According to Michael Rea, naturalism is a combination of a method and an assumption. Methodologically, naturalism is the attempt to understand the world through the natural sciences. The assumption, which, theoretically at least, can be overturned at any moment, is that everything that exists is material. The assumption can, in principle, be overturned at any moment because natural sciences don’t claim to have all data at their fingertips. Fairies could be found at the bottom of an English garden and God might appear from behind a far off planet, but neither of these events seems plausible given only the data that the natural sciences have produced so far.
Rea defines naturalism as a research program. As such it does not fit the bill as a substantive thesis. To say that naturalism is a substantive thesis would be to suggest that it is a thesis about metaphysics, epistemology or methodology. Rea argues that any such thesis is incompatible with naturalism. Consequently, naturalism should be seen as a research project.
Metaphysical naturalism would entail a thesis about what is real. However, on its own view, naturalism asserts that the natural sciences do not predetermine the outcome of observation. Naturalism must be compatible with anything that is discovered. Therefore N is not a metaphysical thesis.
Neither is N an epistemological thesis. On this view the basis for science is empiricism. Rea concludes that this is self-defeating since such a thesis in favor of empiricism is itself unjustifiable by scientific method. Furthermore, such a thesis is unjustifiable by empiricism. Methodological theses are either metaphysical, about the nature of inquiry, for example, or epistemological. Consequently, argues Rea, they fall into the prior categories.
Rea concludes that naturalism is a research program. A research program is “a set of methodological dispositions” by which inquirers discern between trustworthy and untrustworthy sources of information. Research programs are held by individuals and groups, self-consciously or not, and are taken as basic. As such they are subject to defeaters, but cannot be adopted due to new evidence since RPs determine what evidence is trustworthy.
As a research program, naturalism is, according to Rea’s central thesis, without a rational foundation. Since evidence is weighed according to the prior commitments of a RP the strategy for defeat is pragmatic. Rea seeks to demonstrate that the costs of adopting a naturalistic research program are too high.
Naturalists cannot, by their own lights, be justified in accepting two metaphysical views that many philosophers—naturalists in particular—very much want to accept. Those views are realism about material objects (RMO) and materialism. I will also argue that, on the assumption that standard naturalistic arguments against mind-body dualism are successful, naturalists must give up a third thesis: realism about other minds (ROM).
Realism about material objects is the view that material objects have mind-independent existence. Materialism is the view that nothing exists except material objects within space and time. Realism about other minds is the thesis that other minds apart from one’s own exist.
The discovery problem is the fact that “intrinsic modal properties seem to be undiscoverable by the methods of natural sciences.” Modal properties are properties about objects that involve necessities and possibilities. We are somehow able to know what these properties are. Naturalism requires that such properties be discoverable through the natural sciences. However, the problem arises, suggests Rea, when one considers that realism about material objects requires that modal properties be intrinsic.
Material constitution is when an object, b, is constituted by object a. For example, when a collection of particles constitutes a human being. The question is: Is it right to say that a is identical with b or that both a and b exist as two distinct objects? It appears strange to suggest that my particles are distinct from my being a human being. Yet it is also not quite right to suggest that I am a different human being than the one I was when I was two years old, when I was made up of entirely different particles.
Persistence conditions are what properties are essential properties. If I blew up in a puff of smoke and the dust dispersed into the atmosphere would I have persisted to exit? While one might suggest that the particles that made me persist in some other form (dust) it is difficult to say that I persisted and highly implausible to claim that I persist as a human being.
To say that something belongs in a certain kind is to say that I am a certain kind of thing, a human being. If I blow up and turn to dust then I cease to obtain any properties that make me a human being because I fail to belong to that kind. What comes first—knowledge of an object’s persistence conditions or of its kind—is difficult to tell. They, at least, are conceptually distinct, but one seems to depend the other.
The epistemological problem arising from the material constitution problem is what justifies our attempt to solve it. In other words, if we can solve the problem we need to supply adequate grounds for the conclusion. This, suggests Rea, cannot be done on naturalism.
Initially there are four possible answers. First, skepticism suggests that modal properties are not intrinsic, but extrinsic. Rea argues that, given persistence conditions (that we can tell whether an object continues to exist or not), we are forced to admit to knowing at least one intrinsic modal property. To abandon that, we must abandon RMO. Second, one might suggest that beliefs in modal properties are conceptual and tied to linguistic conventions. The latter gives us a truth regardless of empirical observation that is then comported with an empirical observation. Such a view, suggests Rea, commits its proponent to modal anti-realism if they are also naturalists. This is because, “there is no naturalistically acceptable basis for thinking that reflecting upon conceptual or conventional truths is a way of acquiring information about the world’s intrinsic modal structure.” If modal properties are intrinsic then they cannot be grounded in conceptual truths since those truths are not analytic and cannot be justified by the natural sciences. Rea is convinced that the latter two possible answers cannot overcome the problems facing the former. Intuition, the suggestion that beliefs in modal properties can be grounded in rational intuition, does not do the work for the naturalist since the view presupposes that such beliefs can be grounded in concepts that are not analytic. Science, the final view, is troubled by the fact that it is only if modal properties of objects are intrinsic to objects that we might contend for RMO, but, on the naturalist view, scientific method has little chance of showing how we might have access to such properties.
“To be able to apprehend an object before you, you must think it is presently and continuously existing. But to know if something is persisting in its existence, we must know something about what we take it to be—what kind of thing it is—in order to know its essential persistence conditions.”
These four responses do not exhaust the discovery problem’s responses and some of them appear to be open to naturalists. Rea, however, is pessimistic about this claim. One response is to suggest that objects have proper functions. An object, on this view, has intrinsic purpose that is detectable by observation. A proper function is an objective primary function of an object. Objects have parts that they are, objectively, supposed to perform. How do we explain how some object appearing to us within space-time appears to contain some x that at least a part of that is supposed to reach some end? R appears to be supposed to perform in some way. Its “supposed to perform” quality supervenes on its properties such that we assume that the object (or its part) has some objective function. If this is the case we can make claims about properties based on these empirical observations about what a thing is for. And that counts as a modal property. Such a property may even be considered to be an intrinsic property. If the parts function to maintain the object by fulfilling an objective function then such a property can be construed as intrinsic. And since no external object is required in order to empirically know such a property it appears we have a solution to the discovery problem.
First, Rea suggests that though this might be plausible when concerned with biological material, it is not at all clear that this works for inanimate objects. This is because rocks do not appear to have any apparent proper functions. However, Rea notes the lack in his objection since animate objects might provide paradigmatic examples of material objects. We could merely assume that all objects have proper functions even though we do not know what they are. We know enough about proper function within animals to infer that it is likely that other objects have them as well.
Rea’s main objection is based on what is meant by proper function. It is important to notice that proper function is not merely apparent function, but the function an object is supposed to perform, a proper function. If this is the case then proper function is not empirically detectable.
Apart from designed objects (a tacichscope for example), objects do not exhibit a anything we can be sure is an objective phenomena about the function of the object. We might speak about them that way, but we cannot, by empirical method alone, strictly know that that is the case.
“None of the available suggestions are compatible with the idea that biological organisms have natural proper functions if biological organisms do not; and there is no reason think that there are promising accounts of what might make a function a proper function that are not already available in the literature.”
A part of a system has proper function if it has a function and that one of its functions is metaphysically privileged, objective, something it must perform etc. Rea concedes that the former is possible (we can see that something has functions), but argues that the latter is undetectable by empirical method.
Rea’s surveys three suggestions for solving the problem – statistical normativity, etiology and system flourishing. Statistical normativity (SN) theories suggest that object “X has F as its proper part function just in case X belongs to a kind whose members statistically normally have F as their function.” Rea suggests that this may well get us to the point of saying that most Xs have function F as part of S, but it cannot support the idea that F is metaphysically privileged. If Xs began to have not Fs as their most normative function, then that would become the proper function, but that is not what we mean by PF. We mean, rather, that the function is what that object is supposed to be doing. If a genetic disease made human beings blind and, instead of seeing, all eyes produced pleasure by poking, we would not say that pleasure by poking is an objective purpose for eyes.
Etiological Accounts (EA) “the… proper function of X is to F just in case X was designed to F.” To give this a naturalistic spin requires that design is a purpose of species to make offspring fitting for the world. “It is the PF if an item X of an organism O to do that which items of X’s type did to contribute to the inclusive fitness of O’s ancestors, and which caused the genotype, of which X is the phenotypic expression, to be selected by natural selection.” Again, this might tell us what has enabled organism O to be best fitted to function, but it does not tell us that F is what X was supposed to do. Consider divergent evolving creatures. Let’s say two kinds evolving from one kind in two contrasting ways. One kind evolving eyes to see, the other to sense only vague light in otherwise unchanged species. We surely would not say that both evolved to proper function for O. We would say that some O are blind or that their eyes are not properly functioning. “It simply does not follow from the fact that a thing has a certain kind of natural causal history that it is in some metaphysically important sense supposed to do whatever it is that causal history has determined it to do.”
Aristotelian accounts further the claim that a PF of X is whichever F that contributes to the well being of the system. The use of this proposal, for naturalists, depends on their use of Wright’s Law. WL stipulates that X’s F in S must be causally explained by the existence of a causal law linking X to S. So a PF for X is known by linking its cause to the effect of flourishing in a system. Rea suggests such a system might be substituted for “ecosystem.” Then organisms within the system have proper functions that help the ecosystem flourish. If this is what we mean by proper function then all sorts of strange PF become legitimate. A fish serves as food for birds. Is this the proper function of the fish? This seems not to be treating PF in the same way as is under discussion.
Pragmatic arguments: Constructivism is the view that no property of a material object is intrinsic to the object in question. This includes both sortal properties, properties that correspond to sorts or kinds. Constructivism suggests that sortal properties are mind-dependent such that the necessary condition for sortal properties being the way they are is the relationship they bear with minds. Rea concludes that this definition does not exclude the existence of material objects, thus not entailing idealism.
 Michael Rea, World Without Design: The Ontological Consequences of Naturalism (Oxford: Oxford University Press, 2002), 3.
 Ibid., 6.
 Ibid,. 8.
 This is argued convincingly in: Bradley Jay Strawser “Rea’s Revenge and the Persistent Problem of Persistence for Realism” Philosophia (2011) 39: 375-391.
 Which way they depend is unclear “I’m inclined to speak of properties ‘giving’ objects their persistence conditions; but I’m also inclined to say (e.g) that objects are grouped into kinds on the basis of differences in persistence conditions” Rea cited from personal correspondence with Bradley Strawser. Ibid., 386.
 Ibid., 86-87.
 Strawser, “Rea’s Revenge,” 387.
 Neander (1991), 168-173.