Is constructivism an alternative to moral realism?

ch apter five Is constructivism an alternative to moral realism? David Copp Meta-ethical constructivism is sometimes viewed as an anti-realist alternative to moral realism. It is thought to avoid some of the difficulties of realism without lapsing into either an error theory or non-cognitivism. I will contend, however, that this is a misleading way to think of the matter. The distinction between constructivism and the kinds of views that would ordinarily be classified as realist is both fuzzy and insubstantial. Nothing substantive turns on how we use the term “realism,” of course, nor on how we use the term “constructivism.” But there are two substantive meta-ethical issues that are worth worrying about, and on both issues, the most interesting forms of meta-ethical constructivism agree with some of the most interesting and ambitious forms of non-constructivist moral realism. Moreover, many kinds of constructivism can be paired with non-constructivist sister theories that would ordinarily be classified as realist, and in many of these cases the constructivist and non-constructivist siblings do not differ in ways that are of any major philosophical significance. In this essay, then, I will be arguing that meta-ethics should avoid allowing itself to be distracted by a putative distinction between constructivism and realism or between constructivism and non-constructivism. Instead of worrying about where to draw these lines and attempting to defend the idea that the truth lies on one side of one of the lines or the other, we should focus on more substantive meta-ethical issues. In section 1, I explain what I take to be the most important meta-ethical issues that have a bearing on the plausibility of constructivism. In section 2, I provide examples of constructivist theories. In section 3, I provide a A version of this essay was presented in October 2011 to the Department of Philosophy at Bowling Green State University, and in April 2012 to the Department of Philosophy at the University of Vermont. I thank Christian Coons, Terence Cuneo, Don Loeb, Mark Moyer, Sara Worley, and everyone else who contributed to the discussions on these occasions for their helpful comments. I am especially grateful to Carla Bagnoli and Paul Formosa for extensive comments on an early draft. 108 9781107019218c05_p108-132.indd 108 3/21/2013 4:31:33 PM Is constructivism an alternative to moral realism? 109 characterization of constructivism. In section 4, I discuss the issue whether constructivism is a form of realism. In section 5, I discuss the relation between constructivism and moral naturalism. In section 6, I point out that constructivist theories often have non-constructivist sister theories that differ from them in apparently unimportant ways. Several philosophers have proposed arguments that claim to show all at once that the truth lies on the realist side, or that it lies on the constructivist side. As I explain in section 7, I am pessimistic about the prospects of such arguments. In the final section I draw conclusions. 1 Some issues worth worrying about The first issue of major significance concerns the nature of normativity. Some theories aim to explain normativity by giving an account of what normativity consists in, but some theories instead take normativity as a primitive, or as given, and some even appear to view normativity as a kind of illusion. For our purposes, we do not need to settle on a characterization of what exactly such theories might attempt to explain. The point is that constructivist theories are typically motivated by the aim to explain normativity. Many realist theories do not share this aim. Non-naturalist theories typically take normativity as a primitive that is not open to explanation, and some naturalist theories attempt to explain normativity away, as a kind of illusion. But there are naturalist theories that aim to explain normativity in terms of naturalistic phenomena of which we have an independent understanding. From the perspective of this first issue, then, the distinction between constructivist and realist theories is uninteresting. The important divide is between theories that aim to explain normativity and those that do not. It would not be accurate to say that all constructivist theories aim to explain normativity. Some forms of Kantianism take it as given that there are reasons of certain kinds, and some take certain norms of rationality as given. However, the claim that there is a reason to do such-and-such or that rationality requires such-and-such are normative if any are. Hence, a theory that takes reasons or norms of rationality as given is importantly similar to a realist non-naturalist theory that takes normativity as primitive. The important distinction, between theories that aim to explain normativity and theories that do not, cuts across the distinction between realism and constructivism. The second issue of major importance is about cognitivism and truth. Some theories take moral and other normative claims to be 9781107019218c05_p108-132.indd 109 3/21/2013 4:31:34 PM 110 David Copp truth-evaluable.1 Call these theories “cognitivist.” Some cognitivist theories add that some substantive normative claims are true. Call these theories “success theories.” If a success theory is compatible with a robust non-deflationary theory of truth call it a “Success theory” (with a capital “S”).2 Most forms of constructivism aim to be Success theories in this sense since they aim to be non-committal on questions such as the nature of truth. On this level, these theories are similar to realist theories that are also Success theories. There are two additional distinctions that should be noticed in this context. First is a distinction between theories that aim to provide a substantive, non-trivial, and philosophically interesting account of the truth conditions of moral judgments and theories that do not aim to do this.3 In this respect, meta-ethical constructivism is closely allied to certain forms of moral naturalism, since, like most naturalist theories, constructivist theories also aim to provide substantive accounts of the truth conditions of moral judgments. Second is a distinction between theories according to which the truth conditions of moral judgments are “mind-dependent” and theories according to which this is not the case. I will not here attempt to define mind-dependence, but I will count a theory as “Mind-Dependent” if its account of the truth conditions of moral judgments invokes counterfactual claims about what persons would judge or choose, or about their attitudes or desires, in hypothetical circumstances of some kind. Constructivist theories are Mind-Dependent for they hold that, roughly, the truth of a moral judgment depends on whether it is appropriately related to the moral principles that would be endorsed by relevant agents who engaged in an idealized process of reasoning. Some forms of naturalist moral realism are also Mind-Dependent in this sense, as we will see. There are then three important distinctions with respect to the issue of cognitivism and truth, and all interesting forms of constructivism lie on the same side of each of these distinctions with Mind-Dependent forms of moral naturalism. The most interesting constructivist theories, like the most ambitious forms of moral naturalism, are Success theories that aim to provide substantive accounts of the truth conditions of moral 1 2 3 All normative judgment are relevant here, including judgments about practical reasons in addition to moral judgments. I will focus mainly on moral judgments. For the distinction between deflationary accounts of truth and robust, non-deflationary accounts, see Beall and Glanzberg 2008. I think all forms of moral realism are Success theories even though, of course, a realist might have independent reasons to favor one or another account of truth. I return to this issue in section 4. The issue whether such an account of truth conditions can be defended is different from the issue whether truth itself is a substantive property. See Beall and Glanzberg 2008. 9781107019218c05_p108-132.indd 110 3/21/2013 4:31:34 PM Is constructivism an alternative to moral realism? 111 judgments; moreover, constructivist theories, like some forms of moral naturalism, are Mind-Dependent. Given this, it seems that from the perspective of the issues about cognitivism and truth, the distinction between constructivism and realism is uninteresting as such. The distinctions that are most important are between Success theories and other theories, and, among Success theories, between theories that provide substantive mind-dependent accounts of the truth conditions of moral judgments and other theories. As I will suggest in section 4, Mind-Dependent forms of moral naturalism are forms of moral realism, as realism is best understood. And, as I will explain in section 6, many of the most interesting constructivist theories have sister theories that are forms of Mind-Dependent moral naturalism. Moreover, it is difficult to see any philosophical importance in the distinction between a constructivist theory and its sister naturalist theory. For these reasons, I think meta-ethics should not be distracted by the debate between constructivism and realism. It should instead focus on the tenability of Mind-Dependent accounts of the truth conditions of moral judgments and on the tenability of the different theories that aim to explain normativity. Some of the interesting theories are constructivist, and some are not. 2 Kinds of meta-ethical constructivism I am here concerned with constructivist meta-ethical theories that purport to account for the truth conditions of moral judgments, or, more generally, of normative judgments. John Rawls presents his theory of justice as a form of “Kantian” constructivism (1980/1999b: 303), but it is not a form of meta-ethical constructivism since it aims only to specify truth conditions or correctness conditions for “the first principles of justice” (1980/1999b: 304). It does not purport to specify truth conditions for all moral judgments without restriction, much less for all normative judgments. It will nevertheless be useful to consider how Rawls characterizes constructivism. In abstract terms, Rawls says, his theory proposes a conception of the person and a “procedure of construction” such that the principles of justice are related suitably to the specified conception of the person and the specified procedure (1980/1999b: 304). The Rawlsian procedure of construction is the process of deliberation that would be followed by rational persons behind Rawls’ “veil of ignorance” in the Rawlsian “original position”; the principles of justice are the principles that people in the original position 9781107019218c05_p108-132.indd 111 3/21/2013 4:31:34 PM 112 David Copp would choose to serve as the public principles of justice in their society. And so, Rawls says, the principles of justice are “constructed” by the parties to the original position (1980/1999b: 311). In his view, there are no moral facts “apart from the procedure of constructing the principles of justice” (1980/1999b: 307). Rawlsian constructivism is restricted to the problem of “justifying a conception of justice” (1980/1999b: 306).4 Following the Rawlsian schema, we might then say, in general terms, that a restricted form of moral constructivism would contend that the true principles concerned with a given moral subject matter would be the output of a specified procedure of construction. The theory would then claim that there are no moral facts concerning this subject matter apart from those yielded by this procedure. A meta-ethical constructivism would eliminate the restriction to a given moral subject matter. It would claim that there are no moral facts at all, or perhaps no normative facts at all, apart from the procedure of construction. Rawls says that, in the constructivist view, the construction of principles “replaces the search for moral truth interpreted as fixed by a prior and independent order of objects and relations” (1980/1999b: 306). He says it is best to view moral principles, not as true, but as “reasonable for us” (1980: 519). In these places, Rawls does not restrict his claim about the search for moral truth to a claim about justice. The important point is that, for the constructivist, the “procedure of construction” is not merely of epistemological or methodological significance. It is not merely a procedure for discovering the facts. It is metaphysically significant since it is thought to determine what the relevant facts or truths are.5 A variety of meta-ethical theories seem to be sufficiently similar to the Rawlsian view in relevant ways to count as plausible examples of constructivism. Let me roughly sketch four such theories and briefly explain why they qualify as constructivist. First is the Kantian theory that to conceive of oneself as free and rational, one must conceive of oneself as bound by the categorical imperative (Kant 1981). The Kantian then contends that we are in fact bound by the categorical imperative since we cannot but see ourselves as free and rational. The idea seems to be that there are no moral facts independently 4 5 Street (2008a) calls it a kind of “restricted constructivism.” I follow her usage. One can distinguish between metaphysically robust conceptions of facts and deflationary conceptions according to which to say it is a fact that p is simply to say that p is true. My arguments are meant to be compatible with both conceptions and also with both deflationary and non-deflationary conceptions of truth. See Beall and Glanzberg 2008. 9781107019218c05_p108-132.indd 112 3/21/2013 4:31:34 PM Is constructivism an alternative to moral realism? 113 of how we must conceive of ourselves. Moreover the facts as to what we are permitted to do are determined by the categorical imperative, which says that we must act only on maxims that we can will to be universal laws. So the content of morality and the fact that we are bound by morality depend on how we must conceive of ourselves and on what we can rationally will.6 Second is David Gauthier’s idea that rational agents, aiming to maximize their expected utility, would in suitable conditions choose to become “constrained maximizers.” Gauthier argues further that constrained maximizers would reach agreements among themselves on the division of the fruits of their cooperative activity. Moral requirements are the terms of such rational agreements (Gauthier 1986). Gauthier seems to take it as a given that rational agents acting in isolation seek to maximize their expected utility. In Gauthier’s view, as I understand it, there are no requirements to constrain our maximizing activity, neither to enable us to cooperate nor to show respect for each other, unless it is rational to constrain our maximizing activity. Gauthier then argues that it is rational to do so. In his view, the facts about how it is rational to choose determine the content of morality. Third is Christine Korsgaard’s proposal that anyone who values anything at all is committed to valuing rational reflection in all its instantiations, and so is committed to valuing persons. But it is essential to practical rationality that one value something, and so anyone who is practically rational must value persons, and this commits him to morality. This is what binds us to comply with moral requirements (Korsgaard 1996b). In Korsgaard’s view, the fact that we are bound by morality, and the content of morality, are determined by facts about what we are committed to in virtue of our having certain values, given that having values is essential to practical rationality. Fourth is Sharon Street’s idea that the facts about a person’s reasons for action are constituted by facts about the judgments about his reasons that would withstand scrutiny in a suitably constructed reflective equilibrium (Street 2008a: 223, 238–39). Street holds that we have no reasons until we judge that we do. And the reasons a person has are determined by what reasons he would judge himself to have if his overall evaluative perspective were coherent and in reflective equilibrium (Street 2008a: 222). His actual reasons are the reasons he would take himself to have if his state of mind were rationally coherent. 6 Kant can also be interpreted as a non-constructivist. I return to this point in section 6. 9781107019218c05_p108-132.indd 113 3/21/2013 4:31:34 PM 114 David Copp These four theories are very different from one another but they share important features. Most importantly, they share the idea that the existence of moral facts – or the existence of any normative facts at all – depends on facts about what persons would be rational to will or to choose or to value, given that they value something, or given that they must value something in order to be rational. These theories do not all use the metaphor of construction, but they share the idea that the moral or normative facts are mind-dependent and dependent specifically on the “deliberative” intentional states that persons would have in specified hypothetical circumstances. By a “deliberative” intentional state, I mean an intentional state such as a willing, a choosing, or a valuing that has a desire-like direction of fit and that is normally subject to deliberative control. Each of the four theories holds that the moral or normative facts are dependent on facts about what deliberative intentional states we would be rational to have in specified circumstances. This, however, is only an approximate characterization of the constructivist view. In the next section, I will consider some characterizations that are found in the literature. 3 The characterization of meta-ethical constructivism According to Darwall et al., “[T]he constructivist is a hypothetical proceduralist. He endorses some hypothetical procedure as determining which principles constitute valid standards of morality … [He] maintains that there are no moral facts independent of the finding that a certain hypothetical procedure would have such and such an upshot” (1992: 140). This suggestion is based in Rawls’ writings, for, as we saw, Rawls held that there are no facts about the requirements of justice “apart from” the results of the relevant procedure of construction (1980/1999b: 304, 307). The Rawlsian procedure is constrained by norms of practical reason, but as Darwall et al. see matters, this is Rawlsian constructivism, not constructivism in general (1992: 138). In general, what constructivists share is the idea that the moral facts are determined by the outcome of a hypothetical procedure of some kind. Shafer-Landau takes a similar view. In his words (2003: 14): Constructivists endorse the reality of a domain, but explain this by invoking a constructive function out of which the reality is created. This function has moral reality as its output. What distinguishes constructivist theories from one another are the different views about the proper input … What is common to all constructivists is the idea that moral reality is constituted 9781107019218c05_p108-132.indd 114 3/21/2013 4:31:34 PM Is constructivism an alternative to moral realism? 115 by the attitudes, actions, responses, or outlooks of persons, possibly under idealized circumstances. These “attitudes, actions, responses, or outlooks” can be taken to constitute a “standpoint,” so Shafer-Landau adds that, according to constructivism, moral standards are “made true” by being endorsed from a preferred standpoint (2003: 16).7 Accordingly, he says, constructivism denies the “stance-independence thesis.”8 Yet Shafer-Landau’s characterization allows a constructivist to maintain that the constructive function yields the same set of moral truths given any input of the right kind. Constructivist theories can be “objectivist” in this way. The idea of a constructive function is vague, and so is Darwall et al.’s idea of a procedure. Yet it seems that any procedure of the kind that Darwall et al. have in mind could be characterized by a function that takes the hypothetical attitudes of hypothetical participants into such and such moral standards. To be sure, some constructive functions of the kind Shafer-Landau has in mind might be too abstract to count as procedures. Despite this, however, these authors seem to have basically the same understanding of constructivism. For my purposes, Shafer-Landau’s characterization of constructivism seems not to be importantly different from Darwall et al.’s. On both accounts, constructivist theories reject what Shafer-Landau calls the “stance-independence thesis.” In Darwall et al.’s account, the attitudes of the participants in a relevant hypothetical procedure presumably would play a role in determining the output of the procedure, so constructivism in their view appears to yield mind-dependent truth conditions for moral judgments. Hence, in their view, constructivism rejects “stance-independence,” just as Shafer-Landau takes it to. Given this, and given the vagueness of the idea of a constructive function or procedure, one might suggest classifying all theories that propose mind-dependent truth conditions for moral judgments as “constructivist.” This terminological choice would be misleading, it seems to me, because many Mind-Dependent theories have different structures and motivations from the theories that are typically viewed as constructivist. There are, for example, sensibility theories (Darwall et al. 1992: 152–65), 7 8 Shafer-Landau here uses the term “standards” to refer to principles that can be true or false (2003: 15, n. 2). This differs from my usage in other writings but I will follow Shafer-Landau’s usage in this essay. Shafer-Landau 2003: 15. He says, “Realists believe … that the moral standards that fix the moral facts are not made true by virtue of their ratification from within any given actual or hypothetical perspective” (2003: 15). 9781107019218c05_p108-132.indd 115 3/21/2013 4:31:34 PM 116 David Copp response-dependence theories, projectivist theories, and theories of other kinds as well (see D’Arms and Jacobson 2006 and Slote 2006). Constructivist theories are Mind-Dependent, but not all Mind-Dependent theories are usefully classified as constructivist. The idea of a deliberative intentional state can be used to clarify the idea of a procedure. As an example, consider the legislative process, which can be described as a deliberative or reflective process that results in a decision either to favor or to disfavor a piece of legislation. I propose that the hypothetical procedures that Darwall et al. have in mind are, similarly, hypothetical deliberative or reflective processes that the relevant participants could engage in, at least in principle, with the result that they would reach a deliberative intentional state of a specified kind, such as a willing, choosing, or valuing of something. A constructivist theory would specify such a process and propose that the true moral principles stand in a specified relation to the objects of the specified deliberative intentional state. In Shafer-Landau’s terms, a process of this kind could be described as a function from “the attitudes, actions, responses, or outlooks of persons, possibly under idealized circumstances,” to an output of deliberative intentional states, the objects of which determine the content of the true or correct moral principles. For instance, Korsgaard holds, very roughly, that the content of morality is determined by what we would value if we were to engage in a process of rational reflection that led us to value what our initial values commit us to valuing. We may speak, then, of the “proceduralist characterization” of constructivism (see Street 2010: 365). I formulate it as follows: The Proceduralist Characterization of Meta-ethical Constructivism: a constructivist theory defines a hypothetical reflective or deliberative procedure such that relevant agents engaging in the procedure, perhaps in ideal circumstances, would reach a deliberative intentional state of a specified kind, such as a willing, choosing, or valuing of something, where certain moral principles stand in a specified relation to the objects of the specified deliberative intentional state, and where the theory holds that these and only these principles are true and that they are true because they are in this way the outcome of the procedure. In brief, a constructivist theory defines a hypothetical procedure that could in principle be followed, where the outcome of the procedure is a set of standards that the theory holds to be true because they are yielded by the procedure. Street objects that the proceduralist characterization “fails to capture what is philosophically most interesting and distinctive” about 9781107019218c05_p108-132.indd 116 3/21/2013 4:31:34 PM Is constructivism an alternative to moral realism? 117 constructivism (2010: 364). First, constructivists can disagree as to the nature of the constructive procedure, and Darwall et al. suggest that such disagreement seems to be a normative disagreement. So it is not clear that constructivism is a meta-ethical position at all (Darwall et al. 1992: 140–41). Second, setting aside this first worry, it is not clear that constructivism is distinct from a standard kind of naturalist moral realism that reduces normative facts to facts about the responses of hypothetical persons in idealized circumstances.9 Street thinks that these worries are generated by the proceduralist characterization and are not objections to constructivism itself. I address these worries in sections 4 through 6. Street proposes a “practical standpoint characterization” of constructivism (2010: 366). Constructivism holds, she suggests, that there is no normative truth independent of a practical point of view, the point of view of an agent who values something or who takes something to give her a reason or to be valuable (2010: 366). According to meta-ethical constructivism, she says, “the truth of a normative claim consists in that claim’s being entailed from within the practical point of view, where the practical point of view is given a formal characterization” (2010: 369, emphasis in original). To give the practical point of view a formal characterization, she explains, is to characterize it in terms of what is involved in valuing anything at all. The idea, she says, is that standards of correctness in normative judgment are “generated” by the attitude of valuing just as such (2010: 369). On Kantian versions of constructivism, Street suggests, moral values are entailed from within the standpoint of any valuer (2010: 369). On Humean versions, the reasons a person has depend on what she herself actually values together with the constraints on valuing that are implicit in the attitude of valuing just as such (2010: 369). Let me set aside the question of what Street means by saying that some proposition is entailed “from within the practical point of view as such.” In a different paper, Street offers a different characterization of constructivism. She says that, according to meta-ethical constructivism, “the fact that X is a reason to Y for agent A is constituted by the fact that the judgment that X is a reason to Y (for A) withstands scrutiny from the standpoint of A’s other judgments about reasons” (2008a: 223; see 233, n. 41). She explains that for such a judgment to withstand such scrutiny is in effect for it to be a judgment that the person could coherently accept in reflective equilibrium, given the judgments about reasons with which he begins (Street 2008a: 223, 238–39). 9 Street 2010: 365. Street cites David Enoch (2009) for this point. 9781107019218c05_p108-132.indd 117 3/21/2013 4:31:34 PM 118 David Copp What is the relation between Street’s (2010) practical standpoint characterization and her (2008a) “withstanding scrutiny characterization”? They are different accounts. We could accept the withstanding scrutiny characterization without thinking that anything at all is “entailed from within the practical standpoint just as such.” So we could consistently deny the practical standpoint characterization. And we could accept the practical standpoint characterization without thinking that what reasons a person has depends at all on what judgments about reasons the person already accepts or on what judgments about reasons would withstand scrutiny from the standpoint of the person’s other judgments. So we could consistently deny the withstanding scrutiny account. The two accounts seem to be independent of each other. The withstanding scrutiny characterization appears to be a characterization of Street’s own specific view as she develops it in her 2008a paper. It fails to count any of the other three positions I sketched in section 2 as constructivist. Neither Kant, Gauthier, nor Korsgaard would agree with Street that what reasons a person has is determined by what judgments about reasons would withstand scrutiny from the standpoint of the person’s other judgments about reasons. Street imagines an ideally coherent Caligula (2010: 371). According to Street, if this Caligula’s judgments about reasons would survive the relevant kind of scrutiny, he would have the reasons he judges himself to have. But neither Kant nor Gauthier nor Korsgaard would agree, so their views do not count as constructivist according to the withstanding scrutiny account. I therefore set it aside as too parochial to serve as a general characterization of constructivism. Street’s practical standpoint account is of wider application than her withstanding scrutiny characterization since it allows both Kantian and Humean views to count as constructivist. Street also intends the account to embrace Korsgaard’s view (2010: 370). But Kant’s own view would not count as constructivist on this characterization since Kant holds roughly that our reasons are determined by the content of the maxims that we could or could not rationally will to be universal laws, not by what is entailed from within the practical point of view as such.10 Similarly, Gauthier’s view would not count as constructivist on Street’s characterization. Gauthier does not rest his view on claims about what would be 10 On Kant’s “Formula of Humanity” our moral reasons are determined by what is involved in valuing persons or their dignity. Yet this does not seem to be equivalent to, and does not commit Kant to, the idea that our reasons are determined by what is entailed from within the practical point of view as such. 9781107019218c05_p108-132.indd 118 3/21/2013 4:31:34 PM Is constructivism an alternative to moral realism? 119 entailed from within the practical standpoint as such. Someone might contend that Kant’s theory, or Gauthier’s theory, is entailed from within the standpoint of anyone who values anything at all. But this is not something that Kant or Gauthier appears to claim. Hence, even though Street’s practical standpoint characterization does better as a general characterization of constructivism than does her withstanding scrutiny characterization, it is still insufficiently general.11 The proceduralist characterization does better. On the proceduralist characterization, Street’s view can be described as specifying a hypothetical process of rational reflection that would lead rational people who followed the process to form the values that are entailed from within the practical standpoint as such, or to hold only the normative judgments that they could coherently accept in reflective equilibrium, where these values or judgments determine the content of the normative or moral facts. Hence, the proceduralist characterization counts the positions that Street has in mind as constructivist. It appears to have the generality that we want. I shall therefore work with the proceduralist characterization of constructivism. 4 Meta-ethical constructivism and realism Darwall et al. seem to view constructivism as compatible with any traditional meta-ethical position. They suggest that the disagreement between philosophers who propose different constructivist theories is at bottom a normative disagreement (1992: 141). If someone claims that one version of constructivism is correct, then if we dispute this claim, “old metaethical issues reappear.” Constructivists might hope to “render traditional metaethics obsolete,” but, Darwall et al. suggest, a more defensible kind of constructivism would be modest. It would aim to illuminate the nature of morality “without claiming to preempt traditional metaethics” (1992: 143–44). Of course restricted forms of constructivism, such as Rawls’ theory of justice, do not aim to answer meta-ethical questions. But Darwall et al. seem to be suggesting that even unrestricted constructivist theories that fit the proceduralist characterization do not answer any meta-ethical questions. 11 Street describes my society-centered theory as constructivist (Street 2008b), but it is not constructivist on either of her characterizations. My theory says nothing about what is entailed from within the practical standpoint as such, nor about what judgments would withstand scrutiny from the standpoint of agents’ other judgments about reasons. See Copp 1995; Copp 2007a: 13–26. 9781107019218c05_p108-132.indd 119 3/21/2013 4:31:34 PM 120 David Copp One of the important meta-ethical questions, however, is whether moral and other normative judgments have robust truth conditions and if so what these truth conditions might be. The theories I am interested in aim to answer this question. On the proceduralist characterization, a constructivist theory defines a hypothetical procedure that yields moral principles as its outcome where, according to the theory, these and only these principles are true, and they are true because they are the output of the procedure. Proceduralist theories aim, then, to provide an account of the truth conditions of moral judgments. Darwall et al. worry that traditional meta-ethical issues arise in debates between different versions of constructivism. Such a debate is akin to the debate between different versions of non-cognitivism or between non-cognitivism and moral naturalism. But the fact that one must defend a constructivist theory with meta-ethical arguments and the fact that constructivist theories are debatable does not show that these theories are not meta-ethical theories. Non-cognitivist theories and versions of moral naturalism are meta-ethical theories yet debates among them obviously raise meta-ethical issues. Let me turn, then, to the issue whether constructivist theories are an alternative to moral realism. Shafer-Landau and Street both view constructivism as standing in opposition to moral realism, and in this they follow Korsgaard. For instance, Shafer-Landau views constructivism as a kind of cognitivist anti-realism.12 It denies a thesis he takes to be central to realism, the thesis that some moral truths are “stance-independent” (2003: 15).13 He says, “Realists believe that there are moral truths that obtain independently of any preferred perspective, in the sense that the moral standards that fix the moral facts are not made true by virtue of their ratification from within any given actual or hypothetical perspective” (2003: 15). Meta-ethical constructivists deny this. On the proceduralist characterization, constructivists hold that the moral facts are determined by the outcome of a procedure that is stance-dependent. Despite this, as Shafer-Landau says, constructivism “endorses” the “reality” of the moral “domain” (2003: 14). It will be useful to distinguish between “basic realism” and “stance-independent realism,” which I define to be a kind of basic realism.14 I stipulate that basic realism accepts the following five doctrines. First, there are moral properties. The actions that are wrong are similar, of course, in that they are all wrong; that is, they have in common the property of 12 13 14 Shafer-Landau 2003. Regarding anti-realism, see pp. 15 and 17–18. Regarding cognitivism, see p. 39. Street agrees that realist theories accept the mind-independence of moral facts whereas constructivists deny mind-independence (2010: 371). In Copp 2006b (7–8), I equated moral realism with what I am here calling “basic realism.” 9781107019218c05_p108-132.indd 120 3/21/2013 4:31:34 PM Is constructivism an alternative to moral realism? 121 being wrong. There are familiar metaphysical debates about the nature of such similarities, but the moral or normative realist can remain neutral in these debates. She needs to insist merely, second, that moral properties have the same basic metaphysical nature as any other properties, including familiar non-normative properties, whatever that might be. Third, some moral properties are instantiated. Some kinds of action are wrong and some traits of character are virtues. That is, there are moral facts.15 Fourth, moral predicates are used to ascribe moral properties. The predicate “wrong” is used to ascribe wrongness, for instance. Moral predicates pick out respects in which things are morally similar and are used to describe things in terms of these similarities. Moral predicates do not work in any special way that distinguishes them fundamentally from ordinary descriptive predicates. Fifth, moral assertions express ordinary beliefs that have the same basic nature as other beliefs. Stance-independent realism adds the sixth thesis that at least certain moral facts are “stance-independent” and “mind-independent.”16 Theories that Shafer-Landau counts as “realist” appear to accept all six doctrines of stance-independent realism, but some theories that are ordinarily categorized as realist do not accept stance-independence. For instance, Peter Railton has proposed that “normative facts about an individual’s non-moral good” reduce to “facts about what that person would want himself to want if he were fully and vividly informed about himself and his circumstances” (Street 2010: 372; Railton 1986a and 1986b).17 On this view, normative facts about what is non-morally good for a person are not mind-independent, and Railton explains the moral facts in terms of facts about what is non-morally good for persons, so moral facts also are not mind-independent (1986b). Since Railton’s view is normally classed as realist, it is sensible to count all theories that accept the five doctrines 15 16 17 I intend this to be compatible with both robust and deflationary conceptions of facts. Furthermore, I intend basic realism to be compatible with both robust and deflationary conceptions of truth. Korsgaard distinguishes between “procedural realism” and “substantive realism” (Korsgaard 1996b: 34–37). This is not the same as my distinction between “basic realism” and “stance-independent realism.” She says that “substantive realism is a version of procedural realism” (1996b: 37, n. 56) just as, on my distinction, stance-independent realism is a version of basic realism. But as Korsgaard explains matters substantive realism is not a version of procedural realism for she says the procedural realist holds that there are moral truths because there are correct procedures for answering normative questions whereas the substantive realist thinks the moral truths exist independently of any such procedures (1996b: 36). For her, “procedural realism” is just “stance-dependent” constructivism while “substantive realism” is “stance-independent realism.” Hence, substantive realism is not a kind of procedural realism. I follow Street’s interpretation of Railton. Railton proposed the view in “Moral Realism” (1986b). It would be odd to classify it as anti-realist. 9781107019218c05_p108-132.indd 121 3/21/2013 4:31:34 PM 122 David Copp of basic realism as realist. This, of course, is merely a terminological recommendation. On the usage I will adopt, there are forms of moral and normative realism that accept stance-independence and there are forms that do not. Meta-ethical constructivists typically accept the five doctrines of basic realism even though they deny stance-independence. This is an important sense in which they “endorse” the “reality” of the “moral domain.” Divine command theory is a useful example. According to divine command theory, as I shall understand it here, a moral standard is true just in case God has commanded compliance with it, and it is true because God has commanded compliance with it. This is what makes it true.18 This is a constructivist theory where the relevant reflective procedure is the one that led God to issue the relevant commands. Shafer-Landau concedes that the theory seems to be realist (2003: 16, n. 4). I classify it as a kind of basic realism. Constructivists do not need to accept all five doctrines of basic realism, however. First, a constructivist could deny that there are any moral properties at all. Korsgaard thinks we cannot answer the question what to do by citing a property of an action, since we can intelligibly ask why we should take the action’s having this property to give us a reason for anything. She seems to deny on this basis that an action’s being morally required is a matter of its having a property.19 Perhaps she thinks there are no moral properties. Second, a constructivist might agree that there are moral properties but claim that moral properties have a significantly different metaphysical status from other properties. For facts about the instantiation of moral properties are the output of a constructive procedure and are not stance-independent. Third, a constructivist might deny that the state of mind a person expresses in making a normative assertion is an ordinary belief. Street holds that the state of mind so expressed is the state of valuing something, or the state of “normative judgment,” which she says is different from ordinary belief (2010: 376). However, it is hard to see why constructivists would not accept the doctrines of basic realism. They otherwise face familiar problems in semantics, metaphysics, philosophical psychology, and epistemology (e.g. 18 19 I am using “standard” as Shafer-Landau does, to refer to a kind of proposition. In previous publications, I have used “standard” to refer to norms that are expressible by imperatives rather than to a kind of proposition. Korsgaard 1996b: 33–35. She rejects “substantive realism” partly on the ground that it postulates “intrinsically normative entities,” by which she seems to mean moral properties. Korsgaard 1996b: 34–35. 9781107019218c05_p108-132.indd 122 3/21/2013 4:31:34 PM Is constructivism an alternative to moral realism? 123 Schroeder 2010). Better to avoid these problems if possible. Moreover, constructivists agree that there are moral truths. Perhaps, for example, torture and lying are wrong. If so, torture and lying share the characteristic of being wrong, which is otherwise called the property wrongness. This property is therefore instantiated. When we assert that torture is wrong we are ascribing wrongness to torture and we would be said to express the belief that torture is wrong. Moreover, it is hard to see why a constructivist would deny that moral properties, in that they are properties, have the same metaphysical status as other properties. Of course normative properties differ from non-normative properties, but so do psychological properties differ from geological properties. Moral facts are stance-dependent, according to constructivism, but facts about my likes and dislikes are also stance-dependent since they are after all facts about relevant stances. There is no need for a constructivist to hold that the moral properties differ in the way that they are properties from non-normative properties. And although normative beliefs obviously differ from non-normative beliefs, it does not follow just from this that normative beliefs differ metaphysically from ordinary beliefs. So I think it is natural to take meta-ethical constructivism to accept all five doctrines of basic realism and therefore to be a kind of realism. It is not that meta-ethical constructivism must be understood as a form of realism, but it is naturally read this way. 5 Meta-ethical constructivism and moral naturalism Moral naturalism adds to the five doctrines of basic realism the thesis that the moral properties are natural properties in some important sense. Meta-ethical constructivism seems to be compatible with this thesis. It appears that there could be a kind of meta-ethical constructivism that is also a version of moral naturalism. Elsewhere I have argued that there is a reconstruction of Korsgaard’s position that is compatible with moral naturalism (Copp 2007b). Suppose Korsgaard is correct that if an action is required of a person, then the person’s not performing it would conflict with her valuing rational reflective agency. Suppose also that a person cannot but value rational reflective agency unless she ceases to value anything, and thereby ceases to be able to see herself as an agent. On this view, there is a complex property that is possessed by actions that are morally required – the property an action may have of being such that, roughly, the agent’s failure to perform it would conflict with her valuing rational reflective agency where the agent 9781107019218c05_p108-132.indd 123 3/21/2013 4:31:34 PM 124 David Copp must value rational reflective agency if she is to value anything and to see herself as an agent. I have argued that this is a natural property, on a reasonable construal of what counts as a natural property (Copp 2007b: 278–79). This leaves the question whether, in Korsgaard’s view, this property is the property of being morally required or whether, instead, these are distinct properties although an action’s possession of Korsgaard’s property is what explains its being required. A naturalist who accepts Korsgaard’s arguments can claim that these are the same property and that being morally required therefore is a natural property. This would be a constructivist form of naturalism. My society-centered moral theory is also a constructivist form of naturalism, at least as it was originally formulated (Copp 2007a: 18–21). First, the theory is constructivist on the proceduralist characterization.20 As originally formulated, the theory says roughly that the “ideal moral code” for a society is the code the society would be rational to choose to serve in it as the social moral code, given the needs and values of the society and the society’s circumstances; it says that the content of the ideal code determines which actions are wrong in relation to this society, and so on. In effect, then, the theory defines a deliberative procedure that, it claims, a society could follow in principle such that the content of the moral code the society would be rational to choose to serve in it as the social moral code determines the content of the relevant moral facts.21 Second, as I have argued before, the theory is a form of moral naturalism (Copp 1995; Copp 2007a: 13–26). To be sure, it defines the relevant procedure in terms of rational choice, yet it specifies in naturalistic terms what rational choice consists in. Moreover, the output of the procedure determines in effect which properties are construed as being identical to the properties rightness, wrongness, and so on. As I have argued, these are natural properties. If so, the theory counts as a form of naturalism. It seems clear, then, that meta-ethical constructivism is compatible with naturalism. The outcome of the procedure specified by a given theory might be such that the property wrongness and other moral properties are plausibly identified with natural properties. 20 21 I explained in note 11 that the theory is not a form of constructivism on either of Street’s characterizations. On the proceduralist characterization, a procedure of the relevant kind is one that agents could in principle follow, but it might be questioned whether a society can be counted as an agent. But we needn’t worry here about whether to amend the proceduralist characterization. For the society-centered theory, as originally formulated, is recognizably constructivist on a proceduralist construal of the basic idea. 9781107019218c05_p108-132.indd 124 3/21/2013 4:31:34 PM Is constructivism an alternative to moral realism? 125 6 Non-constructivist sister theories I have argued before that many constructivist theories have non-constructivist sister theories (Copp 2005). Given a constructivist theory that plainly fits the proceduralist characterization, we can often define an extensionally equivalent non-constructivist sister theory that specifies the truth conditions of moral judgments without referring to a constructive procedure either to specify the truth conditions or to explain why these judgments are true. A constructivist theory defines a hypothetical reflective or deliberative procedure such that agents who followed the procedure, perhaps in ideal circumstances, would reach a specified kind of deliberative intentional state the content of which determines the content of the true moral principles. The theory holds that these principles are true because they are in the relevant way the outcome of the procedure. Given this characterization, we could produce a non-constructivist sister to a constructivist theory by making the following modifications. First, we could simply eliminate the explanatory claim. We could claim that the procedure yields all and only true moral standards without adding the claim that the standards are true because they are the outcome of the procedure. Second, we could specify truth conditions that are extensionally equivalent to the condition of being yielded by the constructive procedure but without relying on the procedural nature of the truth-maker in formulating the theory. In some cases we may be able to formulate such truth conditions without referring to anything plausibly construed as a procedure. Some properties are clearly procedural. For instance, the property of being legislated into law is obviously procedural, and this means that any plausible theory of the nature of this property would be constructivist.22 A non-constructivist theory of the nature of this property would be artificial at best, since to legislate something into law is to follow a procedure on any plausible understanding of matters. Now a friend of constructivism might contend that a non-constructivist sister of a constructivist meta-ethical theory would be similarly artificial and that its plausibility would be parasitic on the plausibility of its constructivist sibling. To see whether this is so in general, however, we need to consider some examples. I know of no argument that shows the moral properties to be procedural in their nature in the way that the property of being legislated into law 22 An example is the property of being the output of a procedure. Sara Worley made this point and gave the legal example, in discussion. 9781107019218c05_p108-132.indd 125 3/21/2013 4:31:34 PM 126 David Copp obviously is. Few philosophers would claim that the wrongness of an action, for example, depends on the actual output of any actual procedure in the way that whether something has been legislated does. The divine command theory comes close to making such a claim, so let us begin with it before turning to other examples. First, then, according to the divine command theory, any true moral standard is true just in case and because God has commanded compliance with it. The theory construes God’s commanding compliance with a standard as the result of a reflective process that God has engaged in, or that God could engage in, where the standard is true because it is in this way the outcome of this process. The theory’s non-constructivist sister says simply that a moral standard is true just in case God desires compliance with it. It says that God’s desires are the truth-makers for moral standards, but it does not view God’s desiring something as the outcome of a process, and the theory is silent about what explains the moral truths. It is a theory of truth conditions that bears an obvious resemblance to its constructivist sister, but it is non-constructivist (on the proceduralist characterization). First, it does not make the explanatory claim that is characteristic of constructivism. Second, it treats God’s desiring compliance with a standard as a state of affairs rather than as the outcome of a process.23 Second, consider a naturalistic theory that reduces normative facts to natural facts about the responses of agents under idealized circumstances. Railton has proposed that “normative facts about an individual’s non-moral good” reduce to “facts about what that person would want himself to want if he were fully and vividly informed about himself and his circumstances” (Street 2010: 372; Railton 1986a and 1986b).24 A constructivist sister to Railton’s theory would say that if G is non-morally good for a person, the fact that this is so is constituted by the fact that if the person were to deliberate about what he would want himself to want if he were fully and vividly informed about himself and his circumstances, he would want himself to want G. Railton’s non-constructivist theory proposes naturalistic truth conditions for propositions about persons’ non-moral good. Its constructivist sister adds that there is a reflective process that a person could follow in principle such that the outcome of the 23 24 Terence Cuneo and Don Loeb pointed out, in discussion, that it arguably is incoherent to view God as engaging in any processes since God exists “outside time.” If so, the constructivist version is incoherent. Again, I follow Street’s interpretation of Railton although I think it is inaccurate. Street asks what the difference is between Railton’s naturalism and the kind of constructivism she favors. I am asking a different question. 9781107019218c05_p108-132.indd 126 3/21/2013 4:31:35 PM Is constructivism an alternative to moral realism? 127 process determines both what is non-morally good for a person and why these things are non-morally good for the person. Third, Kant holds roughly that what we are permitted to do is determined by the content of the maxims that we could rationally will to be universal laws (Kant 1981). On the constructivist interpretation, we think of Kant as intending to specify a hypothetical procedure in which a person first identifies a maxim she proposes to act from, and then considers whether she could rationally will this maxim to be a universal law. The theory says that a person is morally permitted to act on a maxim just in case, and because, she could rationally will the maxim to be a universal law. On the non-constructivist interpretation, we do not think of Kant as specifying a procedure. Instead, we think of him as specifying formal properties a maxim must have in order, roughly, to qualify as a possible universal law. On this interpretation, the theory says that a person is morally permitted to act on a maxim just in case the maxim has properties such that it qualifies as a possible universal law. Since Kant’s view can be interpreted in both ways, it is understandable that Kant scholars disagree as to whether his view is constructivist.25 Fourth, consider Korsgaard’s theory. On a crude interpretation of her view, an action is morally required just in case an agent’s not performing it would conflict with her valuing rational reflective agency where a person cannot but value rational reflective agency unless she does not value anything at all and is therefore unable to see herself as an agent. We can view this theory as specifying a procedure of ideal rational reflection that would lead a person to see that she must value rational reflective agency in order to value anything at all and that there are types of actions that she must perform in order to avoid conflict with her valuing rational reflective agency. The theory claims that agents are required to do actions of these types because a failure to do these things would be incompatible with the agent’s viewing herself as an agent. The non-constructive sibling of this theory simply says that there is a moral requirement to do actions of type A just in case failing to do actions of type A would be incompatible with the agent’s valuing rational reflective agency. This theory gives us the truth conditions of judgments about what agents are morally required to do, but it does not present these truth conditions as a matter of what an agent would value if she were to engage in a procedure of ideal rational reflection and it does not make the characteristic constructivist explanatory claim. 25 For a non-constructivist interpretation, see Wood 1999: 129, 157–58 and Wood 2008: e.g. 283. For discussion, see Formosa 2011. Carla Bagnoli and Paul Formosa helped me with these ideas. 9781107019218c05_p108-132.indd 127 3/21/2013 4:31:35 PM 128 David Copp Fifth, the society-centered theory also has a non-constructivist sibling. As originally formulated, the theory is constructivist. It says roughly that the “ideal code” for a society is the code that the society would be rational to choose to serve in it as the social moral code, given the society’s needs, values, and circumstances, and it says that the content of the ideal code determines which actions are wrong in relation to this society, and so on. These actions are wrong because the society would be rational to choose the ideal code. There is a corresponding non-constructivist “basic theory” (see Copp 2007a: 18–21). This theory says, roughly, that the “ideal code” for a society is the code the currency of which in the society as the societal moral code would do most to serve the needs and values of the society, given the circumstances, and it says that the content of the ideal code determines which actions are wrong in relation to this society, and so on. The basic theory does not use the idea of rational societal choice and it does not make the explanatory claim that the original theory made. Clearly the difference between a constructivist theory and its non-constructivist sister is subtle, and it is hard to see what philosophical significance it could have in general. In some cases, the constructivist theory might be the more plausible of the pair. A Kantian might contend that only a Kantian constructivist view can provide a plausible account of normativity. In other cases the non-constructivist sibling might be the more plausible. I think Railton’s non-constructivist original theory is more plausible than its constructivist sister and I think the basic society-centered theory is more plausible than its constructivist sister. The non-constructivist sister is in these cases more simple. In general, however, it is hard to see any significance to the difference between constructivist and non-constructivist siblings simply in virtue of the fact that one is constructivist and the other is not. If we begin with a constructivist theory, we can cook up a non-constructivist sister, and if we begin with a non-constructivist theory that proposes substantive truth conditions for moral judgments, we can cook up a constructivist sister. At least we can do this in some cases. In general there seems to be no philosophical mileage to be gained by making either change. 7 Prospects for blockbuster arguments One reason the distinction between constructivism and realism has seemed important is that some philosophers have claimed to have arguments that show either that the truth lies on the realist side or that it lies on the constructivist side. In this section, I briefly consider four examples and I 9781107019218c05_p108-132.indd 128 3/21/2013 4:31:35 PM Is constructivism an alternative to moral realism? 129 speculate about the prospect of success in devising such an argument. If what I have argued so far is correct, I suggest we should be pessimistic. The arguments in the literature presuppose that constructivism is not a kind of moral realism, but this is mistaken or misleading if what I have argued so far is correct. Still the arguments can be reformulated, for there is a distinction between constructivist and non-constructivist forms of realism, and the arguments I am interested in turn on some feature of constructivism that distinguishes it from non-constructivism. The arguments do not aim to show that some specific constructivist or non-constructivist theory is true. Rather, (a) they invoke some feature shared by all versions of constructivism, or some feature shared by all versions of non-constructivist moral realism, and (b) they use this feature in arguing either that no version of non-constructivist moral realism is true or that no version of constructivist moral realism is true. Imagine a logical space consisting of all versions of basic moral realism, and imagine this space as divided into a constructivist region and a non-constructivist region. A blockbuster argument would aim to rule out one of these regions without ruling out the other region and without excluding all but one point in the space. I know of four putative blockbuster arguments. The first three seem to run afoul of either my claim that constructivist theories typically have non-constructivist siblings, or my claim that constructivist theories are most plausibly understood to be versions of basic moral realism. The final argument fails for other reasons. First, Shafer-Landau argues that all constructivist theories are subject to a version of the Euthyphro argument (2003: 42–43). A constructivist theory defines a hypothetical procedure that people could in principle follow where the outcome is a set of moral standards that the theory holds to be true because they are yielded by the procedure. Either the theory says that these standards would be chosen by those following the procedure because they are true, or it says that they would be chosen for some other reason. If it says they would be chosen for some other reason, Shafer-Landau contends, it is implausible that the theory is true. There is no reason to think that the chosen standards will match our “deepest ethical convictions” (2003: 42). But if theory says the standards would be chosen because they are true, then it would be circular to say that the standards are true because they would be chosen. Hence a constructivist theory is either implausible or circular. There are at least two problems with this argument (see Copp 2005). First, there are many constructivist theories, and Shafer-Landau would have to look at them in detail before being entitled to claim that they are 9781107019218c05_p108-132.indd 129 3/21/2013 4:31:35 PM 130 David Copp all either implausible or circular. Shafer-Landau himself suggests that a certain class of constructivist theories, including Kantian theories, arguably can escape his dilemma (2003: 43). A theory in this class would contend that people following its constructive procedure would choose moral standards for some specific non-moral reason. It may be that standards chosen for this reason by people following this procedure plausibly would be true. Second, a constructivist theory (typically) has a non-constructivist sister. If a constructivist theory falls victim to Shafer-Landau’s argument, its non-constructivist sister might or might not likewise fall victim. If it does, then the argument has not isolated an objection to constructivism as such. If it does not, then a version of the original theory can be preserved in a non-constructivist form. For example, the non-constructivist basic society-centered view is not undermined by Shafer-Landau’s argument. Second, Street argues that the fact that human beings are evolved creatures and that our moral psychology has been strongly influenced by evolutionary pressures vitiates all forms of non-constructivist realism (Street 2006). I have responded to her argument elsewhere (Copp 2008). In particular, I contended that the society-centered theory escapes Street’s argument. Street replied that this theory is in fact a constructivist theory (Street 2008b). But although the original society-centered theory is constructivist,26 its sister theory, the basic society-centered theory, is not constructivist. Street’s argument therefore seems to be vitiated by the fact that constructivist theories have non-constructivist sisters. Third, Korsgaard argues against “substantive realism” on the ground that it postulates mysterious “intrinsically normative entities” (Korsgaard 1996b: 34–35). Her objection seems to be that non-constructivist realist theories postulate mysterious intrinsically normative moral properties, whereas constructivist theories explain normativity non-mysteriously on the basis of the nature of the rational will. If I am correct, Korsgaard should admit that rightness and wrongness are properties and that her form of constructivism is best understood as a kind of basic realism. Yet she could still contend that on her theory rightness and wrongness are constructed properties and that this makes their normativity non-mysterious, whereas non-constructivist theories cannot explain normativity. I see at least two problems with the argument. First, it seems to assume that non-constructivist realism must take the form of a non-naturalist theory that does not aim to explain normativity. But nothing prevents 26 It is constructivist on the proceduralist characterization, but not on Street’s characterizations. See above, note 11. 9781107019218c05_p108-132.indd 130 3/21/2013 4:31:35 PM Is constructivism an alternative to moral realism? 131 a non-constructivist theory from attempting to explain normativity, and indeed there are non-constructivist theories that aim to explain normativity, such as Railton’s theory and my basic society-centered theory. Second, the argument is vitiated by the fact that constructivist theories have non-constructivist sisters. Kant’s theory comes in both constructivist and non-constructivist versions and Korsgaard’s own theory has a non-constructivist sibling. It is not obvious why the constructivist version is more successful at explaining normativity than the non-constructivist version. The final argument aims to rule out constructivism on the basis of mind-dependence. It claims that all theories that propose mind-dependent truth conditions for moral judgments are objectionably subjectivist. It seems to me, however, that some theories that propose mind-dependent truth conditions for moral judgments are not objectionably subjectivist. For instance, I do not think that the basic society-centered theory is objectionably subjectivist even though the truth conditions it proposes for moral judgments are mind-dependent. Moreover, it seems to me that, on any plausible account, the truth conditions of moral judgments will be mind-dependent, for it seems to me that the facts about what morality demands of us depend crucially on our nature as sentient, reflective beings. This, however, is an issue I cannot address adequately here. Despite what I have argued, someone might discover a successful blockbuster argument, but I am pessimistic. Of course, I do not claim to have shown that blockbuster arguments are impossible. I have merely reviewed four examples. 8 Conclusion I have argued that meta-ethics should avoid being distracted by a putative distinction between meta-ethical constructivism and meta-ethical realism. On a sensible construal of what meta-ethical realism consists in, constructivism is compatible with realism, for a constructivist could accept all five doctrines of basic realism. Constructivists deny the stance-independence thesis, but so do some non-constructivist realist theories. Debates as to whether meta-ethical constructivism is more plausible than realism presuppose, incorrectly, that a choice must be made between constructivism and realism. I have argued in addition that the distinction between meta-ethical constructivism and non-constructivism is of doubtful philosophical significance. For it appears that many if not all constructivist theories have 9781107019218c05_p108-132.indd 131 3/21/2013 4:31:35 PM 132 David Copp non-constructivist siblings, where the difference between the siblings is subtle and, in many cases, insubstantial. In some cases the non-constructivist sibling seems the more plausible of the two because it is the more simple theory. In other cases the constructivist sibling might be the more plausible. For instance, some might contend that the constructivist sibling provides the more plausible account of normativity. In general, however, we should avoid becoming distracted by fruitless arguments as to whether meta-ethical constructivism is more plausible than non-constructivism. We need instead to focus on the substantive issues, and there are two in the vicinity. The first is about the nature of normativity. Some of the interesting theories that aim to explain normativity are constructivist, but some are not constructivist. Our focus should be on whether a theory can provide a plausible explanation of the nature of normativity, not on whether the theory is constructivist. The second issue is about cognitivism and truth. All interesting forms of constructivism are Success theories that aim to provide substantive Mind-Dependent accounts of the truth conditions of moral judgments. The important distinction is between theories of this kind – whether constructivist or not – and other kinds of Success theories. Given these considerations, then, I conclude that the distinction between constructivism and non-constructivism is philosophically uninteresting. The important debates are between theories that provide substantive accounts of normativity and those that do not, and between Success theories that provide Mind-Dependent accounts of the truth conditions of moral judgments and those that do not. Meta-ethics should focus on these debates and not be distracted by a debate about constructivism. 9781107019218c05_p108-132.indd 132 3/21/2013 4:31:35 PM