DRAFT of nearly 7 of the projected 8 sections. COMMENTS WELCOME!

[Go to individualism, below, for its role, and that of supervenience, in the is-ought problem. For Normativity and Proper Function, see Section IV, but the first three sections are necessary background.]

I.

John Mackie's argument from queerness against objective values has two parts. One is metaphysical, a question of what kind of entities or qualities or relations they could be, and how they could possibly be part of the fabric of the world. The other is epistemological, a question of we could ever know or be aware of such values:

When we ask the awkward question, how we can be aware of this authoritative prescriptivity . . . none of our ordinary accounts of sensory perception or introspection or the framing and confirming of explanatory hypotheses or inference or logical construction or conceptual analysis, or any combination of these, will provide a satisfactory answer; 'a special sort of intuition' is a lame answer, but it is the one to which the clear-headed objectivist is compelled to resort.⁽²⁾

This epistemological argument succeeds if the listed kinds of account all fail and there is no other way we could be aware of the would-be objective normativity. Many philosophers agree that the listed kinds all fail and there is no other way.

Nonetheless it is worth exploring whether there is another way. In particular, suppose we adopt the working hypothesis not only that the normativity is a phenomenon objectively in the natural world, but that it is best approached by means of the kind of theory-construction characteristic of empirical science. Granted, Mackie's list seems to cover this way, say under the heading "framing and confirming of explanatory hypotheses." But it is hard to tell, given how little he says about what counts as acceptable framing and confirming of explanatory hypotheses in science.⁽³⁾ In any event, such theory-construction deserves to be explored much more fully as a way to "be aware of" objective normativity, or at least important kinds of it, even though the sort of theory-construction characteristic of empirical science is alien to much philosophy.

Such method is especially alien when the aim is to resolve whether there is any normativity at all in the world independent of an agent's valuations, whether there is some non-mysterious relation between the would-be objective normativity and the world naturalistically described, and how we could ever know any of this. Indeed, most philosophers are likely to reject the very idea that this sort of theory-construction could possibly be worth exploring to any such end. Nonetheless, I hope to explain how naturalists could use the method to provide another way of knowing the would-be objective normativity, at least for important kinds of normativity. The aim is to undermine the widespread presumption that "the supposedly objective values [are] based in fact upon attitudes which the person has who takes himself to be recognizing and responding to those values," as Mackie puts it,⁽⁴⁾ or that "the only norms are in us and exist only from our point of view,"as John Searle puts it.⁽⁵⁾

To see better what the other way amounts to, it helps to begin with a relatively unproblematic example of the intended sort of theorizing. We do so in the next section (§II), using an example from physics. The point is to review the relevant general features of the method before considering how one might exploit them in order to improve the naturalist's chances of getting from is to ought. §III summarizes the well known obstacles to any would-be naturalistic account, acknowledging them as valid exacting constraints on any adequate naturalistic account of objective normativity. §IV introduces some essentials of the theory under construction, primarily the defeasible working hypothesis that a certain important kind of objective normativity is to be equated with a certain specific phenomenon objectively in the natural world. §V considers how the resulting naturalistic account, despite the severity of the constraints, could satisfy each in turn in the case of the objective normativity some see in biology -- the normative matter of what an adaptation is for, its proper function, something it is supposed or ought to do yet may fail to do.

§VI turns to less primitive kinds of normativity, including a kind involved when, under appropriate conditions, certain adaptations come to be for doing something entirely new under the sun. Another kind is involved when an organism conforms -- or fails to conform -- to a rule. If it can be shown how these kinds of objective normativity are not only possible but actual, that will undermine the pandemic irrealist presumption that the only norms are in us and exist only from our point of view. §VI considers certain more sophisticated kinds of normativity from biology. One is involved when an organism conforms -- or fails to conform -- to an "altruistic" rule, such as Tit-for-Tat (cooperate at the first encounter, thereafter do whatever the other did at the previous one), or, more significantly, such as what Philip Kitcher calls "golden-rule altruism."⁽⁶⁾

The final §VIII suggests how the intended general strategy of theory-construction might be extended beyond the biological to cases in which cultural conditions are what determine the normativity, including fully human moral normativity. The point is not that we can infer moral principles from biological (we cannot), not even that we can infer the objectivity of fully human moral normativity from the existence of important kinds of objective normativity in the biological world (we cannot). The point rather is that the intended general strategy might be made to apply directly to cases in which what determines fully human moral normativity has little to do with biology. The section ends with suggestions for how the strategy might be extended further, to epistemic normativity, and, normativity aside, to certain other philosophically fundamental matters. But the primary aim throughout the paper is to provide a naturalistic explanation of how there can be important kinds of normativity objectively in the world, contrary to those philosophers who think there can be no objective normativity at all, or at least none for which there could possibly be a naturalistically acceptable account.

II.

To see better what the intended sort of theorizing comes to, it helps to consider a concrete example. The point of the example, it needs to be said up front, is not that normativity is a natural kind, and not that it is to be defined by a synthetic identity statement. Rather, the point is to review the relevant general features of the intended sort of theory-construction, prior to considering how such construction might improve the naturalist's chances of getting from is to ought.

In this spirit, then, suppose we adopt the defeasible working hypothesis that temperature is an empirical phenomenon objectively in the natural world which is best approached by way of the sort of theory-construction characteristic of empirical science. This of course is substantially how physicists actually dealt with temperature. As regards temperature of liquids, for example, and oversimplifying for purposes of illustration, we may characterize the strategy adopted by kinetic theorists of temperature as provisionally assuming the identity -- or at least the (nomological) equivalence -- of my coffee's temperature and the mean kinetic energy of its molecules (in standard conditions). Kinetic theorists were already convinced that temperature is an empirical phenomenon objectively in the world (by contrast with a secondary quality or a sensation of temperature); what they were looking for was mainly a unification of temperature with Newtonian mechanics. To this end, they provisionally equated temperature with mean molecular kinetic energy (mmke), in standard conditions, on the ground that doing so might enable the resulting theory to unify the phenomena (mechanical and thermodynamic), and to do so with greater empirical adequacy than alternative accounts. This provisional assumption of identity or at least equivalence was then justified by its consequences -- explanatory and predictive -- under the resulting theory (the kinetic theory of temperature for liquids in standard conditions). So too today. Provided the theory enjoys the greater adequacy and power to unify, we are warranted in claiming to have discovered that the temperature of liquids equates with the mmke of their molecules (in standard conditions), in the sense of being either identical or at least (nomologically) equivalent to the mmke. If temperature thus construed does not conform to the vernacular concept -- or the philosophers' or for that matter earlier physicists' -- so much the worse for it.

Clearly, one consequence of this sort of approach is that conceptual analysis is out of place if not out of order, either for determining whether temperature is something objectively in the world, or for determining its relation to the world described in purely mechanical terms. So too is consulting our intuitions about our ordinary notion of temperature in order to identify necessary and sufficient conditions for its correct application to the target phenomenon in the world. Instead, we cast about for something mechanical in the world that explains temperature, causes it, or, perhaps best of all, is it or at least in some other sense equates with it. Furthermore, analysis aside, we are not bound by received notions of temperature, or by the intuitions of competent speakers of the vernacular in which the received notions are at home. In terms of Strawson's distinction between the descriptive and the revisionary,⁽⁷⁾ our approach to temperature is not be descriptive, in the sense of being bound by an existing concept, but decidedly revisionary, in the sense of putting the concept on trial should the phenomenon prove not to conform to it.

Notice further that given only the bare physical fact that the molecules in my coffee have high mean kinetic energy, one cannot deduce that it is piping hot or any other temperature. Just as, in parallel with what Mackie calls Hume's Law, one cannot deduce an ought from a bare is, one cannot deduce temperature from bare mmke. But to conclude from this, as many do in the case of normativity, that the theorist has not successfully connected would-be objective temperature to mmke would be strikingly beside the point. Kinetic theorists do not aim to deduce temperature from bare mmke (or from anything else non-thermodynamic). Rather, they provisionally propose -- hypothesize, posit, try on for size -- the identity or equivalence of temperature and mmke, then test the proposal by exploring its fruits and failures under the resulting theory. This general feature of such theory-construction -- that Hume's Law is beside the point -- would improve our chances of getting from is to ought.

Another general feature of this sort of revisionary theory-construction is that open-question arguments (OQA's) are rendered as beside the point as Hume's Law. According to Moore's OQA, we can easily imagine ourselves both recognizing that some natural condition C obtains (say that an act or policy x would conduce to the greatest happiness of the greatest number) and nonetheless asking meaningfully -- or, as he says, "with significance" -- whether x has normative property N (say whether x is good). Since this is an open question, it "shows clearly that we have two different notions before our minds"; therefore, N and C cannot be identical or equivalent, or their predicates synonymous.⁽⁸⁾ But so too can we imagine ourselves both recognizing that the coffee's molecules have a high mean kinetic energy and nonetheless asking "with significance" whether the coffee is hot. We do indeed have two notions before our minds here: the (vernacular) notion of temperature and the notion of mmke. Yet to conclude from this that temperature and mmke cannot be identical or equivalent would miss the point of the revisionary theorizing involved in the kinetic theorist's account (not that Moore would so conclude, though he would owe us some justification for the double standard -- that such theorizing is OK for temperature, but not for normativity -- a matter considered more fully in §V(ii)).

To be sure, Moore's OQA here against what Richard Hare calls semantic or analytic naturalism is not the only kind of open question argument.⁽⁹⁾ There is for example the new OQA advanced by Terry Horgan and Mark Timmons, which is endorsed by Hare and targets the "new-wave" naturalist's assertion of synthetic identity between normative and non-normative properties.⁽¹⁰⁾ But in §V(ii) we see that the intended sort of revisionary theory-construction is equally effective against new, improved OQA's. Even if a new, improved OQA succeeds against new-wave naturalism, it fails against the account developed here.

The aim of the intended sort of theory-construction is not to analyze or capture or conform to an existing notion, and not to find necessary and sufficient conditions for whatever property it might pick out if any. Rather, the aim is to find the best account of an empirical phenomenon in the world, one which we may see but darkly or whose existence we may even seriously doubt (as until recently was true of ball lightning, to mention just one example). In this sort of enterprise our received notions do not rule, contrary not only to what is presupposed by OQA's -- both Moore's and new, improved OQA's -- but much conventional philosophical method. A received notion is one which among other things has been around long enough for us to have acquired solid intuitions about its use. The new notion, almost by definition, has not been around long enough. Indeed, competence in its use is among the things the revisionary theory is meant to define. Consequently, even when the aim is not analysis -- and not to capture or conform to an existing notion or find necessary and sufficient conditions for whatever property it might pick out if any -- reliance on the intuitions of competent speakers is again beside the point (so too for intuitions about meaning and reference of terms on Twin Earth, as we see in §V(ii)). Nor is the posited identity or equivalence meant to hold in all possible worlds, only in a proper subset of the physically possible worlds (those in which the laws of physics obtain). The proper subset is made up of the physically possible worlds (ppw's) in which the relevant standard conditions obtain; what happens in other worlds is a don't-care.⁽¹¹⁾

Does any of this mean that the new notion need bear no resemblance to the old, that there are no received constraints whatever on usage of the same name in the new way? Not at all. Einstein's theory of relativity entails that mass is not conserved in all interactions, whereas for the philosophers of the day, and the classical physicists (and perhaps the folk), it was. His theory was therefore not only highly counter-intuitive but subject to the charge that he was changing the subject, that whatever he was talking about, it simply was not mass. But even though for Einstein mass is not conserved in all interactions, his theory entails that mass retains enough of the crucial features of mass classically construed to warrant the name. According to the theory, the unit of mass is still the kilogram, the mass of a body still determines both the action of gravity on the body and its resistance to changes in motion, mass still satisfies the principle that f = ma, and so on. These are received constraints on using the same name in the new way, where the constraints themselves are negotiable in light of evidence and the needs of theory. We retain our competence in applying the surviving classical principles, and, by studying the new theory, we add to this competence a competence in applying the principle that mass is not conserved in all interactions, as well as in applying other hitherto undreamt-of principles about mass that likewise derive from the fundamentals of the new theory.

In general, for any theory, however revisionary, there needs to be some sufficiently strong (family) resemblance between the new usage and the old, if the theory is to count as a theory of a phenomenon by the same name. What counts as a sufficiently strong resemblance varies from field to field and case to case, depending on the needs and aims of theory, on received usage of the relevant terms both in and out of the theory, on various empirical facts about the targeted phenomenon, and more. We should not expect some context-independent, across-the-board criterion for sufficient-strength-of-resemblance between the old and the new.⁽¹²⁾

This general strategy of theory construction, though routine in science and congenial to naturalism, is alien not only to much conventional philosophical method but even to many naturalists. Nonetheless, the strategy is so widespread and useful that it deserves a name. Call it inference to the best identity -- or, where the situation warrants only equivalence, inference to the best equivalence (though in either case the name could mislead, since strictly speaking there is no inference involved, let alone a truth-indicative inference, only a trial proposal to be tested by its fruits and failures under the containing theory). By whatever name, this revisionary strategy of inference to the best identity or equivalence -- the revisionary IBI/E strategy -- looks as though it could free the naturalistic theorist of objective normativity from the oppression of Hume's Law and OQA's, thus improving the naturalist's chances of getting from is to ought.

III.

The trouble, of course, is that even if Hume's Law and OQA's can thus be rendered beside the point -- which has still to be shown -- there remain a number of equally stubborn obstacles to any such would-be naturalistic account of objective normativity. The truly stubborn obstacles, in addition to Hume's Law and OQA, include the argument from queerness (AQ). It is true that Mackie formulates AQ as an argument against objective values, and talks mostly of moral values at that. Thus it may seem that Mackie does not mean AQ to apply to would-be objective normativity in general, but only to would-be objective moral normativity. Yet he also says,

The claim that values are not objective is meant to include not only moral goodness, which might be most naturally equated with moral value . . . . It also includes non-moral values . . . . [T]here would be at least some initial implausibility in a view that gave the one a different status from the other.⁽¹³⁾

Moreover, in his discussion of "functional words" like 'knife' and 'hygrometer,' where there is something the functional item A is for or is supposed or ought to do, Mackie argues as follows. We can indeed define a good A as simply "an A which is such as to be able to do that." Nonetheless, 'good' so defined, and however objective it might seem, "always imports some reference to something like interests or wants." Thus the seemingly objective good is based upon attitudes which the persons have who take themselves to be recognizing and responding to this goodness.⁽¹⁴⁾ Again the only norms are in us and exist only from our point of view. Mackie evidently does mean AQ to apply to would-be objective normativity across the board, and certainly to function words, some of which will figure prominently in the next section and beyond.

Even if Mackie does not mean AQ to apply so broadly, AQ can easily be generalized -- as at least implicitly some philosophers have done -- so as to target kinds of normativity other than those involved in moral values, certainly including the kinds of non-moral normativity at issue here (as we see in the case of Searle in §V(ii) and §V(iv)). According to the metaphysical part of AQ thus generalized, objective normativity -- even the kind supposedly involved in attributing function -- would have to be a queer sort of thing, because its relation to what is clearly objectively the case would be quite mysterious. No allegedly objective normative property N of an item x is deducible from x's non-normative or natural properties, as Hume taught us. Nor can N be reduced to natural properties of x, as Moore's OQA taught us. Furthermore, talk of supervenience of N on the facts is itself in need of naturalistically acceptable explanation.⁽¹⁵⁾ Since no other relation has been spelled out that works, the allegedly objective normativity must be a queer business indeed. "How much simpler and more comprehensible the situation would be if we could replace the [normative] quality with some sort of subjective response which could be causally related to the detection of the natural features on which the supposed quality is said to be consequential."⁽¹⁶⁾ What we call objective normativity is only our subjective valuation projected onto the value-neutral real world, a process Mackie calls objectification, likening it to what Hume calls the mind's "propensity to spread itself on external objects."⁽¹⁷⁾

Clearly, AQ contains a couple of sub-arguments that rest respectively on Hume's Law and OQA. By so much, then, would AQ be rendered beside the point by the IBI/E strategy, assuming the strategy can be made to apply to objective normativity. But there would remain the challenge of explaining just what are the objective affairs in world with which we could equate normativity by way of inference to the best identity or equivalence, and how such affairs could possibly determine, or subvene, the would-be objective fact that some given specific individual has this or that specific normative quality or property. Thus an adequate naturalistic account of objective normativity, in addition to deflecting Hume's Law and OQA, must show that the rest of AQ is unsound, by explaining (a) just what the needed non-mysterious relation is (or relations are) between the would-be objective normative properties and the relevant natural affairs, and (b) why the relation holds between given specific normative properties and the relevant specific natural affairs.⁽¹⁸⁾

The foregoing three conditions -- conform to Hume's Law, escape OQA and AQ -- amount to constraints on any adequate naturalistic theory of objective normativity. A crucial further constraint is that whether an individual x has a normative property N be determined solely by purely natural affairs. After all, one of naturalism's core theses is that all genuine properties whatever are thus determined. But another reason for imposing the constraint is that a property is clearly an objective one only if it is clearly determined by objective affairs, and the theory had better show that it is.

Another constraint on any naturalistic theory of objective normativity is related to the distinction -- or gap -- between what someone or something actually does or is disposed to do and what would be normatively better or worse. We see the distinction at work in our reaction to cynical remarks like "An impeachable offense is whatever the House says it is" (as then Representative Gerald Ford said with regard to Chief Justice Earl Warren). What the House ought to do cannot be equated with or reduced to what it does or is disposed to do (as recent experience has forcefully reminded us). As Kripke says, "whatever in fact I (am disposed to) do, there is a unique thing that I should do."⁽¹⁹⁾ We also see the distinction at work when we speak of a defective heart, or a bad heart, meaning a heart that does not (is not disposed to) do what it should do, or does not (is not disposed to) perform as designed (or a heart "which is not such as to be able to do that," as Mackie would say in this and other cases of function). Let a token heart be so diseased, damaged or deformed as to be totally indisposed to pump blood; nonetheless, we regard pumping blood as what it is for, what it is supposed to do.

This gap between how a thing x actually behaves or is disposed to behave and how it should behave is crucial for normativity, or at least for the kinds of normativity at issue here.⁽²⁰⁾ Any adequate theory of such normativity must preserve the gap. Specifically, the theory should entail that if N is some such normative property -- a property to the effect that x should (or should not) do thus-and-so -- then whether an individual x has N neither equates with nor is entailed by how x actually behaves in this respect, or is disposed to behave, just as what the House ought to do neither equates with nor is entailed by what it does or is disposed to do.

A further constraint is that the would-be objective normative properties play an appropriate explanatory role. Or rather this is a constraint imposed for the duration of the present paper. To require that any adequate naturalistic theory conform to it may strike some as excessive. Even according to naturalism, it is problematic whether to be is to be capable of figuring in a causal explanation of something else (or: "To be real is to have causal powers").⁽²¹⁾ Even for naturalists it may suffice to require only that a genuine property either play a causal explanatory role or be causally explainable or at least determined by other natural affairs.⁽²²⁾ Call the latter kind of property -- one which is causally explainable, or at least causally determined, but not causally explanatory -- a "spandrel property." Spandrel properties, though they themselves play no causal explanatory role, would be "ontologically grounded," in an important sense, and their relation to the causal order would neither be mysterious nor land us in substance dualism or a dualism of matter and spooky entelechies or pulls from the future (no more than spandrels in biology). On the other hand, many naturalists and others seem united in believing that the result would remain unacceptably epiphenomenal from the point of view of naturalism. So to placate them let us require, for the duration, that there be no spandrel properties -- specifically, that the objective normative properties targeted here do play a causal explanatory role. And for good measure let us require that they play a predictive role as well.

The final constraining condition -- unless others should turn up -- is related to the revisionary nature of the strategy involved in adopting the working hypothesis not only that normativity is a phenomenon -- or a family of phenomena -- objectively in the natural world, but that like other such phenomena it is best approached by way of the sort of theory-construction characteristic of empirical science. We see this condition at work perhaps most obviously in the natural sciences. Thus in physics we learn from Steven Weinberg that "Bohr...doubted the [Heisenberg-Pauli] theory would be the great new revolution in physics because it was not sufficiently 'crazy.' "⁽²³⁾ I call this the craziness condition, and I'm confident you'll come to agree that my theory satisfies it. The point of elevating it to the status of an adequacy constraint is to emphasize, beyond mistaking, the extent to which the strategy is revisionary and departs also in other fundamental ways from conventional philosophical method.

Summing up, an adequate naturalistic theory of the targeted kinds of normativity must (i) conform to Hume's Law; (ii) escape OQA; (iii) escape AQ by explaining (a) just what the needed non-mysterious relation is (or relations are) between the would-be objective normative properties and the relevant natural affairs, and (b) why the relation holds in particular between given specific normative properties and the relevant specific natural affairs; (iv) explain how the matter of whether x has a normative property N is determined solely by objective affairs; (v) preserve the gap between how x actually behaves or is disposed to behave and how it should behave; (vi) explain how N is not a mere spandrel property but plays an appropriate explanatory and predictive role, and a role to boot, and (vii) be sufficiently crazy. These constraints, or at least constraints (i)-(vi), constitute an all-too-familiar and hazardous minefield for the naturalist. How can the naturalist possibly negotiate the is-ought minefield?

IV.

Let us begin by adopting the working hypothesis that normativity is an empirical phenomenon objectively in the natural world best approached by way of revisionary IBI/E -- the revisionary strategy of inference to the best identity or equivalence. Such an approach conforms to Hume's Law and promises to block OQA. The problem is how to pull this off while simultaneously overcoming the equally stubborn obstacles represented by remaining constraints (iii)-(vii).

Prudence counsels starting with a relatively simple kind of normativity, leaving the hard cases for later. In particular, consider a kind of objective normativity many think they see in biology. According to Robert Brandon's rigorous account of adaptation, it makes sense to think of an adaptation as for something.⁽²⁴⁾ To think that an adaptation is for something appears to be a sort of normative stance -- call it the what-for stance. The what-for stance is closely related to what Daniel Dennett calls the design stance,⁽²⁵⁾ and it likewise invites various teleological what-for questions, including "What is the heart for?" and "Why is it for this rather than that?"

Of course a given trait might not be for anything, in which case a what-for question is out of order. But "Whenever we hypothesize that some trait is an adaptation, it makes sense to inquire about its function," what it's for. I would add only that an assertion of what something is for is indeed a normative assertion (as seems implicit in Brandon's account), or at least gives every appearance of being normative. After all, to say that the heart is for pumping blood is ordinarily to imply that even when a given heart cannot possibly pump blood, nonetheless pumping blood is what it is for, what it should do, what it is supposed to bring about or effect. We call a heart "defective" or "bad" when it cannot do what it is for; we distinguish between what it actually does or is disposed to do and what it should do. The notion of what x is for is a member of a family of interlocking notions about functional items, such as the notion of what x's function is, its purpose, what it is supposed to do, what it should or ought to do (in a suitably weak sense of "ought"). That is, x is for effecting E if and only if E is x's function, its purpose, what it should or ought or is supposed to do. At least this is how the terms will be used here; in context, to give an account of one is to give an account of them all.

Now with what objective affairs might we equate the apparently normative matter of being for E? Here too Brandon is instructive. It is true, as he says, that "Adaptations in nature seem to call for teleological explanations" -- explanations of what the adaptation is for, of its function. Black color in the peppered moth is an adaptation, and as such it calls for a teleological explanation, an explanation of what the color is for (it is for camouflage). Explanations of an adaptation -- adaptation explanations -- "are teleological in the sense that they are answers to what-for questions." But not only do they tell us specifically that an adaptation A is for E (the black color is for camouflage). They explain how it came about that A is for E: A is for E because E is the effect of A's past instances (often enough) in virtue of which A was selected for (in the relevant environments or niches).⁽²⁶⁾ Nonetheless, even though adaptation explanations are thereby teleological (in virtue of being answers to what-for questions), "they are also perfectly good causal/mechanical explanations." For according to an adaptation explanation, the "adaptation is the direct product of the process of evolution by natural selection," and the process is a causal/mechanical matter of "the ecological consequences of the adaptation, or its precursors, that explain its adaptive advantage over its alternatives."⁽²⁷⁾

This suggests we try equating the apparently normative matter of adaptation A's being for E with the causal/mechanical matter of E's being the effect of A's past instances in virtue of which A was selected for -- the effect in virtue of which A had an adaptive advantage over its alternatives.⁽²⁸⁾ Specifically, let us adopt as a working hypothesis the defeasible bridge principle that where A is an adaptation,

DFOR. A is directly for E if and only if E is the effect of A's past instances in virtue of which A was selected for.

The qualification "directly" is necessary because many adaptions, in addition to being directly for something in the sense defined by DFOR, can also be for something else in a sense not definable by DFOR -- a distinction which will make a big difference to our account of more sophisticated kinds of normativity in later sections.⁽²⁹⁾ For example, the imprinting mechanism in a newly-hatched chick is an adaptation directly for imprinting Junior on its mother. But the mechanism is also, and derivatively, for imprinting Junior on the here-now specific individual that is Junior's mother -- call her Henna. Since the here-now Henna is nowhere to be found in the relevant evolutionary history, imprinting on Henna cannot be the effect of the mechanism's past instances in virtue of which it was selected for. It follows by DFOR, as it should, that Junior's imprinting mechanism is not directly for imprinting on Henna. Nonetheless, the mechanism is objectively normatively for doing so in a derivative sense to be developed in §VI. As Millikan says, "animals that learn can acquire biological purposes that are peculiar to them as individuals, tailored to their own peculiar circumstances or peculiar history."⁽³⁰⁾

According to DFOR, black color in the peppered moth has the normative property of being for camouflage if and only if camouflage is the effect of the black color's past instances in virtue of which it was selected for. Since the latter is a descriptive, causal matter, DFOR expresses a kind of unification of the normative with the causal-descriptive. Provided DFOR is part of a theory that satisfies constraints (i)-(vii), the resulting theory would have considerable power to unify the normative and the causal-descriptive, in the case of this simple kind of normativity. We would be warranted in claiming to have discovered that this kind of normativity is equivalent to E's being the effect of A's past instances in virtue of which A was selected for. If the normativity thus construed does not fully conform to the vernacular concept(s) of what something is for -- or the philosophers' concept(s) or for that matter some biologists' -- so much the worse for it (cf. §V(iv)).

On the other hand, it is a peculiarity of DFOR that it defines a property in terms of something that is not itself a property of what has the property to be defined. What has the property to be defined -- the definiendum property of being for E -- is an adaptation A. What has the definiens property -- the property of being the effect of A's past instances in virtue of which A was selected for -- is not A but E. Even if a here-now (token of) A actually has the effect E, A is not the effect it has, and again the definiens property would not be a property of A (or its token) but of E. It follows further that DFOR entails neither identity nor equivalence between the definiendum property and the definiens property, since identity or equivalence of two properties requires that they be had by the same things. Any objection to DFOR which presupposes that DFOR asserts either property identity or property equivalence is thereby ruled out.

However, this definitional peculiarity conflicts with customary ideals of good definition, at least in certain quarters. Such ideals are reflected in paradigms like the definition of a triangle as a certain closed plane figure, or of a bachelor as an unmarried male. The property of being a certain closed plane figure is a property of what has the property to be defined; likewise for being an unmarried male. In metaethics the ideals these paradigms reflect are reflected as well in the forms taken by would-be definitions of normative properties and by objections to them. The struggles typically are over whether the definiendum normative property N of some individual x is really the same property as the proposed descriptive definiens property. The definition is said to succeed only if the definiendum and definiens predicates express or connote the same property.

DFOR violates this alleged principle of sound definition -- the principle that the definiendum and definiens predicates must express or connote the same property. According to the principle, if DFOR is to succeed then the properties connoted by the predicates "is for E" and "is the effect of A's past instances in virtue of which A was selected for" must be the same property. But two properties are the same only if they are had by the same things. Since A has the property of being for E but not the definiens property, these two properties cannot be the same. Therefore, DFOR fails.

This line neglects the fact that at least one legitimate variety of definition violates the alleged principle. Consider the definition

T. 'Snow is white' is true if and only if snow is white.

The property to be defined is truth, and it is a property of the sentence 'Snow is white'. But the defining phrase 'Snow is white' does not connote a property of the sentence, hence not a property of what has the property to be defined. Instead, the phrase is about snow and being white and snow's being white; it expresses a condition that does not amount to a property of what has the property to be defined.. Hence the left and right limbs of the definition do not connote the same property (indeed the defining phrase may not connote a property at all but a complete state of affairs). Rejecting a definition on the ground that it violates the alleged principle would require rejecting legitimate truth-definitions like T. It is the alleged principle that must go, insofar as it is meant to apply to all legitimate forms of definition.

DFOR's definitional peculiarity -- and T's -- is one of two defining features of what Colin McGinn calls "self-effacing properties," in a forthcoming book and in conversation.⁽³¹⁾

So far as I know, he is the first to focus on such properties. He does so in connection with truth, which he thinks is the only self-effacing property (and indeed defines truth just as the one and only self-effacing property). The other defining feature of self-effacing properties is that the definiens nowhere denotes the definiendum property. DFOR shares this feature as well. It follows that truth is not the only self-effacing property; there are at least as many self-effacing properties as DFOR defines, one each for the indefinitely many adaptations, each of which is for some specific E. (It is an interesting question why both truth and all these what-for or proper-function properties are self-affecting, not least because the question may have an interesting answer: truth is likewise a kind of proper-function property, though more complex.)

V.

Can DFOR be used in a naturalistic theory in such a way as to satisfy the adequacy constraints (i)-(vii)? Let us consider each in turn, bearing in mind that for now the target normativity is only the kind involved in an adaptation's A being for E.

(i)

According to the present account, the normative matter of A's being for E is not deduced from the bare description of how E is the effect of A's past instances in virtue of which A was selected for. It can be deduced only if the description is conjoined with the posited bridge principle DFOR (just as my coffee's temperature can be deduced from its mmke only in conjunction with the bridge principle that temperature equates with mmke, and the truth of 'Snow is white' from snow's being white only in conjunction with T). Thus the account conforms to Hume's Law, in accord with constraint (i). In this respect, as in others, the account follows Darwin, who did not deduce what seemed to many to be a real feature of the biological world -- design, adaptation, function, what A is for -- from a description of certain adaptive effects of past instances. Rather, consistent with revisionary IBI/E, he proposed in effect that this is what design, adaptation and the like amount to. He then justified the proposal by detailing its fruits under the resulting theory.

(ii)

What about OQA? The present theory does not propose necessary and sufficient conditions either for the correct application of some received concept or for whatever property it might pick out if any. The theory no more attempts this than Darwin attempted to capture or conform to what philosophers, theologians or ordinary folk -- or even fellow biologists -- might have meant when they spoke of adaptation or design. Virtually all of them thought of adaptation and design as entailing a knowing designer; their received notion of adaptation required a designer who creates the adaptation so as to suit or fit or be adapted to its environing conditions. Darwin emphatically thought not; the watchmaker is blind. Indeed we folk can easily imagine ourselves both recognizing that E is the effect in virtue of which A was selected for and nonetheless asking meaningfully -- or, as Moore says, "with significance" -- whether A was designed to do E; and this indeed "shows clearly that we have two different notions before our minds."

But from the fact that there are two different notions here, it does not follow that A's being for E, or A's being designed to bring about E, cannot be equated with E's being the adaptive effect of A's past instances. Nor does it follow that A's being-for-E, properly so-called, has been eliminated because being for E, by the new notion, does not entail a knowing designer. To suppose that either follows would be to misconstrue the point of the revisionary theorizing involved in the IBI/E strategy -- or perhaps to understand it, all right, but insist nonetheless that our received notions rule, which as noted earlier is one of the presuppositions involved not only in OQA but in much conventional philosophical method (see further §V(iv)). Darwin's strategy is likewise revisionary IBI/E. Objecting to his resulting notions of design and adaptation that they eliminate design and adaptation properly so called because they are not what the folk meant -- or what the philosophers or the biologists of the day meant -- would be as beside the point as objecting to Einstein's theory of mass that it fails to accord with what the folk or the philosophers or even the classical physicists meant by "mass," thereby eliminating mass properly so called, on the ground that for him mass is not conserved in all interactions, whereas for the philosophers and the Newtonians of the day it was. The presupposition that our received notions rule can be wonderfully useful against new ideas.

What about Brandon's account of what an adaptation is for, and Millikan's of an adaptation's proper function?⁽³²⁾ Neither account proposes necessary and sufficient conditions either for the correct application of some received concept or for whatever property it might pick out if any. Instead, each amounts to a theory of an empirical phenomenon in the world best approached by way of revisionary IBI/E. Granted, Millikan speaks rather of "theoretical definition," but the effect is much the same, as is the intent. Philosophers like John Searle and Alvin Plantinga misread the significance of her method, whatever their differences in other respects. True, neither of them construes her as intending to give an analysis of "function." Nonetheless, both assume that analysis of received notions is the method we ought to use for giving an account of function, both treat Millikan's account as possibly if unintentionally providing materials for such an analysis, and both discard it for failing to do so, on the ground that it suffers various counter-examples based on (intuitions about) received usage.⁽³³⁾ As Searle puts it, "if we take such definitions as capturing the essential features of our ordinary notion, there are counterexamples to the analyses"; there is no sense looking further in this direction for an account of function.⁽³⁴⁾ Both assume that any would-be account of function is useful only insofar as it might provide materials for an analysis of a received notion, which they believe is the method we should use in determining what function is. Both appear to have given little thought to whether analysis might be as absurdly beside the point here as it is in determining what temperature is, or mass -- or an adaptation, hence what it is for.⁽³⁵⁾

As for the present theory, it too is revisionary, along the lines of Darwin's. It proposes that what has seemed to many to be a real feature in the biological world -- the normativity involved in an adaptation's being for something, its function, what it is designed by natural selection to do -- be understood in terms of adaptive effects of past instances. Provided the proposal figures in a theory that is more adequate in relevant ways than competing theories, it amounts to a discovery. Subject to this proviso, which is addressed toward the end of §V(vi), what we will have discovered is that a phenomenon we saw but darkly -- indeed often doubted was objectively there at all, as many still do -- proves to be a matter of the ecological consequences of an adaptation's past instances which gave them an adaptive advantage over alternatives. The answer to the normative question "What is adaptation A for?" is, "A is for producing the consequences of past instances of A in virtue of which A was selected for," and the strategy used to get to this answer is revisionary IBI/E. Hume's Law and OQA are beside the point here (as are the sub-arguments of AQ that rely on them). Constraint (ii) is therefore satisfied, as well as (i).

Objection: "an element of to-be-pursuedness [is] crucial to the very meaning of normative notions," yet by equating the normative property of A's being for E with descriptive affairs, your account must thereby "squeeze out the normativity itself. . . . This is the general moral of Moore's 'open question argument'."⁽³⁶⁾ Well, yes, of course, some sort of to-be-pursuedness is essential to normativity, or at least to the kinds of normativity at issue here. Even though the normative matter of E's being what A is for does not to involve the same sort of to-be-pursuedness as does a moral normative property, it does involve some sort of to-be-pursuedness. After all, adaptations in nature call for teleological explanations (as seen in §IV in connection with Brandon's account). Teleological explanations are answers to questions about an adaptation's telos, the end it is supposed to aim at or pursue, what it is for. The heart's telos of pumping blood is what it is supposed to do or effect, and in this (primitive) sense is what it is supposed to pursue. Thus being for pumping blood involves a kind of to-be-pursuedness. The to-be-pursuedness is a matter of A's being supposed to pursue E. E has to-be-pursuedness in the sense that it is to be pursued by A.⁽³⁷⁾

Nevertheless, so far from "squeezing out" the to-be-pursuedness, the present account makes room for it, in other terms. It does so by hypothesizing that A's being supposed to pursue E, hence E's to-be-pursuedness, equates with E's being the effect of A's past instances in virtue of which A was selected for. That is, where A is an adaptation for E,

DTPE. A is supposed to pursue E (direct sense) if and only if E is the effect of past instances of A in virtue of which A was selected for.

If one thing is equated with another, then if the latter exists or obtains, so must the former; the former is hardly squeezed out. Since DTPE equates A's being supposed to pursue E with E's being the effect of past instances of A in virtue of which A was selected for, and since the latter exists whenever A is for E, DTPE no more squeezes out the to-be-pursuedness than Einstein's theory squeezes out mass or Darwin's squeezes out adaption. Rather, it makes room for it in other terms.

According to Horgan and Timmons, following Mackie, normative properties "that have to-be-pursuedness built into them would be metaphysically extremely queer," from the naturalist's point of view; "they would be quite unlike the naturalistic properties that natural science talks about."⁽³⁸⁾ But while normative properties are quite unlike the properties physics talks about, some of them are remarkably like certain properties biology talks about. What is queer to physics is not so queer to biology, which is up to its neck in talk about the naturalistic property of being an adaptation, and thereby about what the adaptation is for, as in effect Brandon, Sober, Millikan, Sterelny and Griffiths, and others have been telling us for some time. An adaptation A's property of being for something gives every appearance of having a kind of to-be-pursuedness built in. The to-be-pursuedness is a matter of A's being supposed to pursue E. E has to-be-pursuedness in the sense that it is to be pursued by A.

Horgan and Timmons might reply that even if an adaptation's property of being for E had to-be-pursuedness built in, "an . . . account . . . that identifies [such] normative [properties] with natural properties that do not have to-be-pursuedness built into them [would] thereby squeeze out the normativity itself." The reason is that the to-be-pursuedness "is crucial to the very meaning of normative notions," so that the natural notions, which do not have it built into them, will not mean the same as the normative.⁽³⁹⁾ Therefore, your account squeezes out the normativity by trying to equate the normativity with natural properties. Evolutionary biology only appears to be talking about normative properties objectively in the world.

It is true that "the natural properties do not have to-be-pursuedness built into them" in the sense either of deductively implying the to-be-pursuedness or of implying it under some received vernacular semantic rule. But it is also true in the same sense that the bare mechanical property of having high mmke does not have piping-hotness built into it. Again, given only the bare mechanical fact that the molecules in my coffee have high mean kinetic energy, one cannot deduce that it is piping hot or any other temperature. Nor is the inference from one to the other licensed by a received semantic rule based on the vernacular.⁽⁴⁰⁾ Yet it does not follow that equating temperature of liquids with their mmke would squeeze out the temperature itself. For temperature is built into mmke in the sense of there being a real-world relation between them expressed by the empirical bridge principle equating the two, a principle that has been tested by exploring its fruits and failures under the kinetic theory of temperature which it enables for liquids in standard conditions. So too is an adaptation's normative property of being supposed to pursue E built into E's being the effect of A's past instances in virtue of which A was selected for, in the sense of there being a real-world relation between them expressed by the bridge principle equating the two, namely DTPE.

Horgan and Timmons might reply that equating normative properties with natural properties would squeeze out the normativity itself, since the natural notions will not mean the same as the normative.⁽⁴¹⁾ They appear to allow something like revisionary IBI/E in the case of temperature and other natural-science properties, where preservation of meaning is not the aim, but not in the case of normativity.

This is a double standard, and it is question-begging if assumed without argument. But Horgan and Timmons appear to give an argument, at least by implication, which they call "a new open question argument," an argument against "new wave moral semantics." Even though this new argument is explicitly against new-wave moral semantics, it seems designed to work as well against naturalistic accounts of non-moral kinds of objective normativity (as we saw in §III in the case of Mackie's AQ, by which Horgan and Timmons are strongly influenced). In any event, we need to see whether their new open question argument could be made to work against the kinds of normativity at issue here. Specifically, in light of their discussion, what needs to be seen is whether their new argument could be made to work against the new-wave thesis that

CSN*. Each normative term t rigidly designates the natural property N that uniquely causally regulates the use of t by humans.⁽⁴²⁾

The target of their argument thus generalized would be any theory according to which (a) normative terms are natural kind terms; (b) natural kind terms are defined not by analytic identity statements but by synthetic identity statements -- presumably including theoretical definitions like DFOR and DTPE -- where the identity relation Horgan and Timmons have in mind is such that if x = y, then x = y in all logically possible worlds; and (c) natural kind terms rigidly designate whatever natural property causally regulates their use by humans.⁽⁴³⁾ On any such theory, it follows that each normative term rigidly designates the natural property that causally regulates its use by humans. Consequently, if Horgan and Timmons are right that normative terms fail to do so, contrary to CSN*, then any such theory is wrong.

Their argument thus generalized would work against the present theory only if the theory satisfied each of (a)-(c). It does satisfy part of (b); the normative term "is for E," whether or not it is a natural-kind term, receives a "theoretical definition" by the equivalence DFOR in the course of the revisionary IBI/E strategy. But the theory does not satisfy the part of (b) about identity, since as seen the theory neither asserts nor entails identity in the sense of identity in all logically possible worlds, which is what Horgan and Timmons have in mind. To repeat, the theory asserts only that DFOR is true in every physically possible world (ppw) in which certain background conditions obtain; so too for DTPE. Hence DFOR and DTPE are not even nomologically necessary, let alone logically or metaphysically necessary, since the relevant background conditions do not obtain as a matter of law.⁽⁴⁴⁾

Furthermore, insofar as Horgan and Timmons have property identity in mind, again the theory does not satisfy the part of (b) about identity, for the reasons given in §IV in connection with self-effacing properties.

As regards (a), everything depends on whether the term "the effect of A's past instances in virtue of which A was selected for," with which the target normative term is equated, is a natural-kind term in the sense presupposed by Horgan and Timmons (among others). Paradigms of natural kinds in their sense include "H₂O" and "mmke." Hence it is not at all clear that the phrase "the effect of A's past instances in virtue of which A was selected for" is a natural kind term. For one thing, unlike "H₂O" and "mmke," the phrase is not a term for an intrinsic property of the concrete physical stuff (say the molecules) of which something is composed. Nor, in the first place, is it the case that the phrase is a term for a property of (a token of) adaptation A, intrinsic or not, as opposed to a property of an effect of A's past instances; such is the way of self-effacing properties. In addition, and partly for this reason, it is not at all clear that the phrase, even if it does amount somehow to a natural kind term, would sustain the intuitions that persuade so many philosophers that natural kind terms are rigid designators, itself included. Hence it is not clear that the present theory satisfies (c) (that according to the theory, natural kind terms rigidly designate whatever natural property causally regulates their use by humans). Finally, the present theory does not use or presuppose the causal theory of reference presupposed by the sort of theory targeted by Horgan and Timmons, a representative version of which they find in the work of Richard Boyd.⁽⁴⁵⁾ Millikan's theory of meaning and reference, for example, which differs significantly from Boyd's, would be far better suited to the present theory of normativity.

However, rather than struggle with such tangled issues here, let us assume for the sake of argument that the present theory can be shoe-horned into satisfying all of (a)-(c). Now, under this assumption, would the normative term "is for E," as applied to an adaptation A, rigidly designate the natural property that causally regulates its use? Suppose so. The question then is whether the argument advanced at least implicitly by Horgan and Timmons against CSN* would work in the case of this sort of normative term.

According to their argument, the crux of the matter is what semantically competent speakers would say about the relevant Twin-Earth scenario. Having written that "Mastery of the semantic workings of 'water' is reflected in people's strong intuitions about Putnam's Twin Earth scenario," Horgan and Timmons go on to say:

Presumably, competent speakers have a comparable intuitive mastery of the semantic workings of 'good' and other fundamental moral terms. So if causal semantic naturalism is correct, then things should go the same way they go with 'water'. That is, if indeed the term 'good' rigidly designates the unique natural property (if there is one) that causally regulates the use of 'good' by humankind in general, then it should be possible [for the naturalist] to construct a suitable Twin Earth scenario.⁽⁴⁶⁾

So let us see whether it is possible to construct a suitable Twin Earth scenario, one which parallels the familiar scenario for water, H₂O and XYZ. In parallel with the term "water," suppose that the normative term "is for E," as applied to certain traits of organisms, is used by competent speakers of Twin English in the same readily observable circumstances as on Earth. Suppose further that the circumstances on Earth are those in which what Earthlings refer to as an organism trait that is for E is in fact an adaptation such that E is the effect of the trait's past instances in virtue of which it was selected for (here the parallel is to the circumstances in which what Earthlings refer to as water is H₂O). And suppose finally (in parallel with Twin water's being XYZ) that the circumstances on Twin Earth are those in which what Twin Earthlings refer to as an organism trait that is for E is not an adaptation, but was created to be for E by a super-scientist Artificer a mere 6,000 years ago, when He created Twin Earth and its strata, fossils and the rest so as to make it only seem that E is the effect of the trait's past instances in virtue of which it was selected for. For brevity, call this underlying Twin-Earth kind "CBA" (Created By Artificer). On Twin Earth, an organism trait said correctly to be for E is CBA.

In this scenario, "is for E" on Earth and "is for E" on Twin Earth have the same sense -- the sense of a trait's being for E -- but refer to different kinds (granting, as we are, Horgan and Timmons's terms of debate for the sake of argument). Were travelers from Earth to visit Twin Earth, they would translate the Twin term "is for E" as our orthographically identical English term.⁽⁴⁷⁾ Furthermore, I claim, the semantic intuitions of the relevant group of competent speakers would lead them to judge that the normative term "is for E" on Earth, as applied to an organism trait, rigidly designates what causally regulates the term's use by humans, namely the property of being an adaptation such that E is the effect of its past instances in virtue of which it was selected for. Likewise, they would judge that the normative term "is for E" on Twin Earth, as applied to organism traits, rigidly designates what causally regulates its use by Twin Earthlings, namely being CBA. Contrary to Horgan and Timmons, it looks as though a Twin-Earth scenario can be constructed in which the normative term "is for E," as applied to organism traits, satisfies CSN* (the principle that each normative term t rigidly designates the natural property N that uniquely causally regulates the use of t by humans).

Horgan and Timmons would likely object that the semantic intuitions of competent speakers would not go this way. After all, even long, long after publication of The Origin of Species, most competent speakers of English in the United States reject adaptation by natural selection; for them, an adaptation or adaptive trait for E was created by an Artificer to be for E. Thus their intuitions would likely lead them to judge that the normative terms "is for E" on Earth and "is for E" on Twin Earth "do not differ in meaning or reference, and hence that any apparent . . . differences between Earthlings and Twin Earthlings on this matter . . . involve belief or theory, not meaning."⁽⁴⁸⁾

Well, yes, competent speakers in this entrenched group, as in "humankind in general," might well have intuitions that would lead them to judge that the terms "is for E" on Earth and "is for E" on Twin Earth, as applied to an organism trait, do not differ in reference and are not rigid designators. But to make such speakers' intuitions the arbiter here is to let their received notion rule. It would be like running a Twin-Earth scenario against Einstein, by appealing to the semantic intuitions of the folk or the classical physicists who rejected mass not conserved in all interactions, in order to argue that the term "mass" does not rigidly designate the Einsteinian kind, hence that Einstein is wrong about mass (on the ground that any apparent differences between Earthlings and Twin Earthlings involve belief or theory, not meaning). For that matter, suppose that prior to the widespread acceptance of the relevant physical/chemical theories, some philosopher had run a Twin-Earth scenario about the terms for water and H₂O, or about the terms for temperature and mmke. It is doubtful that the intuitions of humankind in general -- including physicists -- would have supported the conclusion that 'water' rigidly designates H₂O and 'temperature' mmke, largely because any apparent differences in intuitions about reference could so easily be chalked up differences in belief or theory, not meaning.

It is all too easy to overlook this matter of which group of speakers is relevant, and at what time, when the relevant theory is not new but entrenched, as in the case today of water and temperature, where even vernacular semantic intuitions have long been schooled by the theory. Prior to the widespread acceptance of the H₂O-molecular theory of water, humankind in general did not have mastery of the semantic workings of 'water' in accord with the bridge principles of the theory. It is correspondingly doubtful that either they or the philosophers of the day would have had the firm intuitions today's philosophers have about the Twin-Earth scenario for water and H₂O, and in particular about whether 'water' rigidly designates H₂O. Nor is it even clear that their use of 'water' was causally regulated by H₂O, since their pre-theoretical beliefs about water may have caused them to use the term in ways we would not. So too for mass prior to Einstein -- and adaptation prior to Darwin.

In general, running a Twin-Earth scenario against a revisionary theory of something -- water, temperature, mass, adaptation for E -- allows the received notions of "humankind in general" to rule at the expense of new ideas, unless the relevant group of competent speakers in the scenario is specified to consist of those who are competent in using the new notions in accord with the principles of the new theory. Twin-Earth scenarios wedded to even new-style OQA's are as out of place here as Moore's OQA against analytic naturalism, and so is any sub-argument of AQ that relies on them. They are even more beside the point if, as argued above, the present theory -- unlike causal semantic naturalism, and contrary to what we've been assuming for the sake of argument -- does not satisfy the above conditions (a)-(c) necessary for theories of the sort targeted by Horgan and Timmons.

(iii)

What about the rest of AQ, namely the challenge to explain (a) what the needed non-mysterious relation is (or relations are) between the relevant natural affairs and objective norms, and (b) why the relation holds between given specific normative properties and the relevant natural affairs? In particular, what exactly is the relation between certain effects of past instances of adaptation A and the normative matter of A's being for E?

According to DFOR, A is for E if and only if E is the effect of A's past instances in virtue of which A was selected for. So there is at least a relation of equivalence between the two. What is the modality of this equivalence? As seen, the theory asserts only that DFOR is true in every physically possible world (ppw) in which certain background conditions obtain, namely those necessary for there to be adaptation via natural selection, and in particular for there to be selection for a trait in virtue of its past instances having had a certain effect or effects in the relevant environments. What happens in other possible worlds, including other ppw's, is a don't-care. Hence DFOR is not even nomologically necessary, let alone logically or metaphysically so, since the relevant background conditions do not obtain as a matter of law.

Nonetheless, the equivalence implies a strong determination claim. Given that the equivalence DFOR is true in the relevant ppw's -- the ppw's in which the relevant background conditions obtain -- it follows that in the same ppw's, the matter of whether A is for E is determined by whether E is the effect of A's past instances in virtue of which A was selected for. More precisely, A's being for E is determined by E's being the effect of A's past instances in virtue of which A was selected for, in the sense that:

Given any two ppw's in which the relevant background conditions are the same, and if, in both, E is the effect of A's past instances in virtue of which A was selected for, then, in both, A is for E.

In other words, in ppw's in which the relevant background conditions obtain, E's being the effect of A's past instances in virtue of which A was selected for determines the normative matter of A's being for E; the former determines the latter in a proper subset of the ppw's. This relation of "focused" determination or supervenience is simply the relation of global supervenience restricted to or focused on conditions relevant to a specific case.⁽⁴⁹⁾

We now have an explanation of why this relation of focused supervenience/determination holds between the given specific normative property of being for E and the relevant natural affairs. It holds because the former is equivalent to the latter in the relevant ppw's, by DFOR. To take a specific instance, consider again the peppered moth and its black coloration. Why does this relation of supervenience/determination hold between (1) the black color's normative property of being for camouflage and (2) camouflage's being the effect in virtue of which black color was selected for? Contrary to Horgan and Timmons, who argue that no such explanation can be given when the upper-level property is normative,⁽⁵⁰⁾ the explanation is simply that the two are equivalent in the relevant ppw's: in the relevant ppw's, black color in the peppered moth is normatively for camouflage if and only if camouflage is the effect in virtue of which black color was selected for. The supervenience holds in this specific case in virtue of this specific equivalence.

It follows further that there is not just one non-mysterious relation between the objective norm and the relevant natural features, but two: equivalence and focused supervenience/determination. Therefore clause (a) of constraint (iii) is satisfied, and also clause (b), according to which there must be a naturalistically acceptable explanation of why the non-mysterious relation -- focused supervenience/determination -- holds between given specific higher-level properties and specific natural affairs.

(iv)

Whether an adaptation A has the normative property of being for E is determined by whether E is the effect of A's past instances in virtue of which A was selected for. The latter is in turn a matter solely of objective causal/mechanical properties of, and relations among, the entities involved in the interactions of A's past instances with things in the relevant environments (or niches). Since A's being for E is also equivalent to this matter of the causal/mechanical adaptive effects of past instances, it follows that A's normative property of being for E is both determined by and equivalent to objective affairs, in accord with constraint (iv).

Thus what an adaptation is for, its function, is not observer-relative. Much of Searle's argument to the contrary is that any natural-selective account of function -- specifically Millikan's -- "cannot be right as far as our ordinary notion of function is concerned."⁽⁵¹⁾ The trouble with this sort of appeal to a received notion is that it is out of order when the strategy is revisionary IBI/E, which allows the received notion to be put on trial. Indeed Searle's argument against any notion of function that is not observer-relative amounts to a variant of OQA:

We can, arbitrarily, define the 'functions' of biological processes relative to the survival of organisms, but the idea that . . . functions are therefore intrinsic, is always subject to a variant of Moore's open-question argument: What is so functional about functions, so defined?"⁽⁵²⁾

Searle's point is that function so defined does not conform to our ordinary notion of function. This is clear from a variant of OQA that Searle uses in an argument which may seem unrelated:

Another, and perhaps decisive, clue that functions, unlike causes, are observer relative is that functional attributions, unlike causal attributions, are intensional-with-an-s. Substitution of coreferential terms in function contexts fails to guarantee preservation of truth value. . . . For example, it is trivially true that the function of oars is to row with, and rowing consists in exerting pressure on water relative to a fixed fulcrum; but it is not the case that the function of oars is to exert pressure on water relative to a fixed fulcrum.⁽⁵³⁾

For Searle, any theory according to which functional terms are non-intensional cannot be right; the received notion rules, and it is intensional. Furthermore, the present theory entails that functional terms are non-intensional. After all, substitution of coreferential terms in the right-hand limb of DFOR, which is in purely causal terms, preserves truth value (in the relevant ppw's); and since the right-hand limb defines "A is for E," substitution of coreferentials in "A is for E" also preserves truth value (again in the relevant ppw's). Hence it is a consequence of DFOR that function contexts or what-for contexts, as applied to an adaptation, are not intensional. Therefore the present theory cannot be right.

Well, note first that the intensionality of Searle's received notion of function derives from the fact that on this notion, having a function is a property only of artifacts, and an artifact's function, what it is for, is what the artificer/designer intends it to be. This requirement that there be an artificer/designer is what provokes Dennett's gibe that "According to Searle . . . [a]irplane wings are really for flying, but eagles' wings are not."⁽⁵⁴⁾

Note second that on Searle's received notion, if the designer intends that rowing-with is the function of oars, what they are for, then even if rowing = X, it does not follow that the designer intends X; hence it does not follow that X is the function of oars; substitution of coreferentials does not preserve truth value; function is observer-relative. As Moore would say, this "shows clearly that we have two different notions before our minds" (the received and the natural-selective).

But to conclude from any of this that the function of an adaptation A cannot be equated with certain effects of past instances in a natural-selective history, as OQA and Searle would have us conclude, would be like concluding that Darwin cannot be right about design because the received notion of design requires a knowing designer -- or for that matter, that Darwin cannot be right about adaptation because the received notion of adaptation requires a knowing designer who creates the adaptation so as to suit or fit or be adapted to some environing condition. Searle's account of what something is for, its function, would ultimately compel us to reject the notion of adaptation at the heart of evolutionary biology. Creationists and fundamentalists in general, by wielding Searle's or other variants of OQA, or by wielding any other strategy according to which the received notion rules, could make short work of adaptation by natural selection, as does Plantinga.⁽⁵⁵⁾

(v)

What about the gap between how something or someone x actually behaves or is disposed to behave and how x should behave? In particular, where x is an instance or token of an adaptation A, and N is the normative property of being for E, does the present theory entail that whether x has N is neither equated with nor entailed by whether x actually does or is disposed to do E?

According to DFOR, x has N if and only if E is the effect of A's past instances in virtue of which A was selected for. It follows that whether or not the here-now token x of A actually does or is actually disposed to do E, whether x has N is a matter solely of whether E was the effect of A's past instances in virtue of which A was selected for. Such past affairs, which determine whether x has N, cannot be affected by what x actually now does or is now disposed to do. Hence x's having the normative property N of being for E is neither equated with nor entailed by what x actually does or is disposed to do.

Or turn things this way. The gap between how something actually behaves or is disposed to behave, and how it should behave, is difficult to preserve, perhaps impossible, for those who adhere not only to naturalism but to individualism, where individualism is the view that an individual's properties are all determined by its own base properties or those of its parts. Such individualism pervades the mechanism present at the creation of modern science, according to which a thing's genuine properties are all determined by kinematic properties of its parts. Individualism survives in much philosophy, particularly in the reductivism according to which a thing's genuine properties are all reducible to its own base properties (in the sense of being identical or at least equivalent to some compound of them). When the base properties are allowed to include relational properties, the individualism is relational rather than non-relational or intrinsic.⁽⁵⁶⁾ Note too how individualist most if not all versions of OQA and AQ are in their focus on -- obsession with? -- the connection between an individual's normative properties and its own natural properties and relations.

Individualism persists as well in so-called "supervenience theories" of the moral, according to which, in Robert Audi's rendition,

first, no two things, whether acts or persons, can share all their natural properties and differ in their moral ones (if they have any); and second, any entity having moral properties possesses those properties in virtue of its natural properties (or certain of them).⁽⁵⁷⁾

The problem with "supervenience theories" is that they are strongly individualist, as indeed are nearly all the supervenience relations currently on offer. According to Jaegwon Kim, among others, it is "highly plausible to regard . . . Weak Supervenience . . . as minimally necessary for any claim of determination or dependency between sets of properties," hence for any claim of supervenience, where WS is the thesis that necessarily, for any individuals x and y, if x and y have the same base properties P, then they have the same higher-level properties N.⁽⁵⁸⁾ That is, there can be no difference in N-properties without a difference in the individual's own base properties -- a starkly individualist thesis that leaves no room for self-effacing properties, among others. The notable exception to such individualist supervenience relations is global supervenience, otherwise known as nonreductive determination, which is the supervenience relation used in the present paper.⁽⁵⁹⁾

Individualist naturalism about normative properties encounters at least two formidable difficulties. First, suppose for the sake of argument that a normative property of x is determined by some natural property of x. In particular, suppose that the heart's normative property of being for pumping blood is determined by some natural property of the heart. Which one? The most plausible candidate is some disposition to pump blood. But this means there would be no gap between the heart's disposition to pump blood and its being normatively for doing so, contrary to constraint (v). As Kripke says, "Is not the dispositional view simply an equation of performance and correctness?" Of course there may be ways around this difficulty, but they would come at a high price (including proliferation of ill-motivated ceteris-paribus clauses). Worse, "The fundamental problem . . . is [this]: whether [the] actual dispositions are 'right' or not, is there anything that mandates what they ought to be?"⁽⁶⁰⁾

The second difficulty is that individualism's starting assumption -- that a thing's genuine properties are all determined by its own base properties or those of its parts -- appears to be false even for many non-normative properties. In social science, biology, biochemistry, even physics itself, one finds properties that on the evidence are not determined by their bearers' own base properties, relational properties included, but only by these together with the base properties of other things at some distance in space and time.⁽⁶¹⁾ Naturalists, at least, would be well advised to avoid supervenience/determination relations that entail individualism, such as any relation that entails Weak Supervenience, as does Strong Supervenience. Unfortunately, as Peter Railton says, it appears to be "widely held that the supervenience of the moral upon the natural is (what has been called) strong supervenience."⁽⁶²⁾ If so, this would go a long way toward explaining why what Audi calls supervenience theories of the moral, individualist as they are, encounter so much trouble preserving the gap between actual behavior or dispositions to behave and what would be morally better or worse.

In any event, non-individualism offers a far smoother way of preserving the gap between actual behavior or dispositions to behave and what would be normatively better or worse (as well as a way of preserving (other) self-effacing properties). For according to the non-individualist, what determines whether an individual x has normative property N could well be natural conditions that include no natural properties or relations of x, or at least none either equated with or entailed by how x itself actually behaves or is disposed to behave. In such cases, what determines whether x has N neither equates with nor is entailed by what x itself does or is disposed to do. Hence constraint (v) is satisfied by the present theory, which is non-individualist.

(vi)

Does an adaptation's normative property of being for E play a causal explanatory role, and, for good measure, a predictive role as well? As seen in §IV, adaptation explanations fit the causal/mechanical or bottom-up model of explanation, where adaptation explanations, in addition to telling us that an adaptation A is for E, explain how it came about that A is for E.⁽⁶³⁾ Thus we may conclude not only that "adaptation explanations are thoroughly mechanistic; but . . . they serve to answer teleological questions," so that "the sense in which what-for questions and their answers are teleological" is explained in purely mechanistic terms.

It follows that an adaptation A's normative property N of being for E plays a causal explanatory role, in accord with constraint (vi). For we can explain A's existence by reference to A's being for E: A exists because A is for E. What "grounds" this explanation, or what perhaps makes it an explanation in the first place, is the longer one it sums up: A's normative property N of being for E explains why A exists, indeed causally explains why A exists, in that for A to be for E is for A's past instances to have had a fitness advantage in virtue of having effect E, which fitness advantage explains why A exists. For instance, why does the black color of the peppered moth in soot-polluted environments exist? Because it is for camouflage. That is, the black color exists because the color's past instances had the adaptive effect of camouflage (the first instances being explained by mutation). More fully still, and generalizing, "adaptation A's existence is explained by a complete adaptation explanation that includes not only the ecological account of the function of the adaptation, but also the other four [kinds of information needed for a complete adaptation explanation]." The explanation is both teleological, in that it tells us what A is for, and causal/mechanical, in that tells us how it came about that A both exists and is for E in virtue of the effects of past instances.

There is a further way N can do causal explanatory work. Biologists are often interested in what causal/mechanical effects x has on other things, assuming x is functioning as designed by natural selection and in design conditions. To attribute to the heart the normative property of being for pumping blood, then, is to give a causal explanation of how things the heart impacts are causally affected by its pumping, assuming the heart is functioning "normally," namely as designed and in design conditions (though of course the explanation will usually require as well some further knowledge about the heart's actual causal powers and mechanisms, assuming it is functioning as designed, and about those of the things it affects).

What of a predictive role for A's normative property of being for E? It may seem there can be none. After all, precisely because being for E is a normative property, there must be a gap between what an instance or token x of A is for and what x actually does or is disposed to do, hence a gap between x's being for E and any prediction about what x actually does or is disposed to do, as well as about how other things are affected by what x actually does. But it hardly follows that there is no predictive role at all for x's normative property of being for E. Often we want to know how x will behave assuming it is functioning as designed and in design conditions; this is as true of adaptations as of artifacts, as true of hearts as of fuel pumps. To attribute to a token heart x the normative property of being for pumping blood, then, is among other things to predict how x will behave, and how things x impacts will be affected, assuming x is behaving as designed and in design conditions. In general, if you know what A is for -- if you know its function or purpose -- you can predict how A will behave assuming it is operating as designed and in design conditions. Thus A's normative property N of being for E plays both a causal explanatory role and predictive one, in accord with constraint (vi).

In these two crucial respects, at least, the present theory enjoys significant empirical adequacy, more so at any rate than accounts according to which there are objective normative properties yet they do no causal explanatory or predictive work. But what about accounts according to which there are no objective normative properties in the first place, hence none to do the work? Such accounts often attribute a causal explanatory and predictive role, but only to a speaker's use of the normative predicates, as part of an explanation or prediction of the speaker's psychological attitudes, states or behavior, or their consequences. Such an account is correspondingly unable to do justice to the objective explanatory and predictive power of such predicates in evolutionary biology, a power which results from construing an adaptation as normatively for this or that, and which is a matter of how things are in the world, not a matter of a speaker's attitudes being projected onto it.⁽⁶⁴⁾ Furthermore, if one is looking for a theory that would unify objective normativity and the causal-descriptive, then the present theory enjoys the greater power to do so, the others having given up the very attempt.

But what about other theories according to which there are normative properties objectively in the world which do play both a causal explanatory role and a predictive one? Well, even if they should prove tied with the present theory on this score, or perhaps even superior, they must first satisfy constraints (i)-(v) if they are to be so much as in the running. Since most if not all such theories are individualist and reductive, in the sense of equating a normative property of x with some (compound) base property or disposition of x, they seem bound not to satisfy constraint (v) of preserving the gap between how x actually behaves or is disposed to behave and how it should behave. Hence they would not be in the running. And of course many of them have serious troubles with other constraints as well, especially constraints (ii) (escape OQA) and (iii) (escape AQ).

By contrast, it looks as though the present theory satisfies all seven constraining conditions (satisfaction of the craziness condition being so obvious as to require no argument). Naturalists appear to have a way to negotiate the is-ought minefield, contrary to those who insist there is no kind of normativity objectively in the world, or at least none for which there could possibly be a naturalistically acceptable account. This amounts to a vindication as well of the specific kind of objective normativity presupposed both by Millikan's theory of proper function and at least implicitly by the theories of what an adaptation is for, or its function, constructed by Darwin, Brandon, Sober, Sterelny and Griffiths, and others.⁽⁶⁵⁾ Call this kind of objective normativity primitive or minimal normativity. Where an adaptation A is for E, A's property of being for E is a primitive normative property objectively in the world.

VI.

Among the questions screaming to get out is this: What could such primitive normativity possibly have to do with moral normativity? In itself, not much. But the general strategy which has brought us this far can be extended, in at least three ways. One way, which is explored briefly in the final section (§VIII), is to apply the same strategy, or what is formally the same strategy, not to the what-forness of adaptations or anything else biological, but directly to kinds of normativity in which non-biological, broadly cultural factors are among those that determine the normativity. For example, suppose we were to construct a theory according to which moral normativity is nonreductively determined not by biological affairs, but say by what would contribute to the greatest happiness of the greatest number; or by what would foster the mutually supporting,"homeostatically clustered" physical and psychological or social human needs together with the mechanisms that unify them;⁽⁶⁶⁾ or whatever. The strategy would then be to show that the resulting theory satisfies constraints (i)-(vii), by (a) using revisionary IBI/E to equate the targeted moral normativity with the hypothesized nonreductively determining objective affairs in such a way as to avoid individualism, thus not only preserving the gap between how an individual behaves or is disposed to behave and what would be morally better or worse, but undermining objections based on Hume's Law, OQA and AQ; and (b) explaining how the targeted moral properties play an appropriate explanatory and predictive role. Provided the resulting theory is successful at each step of the way -- obviously a big proviso -- the result would be a naturalistic theory of objective moral normativity which satisfies the constraints.

A second way of extending the strategy, which is less ambitious, appears in the latter part of the present section. The idea is to apply the same strategy to cases in which one of the things an adaption is for is getting the individual organism to conform to a rule. As noted in §IV, the imprinting mechanism in a newly-hatched chick is an adaptation for imprinting Junior on Mom. But as we shall see, it is also, and thereby, for getting Junior to conform to the rule, "Imprint on Mom." It will further turn out that the true proposition that Junior ought to imprint on Mom is, in an important sense, "intrinsically action-guiding," in that the proposition entails a rule or standard or norm, namely "Imprint on Mom."⁽⁶⁷⁾ This holds even though the rule is not expressed or expressible by the chick, and not otherwise represented by the chick to itself; nor do we find the rule by peering into the chick's head. Nonetheless, it is one of the chick's biological purposes to conform to the rule, meaning only that the chick contains mechanisms which are for producing conformity to the rule, and which cause the chick to conform to it when they are operating as designed and in design conditions. Then there are the organisms that contain mechanisms for getting them to conform to an "altruistic" rule, such as Tit-For-Tat or what Kitcher calls golden-rule altruism.⁽⁶⁸⁾ The normativity involved in being supposed to conform to an altruistic rule is explored in the next section (§VII).

A third way of extending the strategy occupies this and the next few paragraphs. Among the things an adaptation can be for, in certain conditions, is effecting something quite new under the sun. This results in a kind of normativity that is significantly less primitive than what DFOR defines. Again suppose that Junior-the-chick's mom is Henna. Because Henna nowhere occurs in the evolutionary history, it follows by DFOR, as it should, that Junior's imprinting mechanism is not directly for imprinting Junior on Henna. But in a further sense the mechanism is indeed for imprinting Junior on Henna, given that Henna is the specific item that the mechanism is here-now supposed to imprint Junior on, by virtue of Henna's being the chick's mother.

To get at this further sense, we need first to define what it means for an adaption A to be relationally for E:

RFOR. A is relationally for E if and only if A is directly for E and E is the doing or producing something x in relation r to something else y.⁽⁶⁹⁾

For example, the imprinting mechanism in the chick is directly for getting it into the relation "imprinted on" to its mother (since this is the effect in virtue of which the mechanism was selected for). Hence by RFOR, the mechanism is relationally for imprinting the chick on its mother. Note, however, that by RFOR the mechanism is not relationally for imprinting Junior on Henna; if it were, it would be directly for imprinting Junior on Henna, which by DFOR would mean that imprinting on Henna is the effect of the mechanism's past instances in virtue of which it was selected for, which would be absurd. We need a further notion of how A can be for something.

Consider a situation S in which (1) A is relationally for E, where E is the doing or producing something x in relation r to something else y; and (2) the environment (external or internal) supplies the specific item y such that, in S, A is for doing or producing x in relation r to y. Call this specific item y*. In such situations, A is for producing x in relation r to y* in virtue of y*'s being the y such that A is for producing x in relation r to y (just as Henna is the y such that Junior's imprinting mechanism is for imprinting on y, in the situation in which Henna is Junior's mom). Thus

SFOR. A is situationally for E* in S if and only if: (a) A is relationally for E, where E is the doing or producing something x in relation r to something else y, and (b) E* is the doing or producing x in relation r to the specific y in S -- call it y* -- such that A is for producing x in relation r to y.

SFOR is a mouthful, but the basic idea is simple enough. Since Junior's imprinting mechanism is relationally for imprinting him on his mother y (in line with (a)), and the situation we are imagining is one in which Henna is his mother, hence is the specific y with respect to which the mechanism is for effecting the imprinted-on-mother relation (in line with (b)), Junior's imprinting mechanism is situationally for imprinting him on Henna.

Another illustration may be helpful -- a variant of one of Millikan's. There are mechanisms in the rat that are an adaptation relationally for producing an aversion to eating what smells like the stuff it had when it got sick (or its fellows had when they got sick or died). Now suppose the situation is one in which the stuff King Rat had when King Rat got sick is the children's Silly Putty -- a substance nowhere encountered in the evolutionary history of rats, indeed instantaneously new by comparison to that history. In King Rat's situation, Silly Putty is the specific stuff that his mechanisms are for producing an aversion to eating, in virtue of Silly Putty's being the specific stuff in S with respect to which they are for effecting the aversion. Thus by SFOR, King Rat's mechanisms are situationally for preventing him from eating Silly Putty. In this way an animal's mechanisms can come to have an objective normative property that is not only novel but instantaneously so; the animal can acquire a biological purpose -- a biological should or ought -- that is peculiar to it as an individual, tailored to its own peculiar circumstances or peculiar experience.⁽⁷⁰⁾ In an important sense, King Rat should or ought not eat Silly Putty, meaning simply that he contains mechanisms that are for getting him not to eat the stuff, or equivalently, as we shall see, for getting him to conform to the rule "Don't eat Silly Putty."

Even though this situational normativity is significantly less primitive than that of an adaptation's being directly for something, one might well wonder whether it was worth slogging through the complications of RFOR and SFOR. But consider organisms that contain sophisticated cognitive mechanisms for solving complex practical problems by producing rules and behavior in conformity to them which results in an optimal solution. One such organism, we know, is endowed further with language and considerable ability to reason. This organism has frequently found itself in situations nowhere encountered in its evolutionary history, situations in which its mechanisms are normatively for solving complex practical problems by producing rules new under the sun and behavior in conformity to them -- some of the rules looking remarkably like moral rules -- so that in an important sense it is one of the organisms's biological purposes to conform to the rules, one of its shoulds or oughts, not unlike the way it is one of Junior's biological purposes or shoulds or oughts to conform to "Imprint on Mom," or one of the rat's to conform to "Don't eat Silly Putty," or, as we shall see in §VII, one of a certain savannah-dwelling hominid's to conform to Kitcher's golden-rule altruism rule.

Clearly this is to get ahead of ourselves. Preliminaries are in order. One of them is to determine whether the normativity involved in A's being either relationally or situationally for something satisfies the constraints (i)-(vii) on any would-be naturalistic theory of objective normativity. Fortunately this is not so difficult, thanks to the work of previous sections. Like DFOR, RFOR and SFOR are introduced as part of the revisionary strategy of inference to the best identity or equivalence. Consequently, again like DFOR and in light of essentially the same arguments as in §V, RFOR and SFOR satisfy constraints (i)-(v) (conform to Hume's Law, escape open-question arguments, escape arguments from queerness, explain how the matter of whether x has a specific normative property N is determined solely by the relevant objective affairs, and preserve the gap between how something actually behaves or is disposed to behave and how it should behave).

However, constraint (vi) -- that N play an appropriate explanatory and predictive role -- is more complicated in the case of SFOR. Here is how it goes. Why does Junior imprint on Henna? Because Henna is what Jr. should imprint on (in the relevant normative sense of this biological should, the one defined by SFOR). More fully, Junior contains an adaptation A which exists because it is for imprinting Juniors on their Moms, and Junior's Mom is Henna. More fully still, Junior contains an adaptation A which exists because of the effect of past instances of A in imprinting past juniors on their Moms; Junior is in a situation in which Mom is Henna; and A is operating as designed and in design conditions. It follows that the normative property N* of being situationally for imprinting Junior on Henna plays a causal explanatory role (cf. §V(vi)). So too, a fortiori, for its being relationally for imprinting him on his mother in the sense defined by RFOR.

What about a predictive role for N*? By regarding Junior's imprinting mechanism as normatively for imprinting him on Henna, we can predict what Junior will do in the situation. We can predict that Junior will imprint on Henna -- assuming his imprinting mechanism is operating as designed and in design conditions, and that his Mom is Henna. So too, a fortiori, for its being relationally for imprinting him on Mom in the sense defined by RFOR. Thus it looks as though the less primitive normativity involved in an adaptation's being either relationally or situationally for something satisfies constraints (i)-(vii) (satisfaction of the craziness condition again being so obvious as to require no argument).

Another necessary preliminary is to make sense of conformity to a rule. Start with an example from Millikan.⁽⁷¹⁾ Male hoverflies will hover for hours in one spot, darting seemingly at random after sundry objects (passing midges, wind-blown bits of this and that, tossed BB's, distant birds, male hoverflies, female hoverflies). What are they doing? That is, assuming this behavior is for something, what is it for? What is the purpose of the underlying mechanisms? It turns out that the males are "keeping their flight muscles warm and primed so that they are ready to dart instantly after any passing female they sight. This chasing behavior is on such a hair-trigger that all manner of inappropriate targets elicit pursuit."⁽⁷²⁾ Calculations show that if the male is to intercept the target, he must turn at an angle equal to the angle between the center of his retina and the image of the target, minus 1/10 the image's angular velocity, ± 180. According to Collett and Land, the hoverfly does indeed conform to this rule. Because it is a rule about how he should respond to a proximal stimulus (the moving spot on his retina), Millikan calls this "the proximal hoverfly rule."

Conformity to the proximal hoverfly rule is a means of conforming to a less proximal, more distal, rule: "If you see a female, catch it." Such rules represent biological purposes the hoverfly has. They are unexpressed purposes, indeed purposes of which the hoverfly is not and could not be aware, given his primitive nervous system. Nor could they be found by peering into his head. Instead, to say that he has such a purpose is to say that he "has within him a genetically determined mechanism of a kind that historically proliferated in part because it was responsible for producing conformity to the ... rule, hence for getting male and female hoverflies together"⁽⁷³⁾ In other words, the mechanism is directly for getting him to conform to the rule, in the sense of DFOR. In order to find out the purpose or function of some behavior or device -- and to find the rule or rules if any to which it is supposed to conform -- we must often take a long look back in time to see what role it played in such evolutionary success as the creature may have enjoyed. In doing so we discover "how purposes inform the rule-following behavior of the hoverfly, how norms, standards, or ideals apply to his behavior, hence how the hoverfly comes to display competences or abilities to conform to rules rather than mere dispositions to coincide with them."⁽⁷⁴⁾

In short, and as Millikan argues, we have an answer to Kripke's challenge, noted earlier: "The fundamental problem . . . is [this]: whether [the] actual dispositions are 'right' or not, is there anything that mandates what they ought to be?" What mandates -- that is, what determines -- that the male hoverfly ought to conform to the proximal rule, as opposed to some quoverfly rule or Fodorian or disjunctive rule, are natural affairs in the selective history in virtue of which the relevant mechanism is for getting him to conform to the rule.⁽⁷⁵⁾ In general, where A is an adaptation for getting an organism to behave a certain way B -- to turn at such-and-such angle, to imprint on Mom, to imprint on Henna, not to eat Silly Putty -- then A is also for getting the organism to conform to the corresponding rule: "Turn at such-and-such angle," "Imprint on Mom," and so on. Formally, this comes to

RULE. Where adaptation A is for getting organism x to B, A is also for getting x to conform to rule "B".

Note that RULE applies whether A is directly, relationally or situationally for getting the organism to B, as the foregoing examples indicate, so that A can be directly, relationally or situationally for getting the organism to conform to rule "B". It should also be clear that, like the kinds of normativity defined by DFOR, RFOR and SFOR, the kind of normativity defined by RULE -- the normativity involved in A's being for getting x to conform to a rule -- satisfies the constraints (i)-(vii) on any would-be naturalistic theory of objective normativity, and for the same reasons. Note also that insofar as the proposition that Junior ought to imprint on Mom means simply that Junior contains mechanisms which are an adaptation for getting him to imprint Mom, it follows by RULE that the proposition entails a rule, standard or norm, namely "Imprint on Mom." Thus in an important sense the proposition is "intrinsically action-guiding," by virtue of entailing a norm.

VII.

With these preliminaries in hand, let us consider whether there is anything like a moral rule to which some organism is objectively supposed to conform, by virtue of what certain of its adaptations are for. This requires deciding first what sort of rule is to count as sufficiently like a moral rule -- whether the rule is a fully human moral rule, a primitive moral rule, or some kind of moral rule between. What are the conditions a rule must satisfy, if it is to count as at least a primitive moral rule? One such condition is this. According to most philosophers and ordinary folk alike, the main point of morality, if not the whole point, is to overrule acting in one's own self-interest when so acting would harm others. A rule which does not overrule acting in one's own self-interest when so acting would harm others is not even a primitive moral rule.⁽⁷⁶⁾ A second condition is that an individual ought to conform to the rule if and only if any relevantly similar individual in relevantly similar circumstances ought to do so as well. This amounts to an abstract, weak universalizability principle (sometimes called supervenience): an individual ought to do so-and-so if and only if any relevantly similar individual in relevantly similar circumstances ought to do so as well.

These two conditions seem necessary and sufficient for a rule to count as at least a primitive moral rule. The idea is that anything that satisfies these conditions bears a sufficiently strong (family) resemblance to what is customarily called a moral rule to deserve the name. Now is there anything sufficiently like a moral rule in this sense, to which some organism is objectively supposed to conform by virtue of what certain of its adaptations are for?

Consider the guppy.⁽⁷⁷⁾ When a large fish nears a school of guppies in the wild, some of the guppies will approach the stranger to see whether it is a predator. If it is, then even though the scouts can warn the school, they will likely be eaten unless enough of them stick together long enough to confuse the predator by scattering. As they approach the stranger, they generally keep up the first time they swim out, but subsequently hold back if accompanied by (enough) others who held back on the previous occasion.. When accompanied by individuals that kept up on the previous occasion, they will themselves keep up on the present occasion. That is, they keep up or hold back according to what others did on the previous occasion. If new individuals swim out, the other scouts keep up the first time, but thereafter do what those individuals did the time before.

Clearly the guppies are cognitively advanced enough to identify individual guppies and to remember which ones did what on the previous occasion. In addition, the behavior of the scout guppies (often enough) can be described as Tit for Tat (TFT): on the first occasion, keep up, thereafter do whatever the other did on the previous occasion. Evidently there are guppy mechanisms responsible for, among other things, identifying other guppies, remembering what they did on the previous occasion, and (often enough) getting the guppy to keep up on the first occasion while thereafter doing what others did on the previous occasion. Furthermore, it turns out mathematically that such reciprocal cooperative behavior significantly increases each individual scout guppy's chances of survival and reproduction by contrast to always keeping up, never keeping up, keeping up every other time, and so on. The extra risk to the individual of keeping up on any given occasion is outweighed by the benefit to it of TFT behavior in the longer run. And of course the whole school benefits as well.

Presumably these mechanisms are an adaptation. If so, among the things they are for is producing the Tit-for-Tat behavior. More precisely, they are relationally for getting the guppy to keep up on the first occasion and thereafter do what the other did on the previous occasion. According to RULE, this is equivalent to saying that the mechanisms are for getting guppies to conform to the TFT rule, "Keep up on the first occasion, thereafter do what others did on the previous occasion." Clearly the TFT rule is unexpressed and unexpressible by the guppy, and not otherwise represented by the guppy to itself; nor do we find the rule by peering into the guppy's head. Nonetheless, it is one of the guppy's biological purposes to conform to the TFT rule, meaning simply that the guppy contains mechanisms that are for producing conformity to the rule. In this primitive sense, the guppy is supposed to or ought to conform to TFT. Further, the proposition that the guppy ought to keep up on the first occasion, thereafter do what others did on the previous occasion conform, is intrinsically action-guiding, by virtue of entailing a rule or norm, namely TFT.

Tit for Tat is a rule that bears on encounters among individual organisms, in this case among conspecifics. In this and many other cases, TFT is the rule of encounter conformity to which best promotes the organism's fitness; in others, other rules do better.⁽⁷⁸⁾ More importantly for present purposes, the point of TFT is to overrule the individual scout's acting in its own self interest, as when it acts to minimize its chances of being eaten by holding back when the others do not. Furthermore, by so holding back the scout would harm those who do not, by increasing their chances of being eaten. So in the guppy case, TFT is to overrule following self-interest when following self-interest would harm others. Thus TFT satisfies the first condition for being at least a primitive moral rule.

What about the universalizablility condition? Is it true that an individual scout should or ought to conform to TFT if and only if any relevantly similar individual in relevantly similar circumstances ought do so? The normative "ought" involved here is defined by appropriate application of RFOR to the guppy case (or SFOR when the guppy is accompanied by a specific guppy George who did not keep up the last time, so that the mechanisms are specifically for getting the guppy to hold back with George). When this definition is unpacked, we get

GTFT. The guppy ought to conform to Tit for Tat if and only if it contains mechanisms that were selected for in virtue of the effect of their past instances in getting guppies to keep up on the first occasion and thereafter do what others did on the previous occasion.

As we saw §V(iii), an equivalence like this implies a rather strong determination or supervenience claim. Specifically, GTFT implies

GSUPV. That the guppy ought to conform to Tit for Tat is determined by or supervenes on the fact that the guppy contains mechanisms selected for in virtue of the effect of their past instances in getting guppies to conform to Tit for Tat,

where GSUPV means that given any two physically possible worlds in which the relevant background conditions are the same, and if, in both, getting guppies to conform to TFT is the effect of the mechanisms' past instances in virtue of which the mechanisms were selected for, then in both the guppy ought to conform to TFT. But what determines or subvenes the possession of a normative property is also what fixes the relevance referred to in the weak universalizability principle. The relevantly similarities of individuals and circumstances are those that determine or subvene the normative property of being obligated to do so-and-so; no difference in the former without a difference in the latter. Thus relevant similarities in the present case are those set out by GSUPV. It follows that an individual guppy ought to conform to TFT if and only if any relevantly similar individual in relevantly similar circumstances ought to conform to TFT, where the relevantly similar individuals and circumstances are those involved in there being guppy mechanisms that were selected for in virtue of the effect of their past instances in getting guppies to conform to Tit for Tat, including being in a physically possible world in which the relevant background conditions obtain. (When the guppy is accompanied by a specific guppy George who did not keep up the last time, so that the mechanisms are specifically for getting the guppy to hold back with George, then George and his behavior are added to the relevant circumstances.) Thus the universalizability condition is satisfied, along with the first condition, so that in this sort of case TFT is at least a primitive moral rule. It is worth emphasizing that since the relevant circumstances here are not shared by us humans, we have no such primitive moral obligation.

Can we go further? Here is a tempting possibility. Kitcher advances an explanation of how possibly human altruism could have evolved under natural selection.⁽⁷⁹⁾ Suppose he is right that for our savannah-dwelling hominid ancestors, "selection will favor 'golden-rule' altruism of a discriminating kind (treat the other as oneself so long as one has no basis for thinking that the other will not do the same)."⁽⁸⁰⁾ Then by the present account, the hominids objectively have a primitive moral obligation to conform to this rule, and so do relevantly like organisms in the relevant kinds of encounters (which clearly does not include us humans). To say that conformity to this rule is primitively obligatory is simply to say that the organism contains a mechanism or mechanisms -- an adaptation -- of a kind that historically proliferated at least partly in virtue of getting ancestor organisms, often enough, to conform to an at least primitive moral rule. It may be possible to reduce the reliance on genetically determined mechanisms if the organism is able to learn the appropriate rule-conforming competence and pass it on, often enough, to others, including offspring, and to refuse to play with those who either don't learn it or don't conform to it. In this case, we would be dealing with a less primitive kind of objective normativity. (And in either case the proposition of protomoral obligation would again be intrinsically action-guiding in the sense of entailing a rule or standard or norm.)

Of course this falls well short of fully human altruism, as Kitcher emphasizes; you can't get here from there. That is, you can't infer that we humans ought to conform to golden-rule altruism of a discriminating kind from the fact that our savannah-dwelling hominid ancestors had a primitive moral obligation to do so, at least not without helping yourself to further and highly contestable assumptions. Nonetheless, we would at least be entitled not only to say with Kitcher that "a recognizable, if rather minimal, type of human altruism" might evolve under natural selection. We could also conclude that conformity to this altruism would be objectively primitively obligatory for relevantly similar creatures in the relevant kinds of environment. If nothing else, this should disturb the slumber -- or rattle the cage -- of those convinced that there can be no sort of objective normativity in the world, and certainly none that is remotely like a moral ought. . . .

TO BE CONTINUED . . . .

FOOTNOTES

1. This paper is a much revised and extended version of "How to Get an Ought from a Biological Is," presented at the Invited Symposium on the Evolution of Norms, APA Pacific Division meeting, Albuquerque, April 7, 2000, Gilbert Harman commenting. For comments on ancestor drafts, I am indebted to Tom Bontly, Allen Coates, Derek Turner and audiences at Duke, Virginia Polytechnic and Western Michigan universities and the North Carolina Philosophical Society, and especially to Michael Ferejohn for detailed and persistent critique in a follow-up correspondence of some weeks.

2. Mackie (1977), 38-39.

3. Cf. Hampton (1995), 110ff, according to whom Mackie never explicitly defines "what is scientifically problematic about moral prescriptivity," whereas "a good look at science reveals that valuation is not something that is, or need be, excluded from our best scientific explanations, especially in fields such as biology." One consequence of my argument will be that evolutionary biology is up to its neck in philosophically relevant kinds of objective normativity.

4. Mackie (1977), 42.

5. Searle (1992), 238. The notion of a norm or standard in the present paper, which is also what Searle seems to have in mind, is essentially Allan Gibbard's (1990), 46, endorsed by Copp (1995), 196: a norm or standard is "a possible rule or prescription, expressible by an imperative."

6. Kitcher (1993), 513.

7. Strawson (1963), xiii.

8. Moore (1980), 15ff. Cf. Ball (1988).

9. Hare (1995), 340.

10. Horgan and Timmons (1992b).

11. Cf. Post (1995), 88-90.

12. What Horgan and Timmons (1992a), 235, call "hybrid semantic constraints" that govern certain terms seem able, if not also designed to accommodate (among other things) the novel usage typical of revisionary theories. If their constraints prove unable do so, their account as it stands appears to provide no other way of allowing for such usage -- a measure, perhaps, of how strong is the influence of OQA on them. Cf. Horgan and Timmons (1992b), (1993), considered at length in §V(ii) below.

13. Mackie (1977), 15.

14. Mackie (1977), 53-58.

15. Here I adopt the interpretation of Mackie's AQ given by Horgan and Timmons (1992a), at least for the duration of the present paper.

16. Mackie (1977), 4. Another part of Mackie's argument -- that something's being morally good, and/or the judgment that it is, is intrinsically or necessarily motivating and therefore must be ontologically queer from the point of view of naturalism -- is addressed at relevant places in the following sections.

17. Mackie (1977), 42.

18. Thus I take quite seriously, and address in §V(iii), the challenge advanced by Horgan and Timmons (1992a) against earlier efforts to outflank AQ, namely that such efforts fail to discharge the explanatory burden expressed in (b).

19. Kripke (1982), 24.

20. I am indebted to Ümit Yalçin for getting me to see that there may be some esthetic normative properties, perhaps among others, to which this distinction does not apply.

21. Kim (1993), 345, endorses and calls this Alexander's dictum. Post (1999a) argues that both the dictum and Kim's closely related principle of the Causal Individuation of Kinds suffer serious empirical counter-examples, especially from biology.

22. On the relations between determination and explanation, cf. Post (1999b).

23. Weinberg (1976), 13.

24. Brandon (1990); all quoted Brandon passages below are from pp. 139, 165, 185-89.

25. Dennett (1995), 229, 236-238.

26. Cf. Sober (1993), 84: "characteristic c is an adaptation for doing task t in a population if and only if members of the population now have c because, ancestrally, there was selection for having c and c conferred a fitness advantage because it performed task t." I'm indebted to Derek Turner for recalling this passage. See also Sterelny and Griffiths (1999).

27. More fully, an ideally complete adaptation explanation requires supplying five kinds of information: "(1) Evidence that selection has occurred, that is, that some types are better adapted than others in the relevant selective environment (and that this has resulted in differential reproduction); (2) an ecological explanation of the fact that some types are better adapted than others; (3) evidence that the traits in question are heritable; (4) information about the structures of the population from both a genetic and a selective point of view, that is, information about patterns of gene flow and patterns of selective environments; and (5) phylogenetic information concerning what has evolved from what, that is, which character states are primitive and which are derived." Cf. Dennett (1995), 238: "Adaptationist reasoning is not optional; it is the heart and soul of evolutionary biology."

28. Of course there can be more than one effect in virtue of which A was selected for, but talking as if there were just one, as I shall continue to do, simplifies the exposition. Also, A's past instances need not always have had effect E, or even very often, just often enough for there to have been selection for A.

29. Cf. Ruth Millikan's distinction between direct and adapted proper function, in Millikan (1984), Chs. 1-2, especially p. 40ff.

30. Millikan (1993), 226. For simplicity, and where no confusion looms, "A is for E" will mean that A is either directly for E or for E in some further sense; otherwise "A is d-for E" will mean that A is directly for E, and other locutions from §VI will designate further ways A can be for E.

31. McGinn (forthcoming).

32. Millikan (1984), (1993).

33. Searle (1995), 18. Plantinga (1993), 201-204.

34. Searle (1995), 18.

35. According to Post (1998), 233, Plantinga "badly misreads [Millikan] as attempting an analysis, then tries to counter-example accordingly." Taken literally this of course is not true, as Plantinga has stressed in correspondence. What I meant, and should have spelled out, was the misreading at a deeper level, spelled out here, of the significance of the method. I had supposed it would suffice for clarity to say in the next breath, as I did, that "the aim of her definition . . . is to be adequate to the phenomenon. Objecting to the resulting notion [of function] that it is not what the folk mean . . . would be as beside the point as objecting to Einstein's account of mass that it fails to accord with what the folk mean. . . ." Obviously it did not suffice.

36. Horgan and Timmons (1993), 188. I am indebted to Tom Bontly in correspondence for stimulating me to think more deeply about the Horgan and Timmons New OQA, and to try constructing a genuine naturalistic Twin-Earth scenario for certain normative terms, below, contrary to their argument that there cannot be one. Cf. Bontly (1998).

37. Less primitive cases, including those in which what is to be pursued involves conforming to a rule, are considered in the next section (§VI).

38. Horgan and Timmons (1993), 188.

39. Horgan and Timmons (1993), 188.

40. Unless of course the relevant bridge principle of the kinetic theory -- the principle that equates temperature with mmke -- has itself become part of the vernacular, having long schooled our intuitions about the relations between temperature and mmke. In that case, (1) change the example to a time before the theory had yet schooled the vernacular, and (2) see below, at running a Twin-Earth scenario against Einstein, on the problem of what counts as the relevant group of competent speakers, and when.

41. Again see Horgan and Timmons (1993), 188.

42. Cp. Horgan and Timmons (1992b), 159. CSN* is the same as their CSN, except that 'moral' is replaced by 'normative.'

43. Horgan and Timmons (1992b), 157ff.

44. Horgan and Timmons (1992a), 250n6 to p. 224, remark that "In the case of moral supervenience, presumably the operative modality is metaphysical necessity -- truth in all possible worlds." But they appear to give no real argument for this, and in any case it does not follow that the operative modality is metaphysical necessity, when the theory claims only that the specific upper-level properties have necessary and sufficient conditions -- expressed say by DFOR and DTPE -- in natural affairs in the ppw's in which the relevant background conditions happen to obtain.

45. Horgan and Timmons (1992b), 158; Boyd (1988).

46. Horgan and Timmons (1992b), 162, emphasis added.

47. In line with Horgan and Timmons (1992b), 164.

48. Horgan and Timmons (1992b), 165-166.

49. Cf. Post (1995), pp. 89-93 on focused determination.

50. Horgan and Timmons (1992a).

51. Searle (1995), 18.

52. Searle (1995), 16.

53. Searle (1995), 18-19.

54. Dennett (1995), 339.

55. Plantinga (1993), Ch. 11.

56. What Baker (1987) and many others call individualism amounts to non-relational individualism; what they call non-individualism amounts to a kind of relational individualism.

57. Audi (1993), 96.

58. Kim (1993), 58, 79, 84.

59. As for example in defining the relation of focused supervenience/determination. Both relations are defended at length in Post (1987), (1995), and (1999a). Cf. Hellman and Thompson (1975), Horgan (1982), Horgan (1984), Lewis (1983). In the present paper, "reduction" means property-property reduction; a nonreductive account is one in which not all a thing's properties need be identical or (nomologically) equivalent to certain of its own base properties (intrinsic or relational).

60. Kripke (1982), 24, 27-32, 57.

61. Cf. Post (1995), §§2-3, which contains further references, and where the objection that there must be a (compound) physical relational property of the bearer in virtue of which the determination holds is met by showing how the candidate relations all fail to do so in representative cases; the culprit is the tendency of many philosophers to chase propertyhood up the tree of syntax, so that a property is anything projected by the predicate-forming operators of logic and set theory. Note also that WS, as defined by Kim, automatically fails for self-effacing properties, since WS -- and thereby SS -- requires a subvening property to be a property of whatever has the supervening property.

62. Railton (1995), 100n18.

63. They also fit the top-down or unification model, in that "All adaptation explanations, no matter how diverse the adaptations, subsume the explained adaptation under the [Principle of Natural Selection]," where PNS is "If a is better adapted than b in environment E, then (probably) a will have greater reproductive success than b in E." Brandon (1990), 11; see 139ff for what sort of role this schematic and seemingly tautological law plays in the theory of natural selection. See note 24, above, for page references to the following quotations from Brandon (1990).

64. For a vivid account of such explanation and prediction, see Dennett (1995) on adaptationist reasoning and reverse engineering, and especially p. 484, on the prediction of the existence of the naked mole rat.

65. It is also a vindication, or at least a confirming instance, of a crucial part of the nonreductive determinationist metaethics introduced in Post (1987), Ch. 6.

66. Boyd (1988), §4.4

67. This sense of "intrinsically action-guiding" is essentially what Copp (1995), 199, and Svavarsdóttir (1999), 162, mean when they distinguish this or equivalent terms from "intrinsically motivating." The notion of a norm or standard at work in the present paper is, again, essentially Gibbard's (1990), 46, endorsed by Copp (1995), 196: a norm or standard is "a possible rule or prescription, expressible by an imperative." If Copp (1995), 199-200, is right and we apply his account here, then the present account is compatible with both internalism and externalism.

68. Kitcher (1993), 513.

69. That is, in relation r to some y such that y not- = x. RFOR is patterned after Millikan (1984), 39, on relational proper function. Her account of the various kinds of proper function or what-for-ness is far more comprehensive and detailed than the present account is or needs to be, given its different aims.

70. Again see Millikan (1993), 226.

71. Millikan (1993), 218.

72. Collett and Land (1978), quoted by Millikan (1993), 218.

73. Millikan (1993), 219, and 210-239 on purposefully conforming to an unexpressed rule.

74. Millikan (1993), 224.

75. Cf. Millikan (1993), 211-239.

76. Philosophers so inclined are welcome to hold instead that the point is to overrule following one's own self interest when everyone's following self-interest would be harmful to everyone. Cf. Baier (1958), 309: "The very raison d'etre of a morality is to yield reasons which overrule the reasons of self-interest in those cases when everyone's following self-interest would be harmful to everyone." Adopting this condition has no affect on the relevant arguments below; likewise for a number of other variations on the general theme.

77. And the stickleback, among others. Cf. Pool (1995), from which the following is drawn.

78. Cf. Axelrod and Hamilton (1981); Axelrod (1984); Axelrod and Dion (1988); and Kitcher (1993), which contains further references.

79. As opposed to how actually it evolved. Kitcher (1993), note 6.

80. Kitcher (1993), 513. He labels this rule DA_0.5R. An Altruist rule is Discriminating if it requires the organism to play with any organism that has never defected on it, and, when playing, always cooperate. DA_0.5R is also unforgiving: the organism is never to play with anyone who has ever defected on it. Both more and less unforgiving strategies are possible, though probably dominated by DA_0.5R.

REFERENCES

Audi, Robert (199). "Ethical Naturalism and the Explanatory Power of Moral Concepts" in Naturalism: A Critical Appraisal, ed. Steven J. Wagner and Richard Warner (Notre Dame, University of Notre Dame Press), 95-115.