Constructivism and Conventionalism
CONSTRUCTIVISM AND CONVENTIONALISM
"Conventionalism" and "constructivism" are kindred, often overlapping positions, asserting that the subject matter of some area of inquiry is not fully mind-independent. Conventionalism and constructivism are not well-defined names of positions but labels adopted—as often by critics as by advocates—to emphasize one positive aspect of positions in a wide range of areas; consequently, these terms group together a variety of positions with varying motivations. In general, the label "conventionalism" is applied to positions that claim the truths in some area are so in virtue of the conventions of a linguistic or conceptual scheme, while "constructivism" emphasizes that a position assigns to the cognitive faculties of humans some role in "making" the objects or facts in the area in question.
Conventionalists claim either that the truths of some subject matter—such as mathematics or logic, or of a certain sort, such as necessary truths, or some dispute, such as whether Euclid's parallel postulate holds of our physical world—are matters of convention rather than of how the world is independent of mind. Some extreme versions of conventionalism take the fact that it is a matter of convention what our words mean (we could have used cat to designate Napoleon Bonaparte) to show that all truth is conventional. However, its being a convention that "Napoleon" names Napoleon hardly makes it conventional that Napoleon was defeated at Waterloo. An interesting conventionalism must assert something more than the conventionality of word meaning and must rest on something more than wild inference from it.
One area in which conventionalism is familiar, though controversial, is necessary truth. This was one cornerstone of logical positivism; from the seeming a priori nature of necessary truths, the positivists argued (some would say claimed) that since a priori knowledge cannot be of (mind-independent) facts, necessary truths must be analytic, which they understood as true by definition. Given that mathematics is also a priori, their argument was applied there as well. This sort of epistemological argument is typical of conventionalist views: Arguing that our methods for ascertaining what is so in some area could not give us knowledge of a mind-independent world, they claim that this knowledge would not be problematic on the assumption that what is fundamentally under investigation are our conventions. Some sorts of conventionalism are also supported by metaphysical considerations such as naturalistic concerns about what, in the mind-independent world, could make for the relevant sort of truth. This sort of argument is common to necessary truths, mathematics, ethics, and other areas with normative import; plainly, such arguments need to be supplemented with an account of how it is that conventions can provide the relevant features.
Saul Kripke's arguments that there are necessary truths that are a posteriori—and, so, not analytic—seemed to some to undermine conventionalism about necessity (Kripke 1980). It has, however, been argued that conventions could explain the necessity of these truths without the truths themselves being analytic—that is, true by convention (Sidelle 1989). This may indicate that in general conventionalism, with respect to a subject matter, does not require that all target truths themselves be analytic but only that conventions be responsible for the features that purportedly cannot be adequately handled within a realistic interpretation.
Aside from the claim that certain truths are so by convention, another common conventionalist position is that some dispute is a conventional rather than factual matter. Jules Henri Poincaré's famous conventionalism about geometry is of this sort. He claims that the choice among systems of geometry, for describing the physical world, is not an issue of which is true but of which is most convenient or useful. By adopting any of them, we could modify our physics so as to have equally full and correct descriptions of the world; indeed, this last claim is the basis for his view that the issue is conventional rather than factual. Rudolf Carnap offers a similar view about ontological disputes between, for instance, phenomenalists and materialists. Both of these views illustrate that "conventional" does not as such imply "arbitrary," as pragmatic differences may be quite genuine; we can also see that the plausibility of conventionalism in some area depends largely on how implausible it is to claim that the issue, or truth in question, is a matter of mind-independent fact.
On a more local level, some disputes can appear "purely verbal," as perhaps whether some politician is conservative. When this is plausible, the issue may be said to be a matter of convention or choice rather than fact. The conventionalism of Poincaré, Carnap, and others is akin to this, only in a wider application. In book 3, chapters 7 and 11, of his Essay, John Locke speculates that many of the "great disputes" are of this sort.
As applied to areas in which the truths are well established (mathematics or logic, for instance), conventionalism is fundamentally a deflationary interpretive position, urging that we not mistake the metaphysical status of these truths. Applied to areas of controversy—ontological or essentialist claims, or whether whales are fish—conventionalism claims that disputes here can only be over what our conventions in fact are, or what they should be, either pragmatically or perhaps morally. In either case, if conventionalism is right, our focus and methods of investigation—and certainly our understanding of what is at stake—for the questions at hand would probably require alteration.
Thomas Kuhn, by virtue of his The Structure of Scientific Revolutions, may be considered constructivism's leading protagonist of the mid-to-late twentieth century, despite not adopting the label himself and expressing unease at having it assigned to him. He writes of scientists within different paradigms—roughly, methodological and theoretical traditions or frameworks—as studying different worlds and of their paradigms as in "a sense … constitutive of nature" (Kuhn 1970, p. 110, chaps. 10, 13), at least suggesting a constructivism about the world studied by science. Kuhn's major concerns are epistemological; he argues that scientific procedure is deeply theory laden and encodes ontological and theoretical commitments that it is incapable of testing. How, then, can such a method give us knowledge of the world? Those who see a constructivist in Kuhn have him answer that the world under investigation is itself partly a product of the investigating paradigm. This puts Kuhn in the tradition of Immanuel Kant, except that the features we "impose" upon the phenomenal world are not (as for Kant) necessary for the possibility of experience, but, rather, contingent features of current science. It is important to note that, even as interpreted, this constructivism does not have scientists making the world out of whole cloth with their paradigms; rather, there is something mind-independent that "filters through" the conceptual apparatus of the paradigm. This is a central difference between constructivism and idealism. The object of scientific study is, however, not this mind-independent world, but rather that which results "through the filter."
Other philosophers, as well as historians and sociologists of science, have taken the supposedly arational or nonobjective features guiding scientific judgment to establish that scientific truth is relative to one's background theory or paradigm. This is sometimes then articulated as the view that these theories or paradigms in part "make" the objects of study—that is, as constructivism. Indeed, many positions that formerly would have simply been called relativist came, in the late twentieth century, to be called constructivist by their protagonists; arguments in their support tend to be of the familiar relativist sort and thus have the same strengths and problems. It should be noted that neither constructivism nor conventionalism need take a relativistic form.
Another philosopher associated with constructivism is Nelson Goodman, due largely to his Ways of Worldmaking. Goodman argues that no sense can be made of the notion "the (one) way the world is"; rather, there are lots of ways the world is, depending on the conceptual apparatus one brings. This sort of position is found in many philosophers since Kant, often argued on the trivial ground that one cannot describe or investigate the world without using a system of representation, therefore (sic ) the world investigated is not mind-independent but partly constructed by our conceptual scheme. This is sometimes added to, or confused with, the relativistic considerations mentioned above. What needs to be explained is how we are supposed to get this substantive conclusion from the banal premise. Why can't the objects represented by the elements of a system of representation—by the name "Tabby," say—be wholly and utterly mind-independent? And even if we add the fact that there can be different schemes of representation, why can it not simply be that they pick out different features of a mind-independent reality? What gives Goodman his special place is that he supplements this argument with the claim that different schemes may be such that their claims conflict with each other, but there can be no grounds for maintaining that one is correct and the other not. Goodman uses as examples the claims that the planets revolve around the sun and that the sun and other planets revolve around Earth. Both, he claims, must be judged as correct (within the appropriately formulated total systems), but they cannot simply be seen as two notationally different descriptions of a single world (thus differing from Poincaré's conventionalism). The success of this argument depends on whether one can simultaneously make out that these claims genuinely do conflict with each other and that, so understood, neither of them can be judged to be true while the other is false.
While both Kuhn and Goodman offer relatively global constructivist positions, there are constructivists about essences, moral and aesthetic properties, mathematical objects, and in principle anything. The same is true of conventionalism. Both conventionalism and constructivism are motivated primarily by negative considerations against a realistic understanding of the subject matter in question; this is sometimes supplemented with positive arguments that by understanding the matter as concerning our conventions or choices we can get a better explanation of the phenomena at hand. Often, the negative arguments are very quick and fail to fully consider the range of options available to realists (Scheffler  presents good discussion), and sometimes they fail to consider whether their positive proposals actually fare any better. Plainly, the plausibility of these positions depends on how well these arguments can be made out, and this may vary drastically across the different subject matters for which conventionalist and constructivist proposals have been offered. Additionally, if these positions are even to be candidates for serious consideration, defenders must be prepared to offer further proof. Conventionalists must specify some sense in which that which is purportedly so by convention would have been otherwise had our conventions been different, and constructivists must describe some sense in which the purportedly constructed objects would not have existed without our input.
Carnap, R. "Empiricism, Semantics, and Ontology." In Meaning and Necessity. Chicago: University of Chicago Press, 1947. A classic exposition of the conventionalist's treatment of apparently ontological questions as fundamentally linguistic.
Dummett, M. Frege: Philosophy of Language. 2nd ed. Cambridge, MA: Harvard University Press, 1981. Chap. 16, esp. pp. 560–583. An interesting presentation of what may be seen as a constructivist position on objects or a conventionalist position on identity criteria.
Goodman, N. Ways of Worldmaking. Indianapolis: Hackett, 1978. For useful discussion, see C. G. Hempel, "Comments on Goodman's Ways of Worldmaking, " and I. Scheffler, "The Wonderful Worlds of Goodman," both in Synthese 45 (1980): 193–209.
Horwich, P. "A Defence of Conventionalism." In Fact, Science and Morality, edited by G. MacDonald and C. Wright. Oxford, 1986.
Knorr-Cetina, K. "The Ethnographic Study of Scientific Work: Towards a Constructivist Interpretation of Science." In Science Observed: Perspectives on the Social Study of Science, edited by K. Knorr-Cetina and M. Mulkay. London, 1983.
Kripke, S. Naming and Necessity. Cambridge, MA: Harvard University Press, 1980.
Kyburg, H. E. "A Defense of Conventionalism." Noûs 11 (1977): 75–95.
Latour, B., and S. Woolger. Laboratory Life: The Social Construction of Scientific Facts. 2nd ed. Princeton, NJ: Princeton University Press, 1986.
Lewis, D. Convention. Cambridge, MA: Harvard University Press, 1969. Argues that sense can be made of the idea of nonexplicit conventions, contra Quine (1953).
Locke, J. An Essay concerning Human Understanding (1689). Locke's discussion of substance in book 3, chap. 7, classically expounds his conventionalism about essence and kinds.
Niiniluoto, I. "Realism, Relativism, and Constructivism." Synthese 89 (1991): 135–162. Critical discussion of constructivism in the social sciences.
Poincaré, H. Science and Hypothesis. London: Walter Scott, 1905.
Putnam, H. "The Refutation of Conventionalism." In Mind, Language, and Reality. Cambridge, U.K.: Cambridge University Press, 1975.
Quine, W. V. O. "Truth by Convention." In The Ways of Paradox and Other Essays. New York: Random House, 1966. A classic argument against the possibility of truth by convention alone.
Quine, W. V. O. "Two Dogmas of Empiricism." In From a Logical Point of View. Cambridge, MA: Harvard University Press, 1953. Classical argument against the analytic-synthetic distinction and so against truth by convention.
Reichenbach, H. Experience and Predication. Chicago, 1938.
Scheffler, I. Science and Subjectivity. Indianapolis, IN: Bobbs-Merrill, 1967. Responds to arguments based on the history of science against objectivity in science, particularly Kuhn's, which are often used to bolster constructivist positions about the objects of scientific study.
Sidelle, A. Necessity, Essence, and Individuation: A Defense of Conventionalism. Ithaca, NY: Cornell University Press, 1989.
Alan Sidelle (1996)