What is Truth in Science?
by Chetan Parikh
 Mail this article to a friend
Previous Back  

In a great book, “The Truth Of Science”, the author, Roger G. Newton, writes on the meaning of truth in science.


“When we speak of the truth of something, the first point to note is that this something has to be a statement or an assertion; contrary to frequent usage, it makes no sense to speak of the truth of a fact or of a property. “The ‘facts’ themselves . . . are not true. They simply are,” William James reminds us. To insist on this is not pedantry or hair splitting. Formulating an assertion is attempting to communicate and therefore requires transmissible concepts and language: truth thus cannot be separated from human concepts and our linguistic apparatus. The history of philosophy is permeated with controversies that are of purely linguistic origin; apparently “deep” statements made in the language of one culture sometimes are untranslatable into that of another for want of a grammatical construction. Awareness of the pitfalls of language should make us cautious when dealing with “truth.”


Furthermore, it is useful to distinguish between the definition or the meaning of truth on the one hand and criteria for discerning it on the other. To be given a roadmap for finding the Grand Canyon is not the same thing as being told what it is. You may ask, Unless, I know what it is, how will I recognize it when I get there? But science will not provide a shiny plaque, engraved with an eternal description of The Truth; it will furnish only a powerful torch to help in the search and a sense of smell to tell when you get closer. Since this is not a treatise of philosophy, my primary concern will be with criteria, not with an abstract definition of truth, though it will help us also to look briefly at the latter.


What we most commonly mean when we claim an assertion is true is what Aristotle meant: it “corresponds to the state of affairs” described in it; a true statement is a “representation of nature,” or a mirror of it. The statement “this flower is red” is true if and only if, in fact, this flower is red. The logician Alfred Tarski even incorporated the “correspondence theory of truth” into formal logic. While to non-philosophers such a definition appears so obvious that it hardly needs mentioning, its meaning becomes more obscure in the context of fundamental science. The “state of affairs” described by “this flower is red” is clear enough, but it becomes less so when the statement is a scientific theory, especially when we recall that most theories are abstractions and simplifications with only approximate applicability in Nature. Is the “state of affairs” described by Newton's laws of motion a clear concept? What about the state of affairs in the EPR experiment? Although there is surely something right about the correspondence theory, as we ponder its significance it grows dim and murky. Because defining truth is not my main interest here, I won't dwell on its meaning further, especially since correspondence is difficult to use as a criterion.


One of my primary concerns is to ask whether truth is relative: are there, within science, varieties of truth is-truth in the eye of the beholder? An affirmative answer to this question lies at the core of the “perspectivist” theory of science and all the other relativistic opinions. The theory of truth at the basis of the constructivist approach is that of consensus. In this view, consensus among the members of a culture is not only a good way to get to the truth-within limits, a defensible proposition-but defines its meaning; in other words, consensus is its only criterion of validity. This is why constructivists, who consider science but an edifice erected by no other means than agreement among an in-group, regard Nature “out there” as irrelevant to the final determination of scientific truth. Indeed, according to them, Nature is nothing but this construction. Recall the proclamation by Latour and Woolgar: “Scientific activity is not ‘about nature’, it is a fierce fight to construct reality.” From this viewpoint, science is like a scaffold put up ostensibly for the close study of prehistoric paintings on the walls of a large cave; when you manage to penetrate the dense grid of girders and ladders, you discover that there are no paintings-the scaffold has been erected for its own sake and the artwork consists of no more than reports from the experts studying it.


Such a definition of truth leads inevitably to relativism, since the inhabitants of various eras or cultures are likely to adopt different truths and there is no way to compare or distinguish them transculturally. “We are prisoners caught in the framework of our theories; our expectations; our past experiences; our language,” Karl Popper grants.


But we are prisoners in a Pickwickian sense: if we try we can break out of our framework at any time. Admittedly, we shall find ourselves again in a framework, but it will be a better and roomier one; and we can at any moment break out of it again. . .


The Myth of the Framework is, in our time, the central bulwark of irrationalism. My counter-thesis is that it simply exaggerates a difficulty into an impossibility. The difficulty of discussion between people brought up in different frameworks is to be admitted. But nothing is more fruitful than such a discussion, than the culture clash which has stimulated some of the greatest intellectual revolutions?


For the philosopher Richard Rorty, the search for truth is nothing but “edifying conversation”; he ignores the fact, emphasized by Ian Hacking, that it forms a basis of action. There is little sense in pretending that truth has no consequences; it has different kinds of consequences in different realms. In science, “the influence of knowledge over action arises from its power of prediction,” John Ziman rightly points out, and this influence is considerable. “Conversation,” or story telling, is hardly an appropriate way to describe our search for it.


There is another reason why the consensus theory of truth is inadequate. I grant that the assertions of science, accepted into its body of truths after much testing, eventually become consensual, and that they must therefore be, in Ziman's terminology, consensible-formulated so as to be capable of being accepted or rejected on the basis of a scientific consensus. This is rarely their final status, however; they are always subject to a residue of potential doubt. If agreement were the definition of truth, consensible scientific statements would not be as precise as they usually are but would tend to be as vague as possible. We see examples of the value of vagueness every day in politics and international diplomacy-the less precise, the more likely an agreement. Furthermore, while general accord might do justice to the internal usefulness of truth among those consenting, it cannot account for the great functional value of scientific truth in the external world.


From the viewpoint of psychology, the consensus theory may be regarded as a manifestation of behaviorism grown rampant. Just as behaviorists define psychological states, feelings, and thoughts entirely by their outward signs in behavior, so the truth is defined by its behavioral demonstration-communal agreement. As a definition or as a sufficient criterion of truth, the consensus theory has to be rejected as a chimera.


The proverbial attempt by three blind men to describe an elephant may serve as a metaphor for our purpose: one man touches the beast's trunk, the second its tail, and the third a leg. The resulting descriptions of the animal will, of course, be incongruous and difficult to reconcile. Truth has many aspects, and even within science it may be approached from distinct angles with varying outcomes. Recall, for example, the two apparently different versions of quantum mechanics developed almost simultaneously by Heisenberg and Schrodinger. I rejected the inductive model of scientific theories-a clear-cut, unique process of induction leading from the data to a theory-advocated by philosophers of science for many years but no longer in fashion. More congenial was Popper's deductive view that theories are products of the human imagination, their validity depending on agreement between their consequences and observational facts. We therefore have no guaranty that there might not be several rival theories of equal validity. Are they all equally true? Would we not expect the scientific theories of Barrow's aliens from Alpha-Centauri to differ from ours? Does that leave the truth in the eye of the beholder?


In answer, let me begin with the observation, agreed upon by most scientists, that a theory is never true in an ultimate sense: it can, at best, only approach the truth. Thomas Kuhn initiated the contemporary vogue of relativism among sociologists of science by drawing on the analogy of Darwin, whose most revolutionary concept was the total absence of a goal in biological evolution. Similarly, Kuhn argued, the progress of science does not move toward anything such as truth; it simply evolves. This implies that we should expect the science of Barrow's aliens to be incommensurate with ours, which would be hard to reconcile with their ability to build space ships or long-distance communications, as we do. The science philosopher, Abner Shimony, replied to Kuhn's argument with an analogy (going back to Descartes) between attempts to discover the laws of Nature and attempts to decipher a message encoded in a text.


Suppose that we have such a text and that after numerous conjectures the tentative decipherment has become more and more coherent. The success may be no more than a series of coincidences, so that the tentative decipherment is completely on the wrong track. But somehow it is more plausible that a good approximation to the correct rules of encoding has been found than that the long run of successes has been coincidental. Kuhn's thesis that the truth plays no role in the progress of science is analogous to maintaining that progressive coherent decipherment could occur even though no such things as the initial message and the rules of encoding exist.”