What is constructive empiricism, and is it defensible?
May 12, 2019
Introduction
Constructive empiricism, an understanding of science pioneered by Bas Van Fraassen, is best summarized in Van Fraassen’s own words: “Science aims to give us theories which are empirically adequate; and acceptance of a theory involves as belief only that it is empirically adequate”. This articulation of constructive empiricism begs the question of how the unobservables posited by theories ought to be understood. In clarifying this query, Van Fraassen utilizes a semantic approach rather than a syntactic one. A focus on the syntactics of theories would differentiate them by the specifics of their linguistic expression; a theory’s syntactic identity is particular to the language and symbolism it is expressed in. A semantic focus, on the other hand, is model-centered: it is primarily concerned with what structures satisfy the axioms of the theory. Theories expressed in different syntaxes can accommodate the same models. Thus, the semantic approach does not render theories language-specific. Because theories accommodate certain structures and rule out others, they are capable of truth or falsity; they are statements about observable and unobservable phenomena that should be taken literally. A literal construal of a theory may elaborate on what an entity or relationship is but cannot change the logical relationships among the entities that the theory claims. As such, Van Fraassen accepts that theories make meaningful claims about metaphysics. However, he denies that scientists have any epistemic access to the unobservables claimed by theories.
In part, this epistemic claim frames what is involved in theory acceptance. Van Fraassen further argues that pragmatic considerations motivate theory acceptance above and beyond belief in unobservables does. In fact, one need not have any beliefs about the truth of a theory’s unobservable claims in order to accept it. One need only think that a commitment to the theory’s research programme will be vindicated. A belief that a theory will be vindicated only implies belief that the theory will prove useful, not that it is approximately true.
I have roughly articulated constructive empiricism as Van Fraassen describes it. I proceed to defend constructive empiricism against the no miracles argument put forth by Hilary Putnam, criticisms of its descriptive claims about science, and various charges of circularity. About the no miracles argument, I argue that empirical adequacy is sufficient to explain the success of science. Criticisms about constructive empiricism’s descriptive accuracy, I argue, mistakenly attribute to Van Fraassen the view that science has no place for gnosticism about theories. Moreover, critics devote too much attention to the idiosyncratic motives of individual scientists rather than to the actual criteria that theories are seen to compete based on. Finally, I pose three versions of the hermeneutic circle argument and outline possible solutions to them.
No Miracles Argument
Putnam, a proponent of the “no miracles” argument, claims that scientific realism is the only way for the success of science to be explained. Van Fraassen offers an alternative explanation: the scientific enterprise selects for those theories that are most empirically adequate and pragmatically virtuous. As an interlocutor in Ladyman’s Understanding Philosophy of Science argues, “Science is successful because thousands of people dedicate their lives to it and lots of what they do doesn’t work. It’s just that we only remember [and keep] the good bits”. Theories are not selected on the basis of their truth, as the truth of their claims about the unobservable is unknowable. Competition between theories hinges on which theories best describe the observable world—empirical adequacy.
A second response is to argue that explanation must end somewhere. To account for the success of science, one need go no further than to recognize the empirical adequacy of widely accepted theories. It is not apparent why the empirical adequacy of these theories must be further accounted for. To suggest that a theory’s claims about unobservables must underlie its adequacy in predicting observable phenomena begs the question of what accounts for these unobservable phenomena. The demand for explanation must have some end, and there is no less arbitrary explanatory limit than the theory’s conformity to observed regularities.
Descriptive Deficiencies of Van Fraassen’s Account
A critic of constructive empiricism might argue that it is not descriptively accurate; many—if not most—scientists conceive of their profession as truth-seeking. This assessment is premised on an interpretation of constructive empiricism as a sociological thesis regarding the attitudes and beliefs of scientists about their practice. Gideon Rosen suggests that such a sociological interpretation might hold that constructive empiricism implies widespread suspension of judgment about the truth of unobservable claims. Van Fraassen distances his position from this reading by clarifying that the constructive empiricist picture of science need not exclude scientific gnostics—those who have beliefs about the veracity of unobservable theoretical claims. In Van Fraassen’s words, “it is not part of constructive empiricism to say that the adoption of such additional beliefs is irrational—just that it is more than what is involved in scientific theory acceptance”. This clarification aside, Rosen argues that if Van Fraassen’s account were meant to describe science as it is conceived of by its practitioners, he would have offered relevant sociological evidence. However, as Van Fraassen argues, scientists’ beliefs about what they are doing may perhaps be misaligned with what they are actually doing—pursuing empirical adequacy. By this claim, Van Fraassen does not mean to take a behaviorist position that intentional aspects of science are unobservable and therefore irrelevant to constructive empiricism’s empirical adequacy. Rather, he means that scientists’ aims, intentions, and beliefs about their own practice is not synonymous with their shared understandings of scientific practice in general. For instance, consider some scientist, Joe, who claims to be motivated by the pursuit of truth. If told that some respected contemporary, Elizabeth, is agnostic about the truth-likeness of her theoretical commitments, Joe would likely not dismiss the work of his contemporary as unscientific. Van Fraassen suggests that a result of this kind would illustrate the difference that he alleges between self-understandings about idiosyncratic practices and the defining characteristics of science itself. He offers as another example of this difference Clausewitz’s doctrine of war: “[the aim of war] is the continuation of diplomacy by other means”. The criterion of success for a soldier is the soldier’s success in continuing her country’s diplomacy. Each soldier’s motivations are perhaps divergent from the proper goal of war. Some may think they are fighting for glory, others may seek the domination of a despised enemy, etc. In any case, the enterprise that they participate in is concerned with the continuation of a country’s diplomacy. Similarly, a scientist may think that he is concerned with what is true when, in fact, the economy of scientific research that he participates in and conforms to is guided by empirical adequacy among other pragmatic criteria. While the understandings of practitioners in the field certainly influences how one ought to define the practice, what is widely recognized as essentially within the spirit of the practice is more revelatory about what is minimally required of a participant than is what participants generally regard themselves to be doing. Van Fraassen seems to be arguing that the pursuit of empirical adequacy is essential to the spirit of science, and scientific gnosticism is an idiosyncratic and non-essential ancillary to theory acceptance. In fact, he says as much in accepting the following interpretation put forth by Rosen: “‘The aim’ of attempting to carry through the constructive empiricist interpretation of science ‘is to show that even though he sees no reason to believe what they say, the [scientific agnostic] need not be driven out from the paradise that Boyle, Newton, Mendel, Rutherford, Bohr and the rest have created’”. This admitted aim begs the question of which scientific phenomena Van Fraassen uses to abstract his vision of science, a consideration which gives rise to a version of the hermeneutic circle.
Hermeneutic Circle
The hermeneutic circle refers to the problem of defining science in relation to its constituent phenomena. In order to answer the question of what science is, one must examine the phenomena understood to be scientific. Whatever these phenomena essentially share in common is what science is defined by. Further, what phenomena one looks at to make this determination depends on whether one is a scientific realist or a constructive empiricist. But, to hold such a position is to assume what is under question: what science is. The question is how one can identify scientific phenomena without having some preconceived notion about what one should be looking for.
I find Van Fraassen’s response to this ambiguity wholly satisfactory, as any other linguistic concept is identifiable in the way that Van Fraassen describes. Philosophical dialogue is only possible insofar as there is common ground from which participants can proceed. This common ground can be found in cultural presuppositions, values, and the identification of relevant facts. This framing has a few implications on how Van Fraassen’s defense of constructive empiricism should be understood. He is conditioned to see science through the constructive empiricist’s lens and thinks that this lens is appropriate for identifying rationality in the scientific practices observed through this lens. Though the lens influences what phenomena are examined and accordingly how science is understood, this circularity poses no problem. Any participant in a scientific dialogue is bound to have some conditioning that predisposes him to study phenomena of interest, guided by a preexisting admiration for particular qualities and rationalities.
Two additional versions of the hermeneutic circle are articulated in the Stanford Encyclopedia of Philosophy. First, if a theory of observability is necessary to identify what is observable and empirical adequacy is assessed in terms of observable phenomena, a theory of observability can name its own terms of empirical adequacy. Empirical adequacy becomes relative; one could adopt a theory of observability that affirms the empirical adequacy of any theory that one is committed to accepting. Van Fraassen’s reponse to this relativity concern is simple but sufficient: observability is a theory-independent notion. A second concern is that to justify a theory of observability requires circular reasoning. Whether something is observable depends on a theory of observability, which serves a guide to empirical adequacy and, thus, weighs on the acceptability of that very theory. I am not prepared to comment on this claim of circularity, but, on face, it seems that this circularity can be resolved in a similar way that the circularity implicit in defining science can be. Some common intuitions are shared between people about what is observable. This common ground forms the basis of any theory of observability, which provides grounds for assessing empirical adequacy. Even insofar as views on observability diverge, there is sufficient related common ground for deliberation to occur about the merits of competing theories. For instance, one theory of observation might fail to account for the importance of the subject to whom a phenomenon is observable. Another might have the implication that a pattern seen through a kaleidoscope is afforded similar epistemic status as something seen using bare eyesight. The circularity can be resolved insofar as dialogue participants have, as a result of their conditioning, common intuitions about various theoretical implications.
Conclusion
I find constructive empiricism to be a defensible and attractive account of science. Putnam’s no miracles argument takes for granted that some explanation for the success of science is needed beyond the empirical adequacy of theories. The descriptive elements of Van Fraassen’s account also seem to successfully distinguish idiosyncratic motives from the collective practice’s manifest criteria of success. Finally, I find the alleged hermeneutic circularity to be a non-issue; while it is perhaps inevitable, the common ground needed for philosophical dialogue to take place is sufficient to resolve the problem that the hermeneutic circle may pose.