day, the choice between the Tychonic (anthropocentric) and Copernican (heliocentric) models of the universe was made on philosophical grounds. Mathematically, both models were identical, so no experiment could have distinguished one from the other. ² Similarly, the choice between the deterministic and indeterministic versions of quantum mechanics was made not because of experiments but because the second is consistent with the philosophical environment that permeated the early 20th century.

Denying the role of the metaphysical backgrounds of many theories and models is not only a rough and inadequate approach to the study of the history of science; it is also a dangerous premise for the future, in the same way that Riordan’s “Platonic physics” is considered a dangerous relaxation of scientific criteria. If science is independent of philosophy in that it unveils truths, then scientific facts are unquestionable. That view of science puts researchers in the position of priests of the oracle’s temple and confers on them a special connection to the mysteries of nature. History has shown that such strict discourses are threatening because they entail the rejection of many alternative research programs. With the complex challenges facing our societies in the future, it is our obligation as scientists to explore the viability of alternatives. Adopting a reductionist posture on scientific discovery will not allow us to explore with calm the options that are on the table.

Riordan makes a compelling point about the nature of science and its connection with reality. I agree that too much metaphysics results in the untimely death of physics. However, too little produces a rigid body of tautologies and universal facts, which can in turn evolve into a dangerous discursive weapon for those with a lax scientific background. A clear example of this occurred in the history of economics. When Leon Walras, Carl Menger, and William Jevons developed the foundations of what is now called neoclassical economics, they used some ideas from physics—specifically mechanics—to justify the scientific nature of their approach. ³ Almost 150 years later, many economists (Pierre Bourdieu, to mention one) question both the validity of that connection and the argument that underlies it. ⁴ Balance is an essential matter. The question is whether we can reach such a balance; as I see it, we can. In my opinion, the problem is solved by distinguishing what is empirically proved from what is pure metaphysical speculation. One cannot say with total certainty whether today’s metaphysics will become tomorrow’s everyday physics. By clarifying the distinction and avoiding quantifiers like “more” or “less” scientific, we leave the doors open for future revolutionary progress.