Review of Michael Martin’s:
Atheism, Morality, and Meaning (2003)
(Buffalo: Prometheus Books, 2002, 330 pages)
Jeffery Jay Lowder
What is the relationship between God and morality? Is morality dependent in some way on the existence of God? Arguments for an affirmative answer to that question may be classified into one of three categories: ontological, epistemological, and prudential. According to ontological moral arguments, the existence of moral facts or properties implies that they are somehow dependent on God’s existence. For example, perhaps nothing would be obligatory unless God commanded it. Epistemological moral arguments are concerned with whether moral facts or properties are knowable. According to such arguments, at least some moral knowledge requires God. Finally, prudential moral arguments attempt to show that unless God exists, morality would not have a motivational basis.
In the last thirty years, there has been a resurgence of moral apologetics, with theists of various stripes offering increasingly sophisticated defenses of each type of moral argument for God’s existence. Arguably, Robert Merrihew Adams deserves the credit for starting this revival with his 1973 article, “A Modified Divine Command Theory of Ethical Wrongness,” and his 1979 follow-up article, “Divine Command Metaethics Modified Again.”[1] In those articles, it became apparent that an honest, professional philosopher–indeed, one who is competent in both the philosophy of religion and metaethics–could defend an original approach to theistic moral philosophy.[2] Adams was far from alone in this respect. Theistic philosophers even created an entire philosophy journal, the Journal of Religious Ethics, dedicated solely to the topic of theistic moral philosophy.
Yet most contemporary naturalists give moral arguments for theism scant attention, believing that the relevance of theism to moral philosophy had been fatally discredited long ago by the likes of Socrates, Hume, and Kant. One welcome exception to this trend is the renowned philosopher Michael Martin. In his recent book, Atheism, Morality, and Meaning, Martin provides an up-to-date, naturalistic assessment of contemporary moral apologetics. But Martin does not just defend atheism against moral apologetics; he goes on the offensive, presenting an affirmative case for an atheistic moral realism. He also makes many interesting points about the idea that Christianity (and not just God in general) is the foundation of ethics, and offers a detailed discussion of the implications of Christianity and naturalism for the meaning of life.
The breadth of the territory covered by Martin is impressive; I know of no other book on the market that covers such a variety of metaethical issues from an overtly atheistic perspective. Anyone interested in the relationship between naturalism, theism, and morality will find Martin’s book useful. Moreover, unlike many books on metaethics, Martin’s book is not highly technical, which helps to make it accessible to the lay reader.
However, the book does have its limitations. In this review, I will briefly highlight what I consider the most significant shortcomings.
Atheism and the Motivation for Being Moral
According to one objection to atheistic ethics, atheism provides no assurance that justice will triumph over evil, and thereby undercuts one’s motivation to behave morally. In reply to this objection, Martin asks, “But why demand certainty? Why is it not enough that the triumph of justice is possible?” (p. 33). Clearly, one need not be certain of the desired outcome of an action in order to be motivated to perform it. But the issue is one of probability, not certainty. If the desired outcome of an action is unlikely, then one may well be unmotivated to perform it. Indeed, in the event that the desired outcome is highly unlikely, one may even be highly motivated not to perform it!
It is unclear whether Martin believes the triumph of justice is likely, since he does not say. Instead, he considers the possibility that atheists could know with absolute certainty that evil, not justice, will ultimately triumph. He writes:
However, suppose that atheists could know with absolute certainty that evil will triumph in the next millennium. … Why should this mean that people should not be virtuous? Would not courage be essential for a morally dignified survival in such a desperate time? Would not mutual compassion and charity help people cope with the reign of terror? Would not wisdom of how to live in such a hell on Earth be needed? (p. 33, my emphasis)
However, there are two problems with this reply. First, notice that this reply doesn’t address the issue of motivation. For example, a person might recognize that she “should” be virtuous, yet be utterly unmotivated to be virtuous. Indeed, Martin’s own externalism about motivation allows just such a possibility. Second, this response, along with the rest of Martin’s book, neglects a much more interesting philosophical problem, namely, the issue of justification. From the moral point of view, virtue, courage, compassion, charity, and wisdom are morally good. However, the fact that such things are morally good provides no non-question-begging reason for why an amoralist should be moral. This leads to my next complaint about the book.
Atheism and the Justification for Being Moral
Assuming that conflicts between morality and self-interest are possible, one can coherently ask the question, “Why should I be moral, especially when it conflicts with my self-interest?” Those of us who already accept the authority of the moral point of view are tempted to dismiss the question as nonsensical, since from the moral point of view, it is a truism that a person should do what is morally required. However, such a reply would be superficial. The point of asking “Why should I be moral?” is to question the authority of the moral point of view in the first place. If one answers the question by saying, “Because the moral point of view requires that you be moral,” the questioner could simply ask, “Why should I adopt that point of view?” This leads to a very common objection to atheistic ethics. According to the objection, if atheism is true, then moral behavior is not rationally required. Indeed, in the event of a conflict with self-interest, moral behavior may even be positively irrational for an atheist.
Although Martin discusses the related issue of the atheistic account of the motivation for being moral, he does not directly address the atheistic justification for adopting the moral point of view. This is significant, since clearly motivation does not entail justification: even if you strongly motivated to do such-and-such, that says nothing about whether you are right to do so. Unfortunately, this renders Martin’s comments about the former largely irrelevant to the latter. One is still left looking in Martin’s book for an atheistic justification for adopting the moral point of view.
This leads to the follow-up question, “On the assumption that God does not exist (and the assumption that conflicts between morality and self-interest are possible), how likely is it that the demands of morality will converge with self-interest?” Unfortunately, given both assumptions, it seems highly unlikely that the demands of morality will converge with self-interest for everyone all of the time. Why? Because if atheism is true, then metaphysical naturalism is probably true. (Although atheism is logically compatible with the existence of supernatural beings other than God, the prior probability of the supernatural given atheism is low. Metaphysical naturalism has the highest prior probability of all atheistic hypotheses.) If metaphysical naturalism is true, then there is no God and no life after death. And if there is no God and no life after death, then there are cases in which the cost of moral behavior greatly outweighs the benefits. In such cases, why wouldn’t a person be justified in satisfying their own self-interest instead of the demands of morality? Unfortunately, as far as I can see, Martin doesn’t discuss this question.
Moral arguments for God’s existence
Although Martin manages to cover a lot of territory, there are still some significant gaps. Theistic philosophers and apologists are prolific writers; it is probably impossible to adequately address every variation of moral arguments for theism in a mere 300-page book. However, I think it is reasonable to expect such a book to discuss at least the most popular and the most sophisticated of such arguments. Unfortunately, Martin does not consistently do this. Some examples should make this clear.
(i) The “laws imply a lawgiver” argument. I assume the intuitive appeal and relevance of this enormously popular argument is obvious. Yet nowhere in his book does Martin explicitly address this argument. To be sure, Martin has (briefly) addressed the lawgiver argument elsewhere,[3] but the lack of an explicit discussion of that argument in a book-length treatment of ethics without God is odd.
(ii) Robert Adams’s Argument from the Social Nature of Obligation. In an interesting article,[4] Adams argues, with some intuitive support, that the notion of obligation is irreducibly social. By “social,” Adams means that obligation occurs in the context of a relationship between persons in which a demand is made. This attribute of moral obligation makes it different from other moral concepts (e.g., goodness, virtue) that do not seem to have this feature.[5] Although humans could impose moral obligations on one another, humans cannot be the source of all moral obligations. Thus, on Adams’s view, if there are any objective moral obligations, they must be the demands of an unchanging, loving God.
Note that whereas Adams is arguing from moral objectivism to the divine command theory, an atheist could argue from atheism to the denial of moral realism. That is, an atheist might agree that the dual objective and social nature of moral obligation requires a divine obligator, but since there is no such obligator, moral objectivism is false. Such an atheist would be a kind of error theorist: the atheist would agree that the concept of moral obligation is irreducibly social, but deny that there exists any moral obligator suitable to the task of underpinning objective moral obligations.[6]
Although initially it might seem that the Ideal Observer would be a suitable moral obligator, this won’t work. As Martin points out, “An Ideal Observer does not create obligations. Rather, the Ideal Observer analysis tells us what moral obligation means” (p. 87). But as far as theories of meaning of moral obligation are concerned, why should we prefer the Ideal Observer Theory to Adams’s social theory of obligation? Whereas Adams has argued that his theory is superior to the Ideal Observer Theory,[7] Martin does not explicitly discuss the social theory of obligation and it isn’t clear how he would respond to such an argument. In this sense, his discussion is incomplete.
(iii) J.P. Moreland’s Hybrid Moral Argument. J.P. Moreland has formulated an abductive moral argument that is a hybrid of ontological and prudential moral arguments for theism. Combining Robert Adams’s ontological moral argument with George Mavrodes’s normative moral argument, Moreland argues that the combination of the existence of objective moral facts and the falsity of ethical egoism is more probable given theism than given naturalism.[8] When properly understood, Moreland’s argument avoids many of the objections Martin raises against weaker arguments. Consider the ontological part of Moreland’s argument, which claims, “Irreducible, nonnatural value properties do exist and are a part of the furniture of the universe.”[9] In a nutshell, Martin’s objections to such an argument amount to the defense, “But atheism is logically compatible with irreducible, nonnatural value properties.” Yet, as Martin himself knows from his own work on atheological arguments from evil, the fact that A is (merely) logically compatible with B does nothing to undermine the claim that A is improbable given B. In other words, Martin (correctly) refutes what I call the logical version of the ontological moral argument, but he fails to discuss evidential versions.
Likewise, consider the normative part of Moreland’s argument: “We have moral obligations that seem to require duties that often go against our own best interests.”[10] This premise is based upon George Mavrodes’s argument to the queerness of materialistic morality, an argument that is highly influential among philosophical theists. Although Martin is familiar with Mavrodes’s argument, Martin only mentions it in passing in an endnote (p. 46, n.11). Instead, he discusses what he calls the “Argument from Culmination,” an “argument from motivation” that Martin considers a “similar” argument. Yet the Argument from Culmination and Mavrodes’s argument are not similar enough: the former is an argument from motivation while the latter is an argument from justification. Even if Martin’s defeaters to arguments from motivation are entirely successful, those defeaters are not successful as defeaters to arguments from justification, including Mavrodes’s.
Atheism, Moral Realism, and Moral Objectivism
Martin makes several comments about both moral realism and what he calls “objective morality.” Martin defines moral realism as follows:
1. There are moral truths or facts.
2. These truths or facts are independent of the evidence for them.
While his definition of moral realism is acceptable, his definition of objective morality leaves much to be desired. Indeed, one searches his book in vain for a rigorous definition of “objective morality.” Martin offers multiple definitions of objective morality (p. 36), but his preferred definition seems to be the belief “there are moral facts” (p. 28). Unfortunately, this definition does not really capture the intended meaning, for it fails to exclude subjectivism. As Geoffrey Sayre-McCord explains, “subjectivism about value gives perfect sense to there being a fact of the matter (you might even say ‘an objective fact of the matter’) about what is good or valuable.”[11] What distinguishes moral objectivism from subjectivism is not a belief in moral facts, but the belief that “the appropriate truth-conditions [of moral claims] make no reference to anyone’s subjective states or to the capacities, conventions, or practices of any group or people.”[12]
Semantical worries aside, what reasons does Martin offer to believe that moral objectivism is true? Although Martin does address arguments against an atheistic objective morality, he offers little in favor of moral objectivism over and against subjectivism. He hints at one argument for moral realism in his Introduction, where he states, “Ordinary language and common sense assume that morality is objective” (p. 12). However, he neither formulates this as an argument for moral realism nor defends it against well-known criticisms.[13] Indeed, it appears that Martin largely takes the truth of moral realism for granted. In his defense of Ideal Observer Theory, he states that it is “attractive” in part because it “allows for moral facts [and] excludes ethical relativism” (p. 51). But allowing moral facts and excluding ethical relativism is attractive in a metaethical theory only if one has a prior reason to believe there are moral facts. As far as I can tell, Martin does not provide such a reason in his book.
In addition to the “old-school” argument from ordinary language and common sense only hinted at by Martin, there is another, “new-school” approach to defending moral realism. The recent resurgence of Aristotelian ethics is based in large part on recent work showing how facts about evolutionary biology, developmental psychology, social cognitive theory, and philosophy of biology support an Aristotelian ethical naturalism.[14] Unfortunately, Martin does not interact with this literature or the evidence it discusses.
Atheism and the (Firthian) Ideal Observer Theory
Martin is not only a moral realist and a moral objectivist, but he is also an ideal observer theorist. He accepts the Ideal Observer Theory (IOT) of ethics as both a theory of meaning and a theory of justification. As a theory of meaning, the IOT holds that “X is morally wrong” means “If there were an Ideal Observer, it would contemplate X with a feeling of disapproval” (p. 50). As a theory of justification, it is the view that one’s ethical judgment is rationally justified if the judgment is based on “one’s [estimation] of the reaction of an Ideal Observer” (p. 51).
Before I discuss Martin’s defense of the IOT, I first need to make two points of clarification. First, when I refer to the IOT as a theory of justification, this aspect of IOT doesn’t answer (and isn’t intended to answer) the justification issue discussed above. Second, there are many different versions of the IOT;[15] Martin accepts Roderick Firth’s version of the Ideal Observer Theory (IOT). With these clarifications in mind, then, let’s consider Martin’s IOT.
What are the properties of a Firthian Ideal Observer? In a bizarre catalog of pseudodivine attributes, Martin explains:
An Ideal Observer has the following characteristics: omniscience with respect to nonethical facts, omnipercipience (which will be discussed shortly), disinterestedness, dispassionateness, consistency, and normalcy in other respects. (p. 55)
One doesn’t have to be a philosopher in order to recognize that when Martin describes the “Ideal Observer,” many people will think “God.” Indeed, this is precisely what we find in the secondary literature. Many commentators on the IOT in general–and Martin’s acceptance of it in particular–have remarked that the Ideal Observer (IO) sounds like another name for God.[16] In reply to such worries, Martin states, “the Ideal Observer is hypothetical–it does not exist” (p. 86). While this may be the case,[17] there does seem to be a certain oddity about an atheist affirming a moral theory based upon the reactions of a nonexistent person who is omniscient with respect to nonethical facts. Moreover, it seems to me Martin’s reply simply trades one problem for a group of other problems. If the IOT is properly understood as being founded on the feelings of a nonexistent being, then such a view seems dubious. Here are three reasons why.
First, consider moral epistemology. Since the IO does not exist, it is far from clear, on the basis of the Firthian IOT alone, how a Firthian IO would feel about any action. This is because the IOT tells us nothing about which fact or property merited the approval of a Firthian IO. For example, consider the moral claim (MC), “Honesty is morally good.” According to IOT, that claim means, “If there were an IO under ideal conditions, it would contemplate honesty with a feeling of approval.” But how would one know that MC is true, given IOT? Why couldn’t a Firthian IO be omniscient with respect to nonethical facts, omnipercipient, disinterested, dispassionate, consistent, normal in other respects, and yet contemplate honesty with a feeling of indifference? Better yet, why couldn’t an IO contemplate honesty with a feeling of disapproval? If we are worried that God might command cruelty for its own sake, should we not be equally worried that a Firthian IO might contemplate dishonesty with a feeling of approval? In the absence of information about what would influence a Firthian IO’s reactions (i.e., whether the Firthian IO is honest and values honesty), it is unclear how Firthian IOT justifies MC over its denial.[18] Indeed, one is left wondering if proponents of Firthian IOT aren’t just projecting their own moral beliefs onto a hypothetical IO.
Moreover, a disapproval of honesty by a (Firthian) IO would conflict with widely held moral intuitions that honesty is prima facie morally required. While conflict with moral intuition is not fatal for an ethical theory, nevertheless it is problematic for a philosopher like Martin who justifies his belief in moral realism at least in part based on moral intuitions (p. 12).
Similarly, any attempts to estimate the reactions of an IO are equally doubtful. The problem is made worse by the fact that all human beings are “considerably less than ideal” and “rarely come near impartiality.”[19] But if the IOT were the correct theory of ethics, then, since the hypothetical feelings of an IO are unknown, we would have no way to know what is morally right or wrong.
Second, consider the ontology of moral properties. According to the IOT, the reactions of a hypothetical IO provide the metaphysical foundation for moral properties. How, precisely, is IOT supposed to be different from error theory? Isn’t the whole point of error theory to grant that moral statements have truth values, but deny that any moral statements correspond with objects, states, or properties that actually exist? This seems too close for comfort to the antirealism that the IOT was supposed to help us avoid. But suppose that the Ideal Observer theorist argued the IOT is a form of ethical naturalism. Perhaps the ontological foundation for the counterfactual, “If there were an IO, then it would disapprove of X,” is a complex and gerrymandered natural fact F. In that case, it would seem that fact F–not the reactions of a hypothetical IO–provides the ultimate ontological foundation for moral properties. This leads to my final point.
Third, IOT is not a complete metaethical theory, since it does not offer an explanation for moral facts and properties. This can be seen by using a Euthyphro-like dilemma to clarify the IO’s role: is X good because an IO would contemplate it with a feeling of approval, or would the IO contemplate X with a feeling of approval because it is good? If X is good because the IO would contemplate it with a feeling of approval, then goodness is, in an important sense, subjective: goodness would not be a property of actions but instead a property of a mind (albeit a hypothetical one). The IO would be an inventor of moral facts, not just a passive observer of them. If, on the other hand, the IO would contemplate X with a feeling of approval because it is good, then moral goodness is independent of the reactions of a hypothetical IO. The IO would merely discover preexisting moral facts or properties and then approve them. In that case, the IO would seem quite superfluous as an explanation for moral goodness. Some other ethical theory–not the IOT–would have to be introduced to do the real work in explaining moral facts and properties.
Concluding Remarks
Far too many naturalists neglect moral arguments for God’s existence. Atheism, Morality, and Meaning is a welcome counterexample to this trend, not only because Michael Martin challenges moral arguments for theism head-on, but because he does so as a moral realist. Unfortunately, his book is not comprehensive; there are some fairly significant gaps in his discussion of moral realism and of objections to atheistic morality. Moreover, many readers will find his defense of the Ideal Observer Theory counterintuitive and unconvincing. Nevertheless, this book would make a useful addition to the library of anyone interested in the relationship between atheism and morality.[20]
Notes:
[1] “A Modified Divine Command Theory of Ethical Wrongness,” Religion and Morality: A Collection of Essays, edited by G. Outka and J. P. Reeder, Jr. (New York: Doubleday and Company, Inc., 1973), pp. 318-34; “Divine Command Metaethics Modified Again,” Journal of Religious Ethics, Vol. 7, No. 1 (Spring, 1979), pp. 71-79.
[2] Cf. Quentin Smith, “The Metaphilosophy of Naturalism” Philo 4 (2001), pp. 195-215.
[3] See his Atheism: A Philosophical Justification (Philadelphia: Temple University Press, 1990), pp.; The Fernandes-Martin Debate 1997; Michael Martin and Peter Williams, “Head to Head Debate: The Theist God” The Philosophers’ Magazine No. 8 (1999), pp. 19-23.
[4] Robert M. Adams, “Divine Commands and the Social Nature of Obligation” Faith and Philosophy 4 (1987), pp. 262-275; cf. Robert Merrihew Adams, Finite and Infinite Goods: A Framework for Ethics (New York: Oxford University Press, 1999), pp. 245-246.
[5] Mark Murphy, “Theological Voluntarism” Stanford Encyclopedia of Philosophy, 2002.
[6] Cf. a similar point made by Dale Tuggy in “Necessity, Control and the Divine Command Theory,” forthcoming.
[7] Adams 1999, pp. 245-246; cf. Robert Merrihew Adams, “Moral Arguments for Theistic Belief,” in Rationality and Religious Belief (ed. C.F. Delaney, Notre Dame: University of Notre Dame Press, 1979), pp. 121-122.
[8] J.P. Moreland, “Ethics Depend on God” in Does God Exist? The Debate Between Theists & Atheists (Buffalo: Prometheus, 1993), p. 120, n. 1.
[11] Geoffrey Sayre-McCord, “Introduction: The Many Moral Realisms” in Essays on Moral Realism (ed. Geoffrey Sayre-McCord, Ithaca: Cornell University Press, 1988), p. 17.
[12] An Aristotelian naturalist might object that this definition excludes Aristotelian naturalism from moral objectivism. However, at least on Sayre-McCord’s teminology, there is a meaningful distinction between realism and objectivism. Moreover, Sayre-McCord distinguishes between subjectivism, intersubjectivism, and objectivism. On Sayre-McCord’s terminology, Aristotelian naturalism is a realist theory since it allows for moral facts, but not objectivist since it holds that moral facts depend upon facts about people’s subjective states. Instead, Aristotelian naturalism is an intersubjectivist version of moral realism. See Sayre-McCord 1988, pp. 14-15, 19-20; cf. Larry Arnhart, Darwinian Natural Right: The Biological Ethics of Human Nature (Albany: State University of New York Press, 1998), p. 70.
[13] For a defense of the argument that ordinary language and common sense provide some evidence for objective moral facts, see Quentin Smith, Ethical and Religious Thought in Analytic Philosophy of Language (New Haven: Yale University Press, 1998), pp. 170-179.
[14] William D. Casebeer, Natural Ethical Facts: Evolution, Connectionism, and Moral Cognition (MIT Press), forthcoming; Arnhart 1998; and William A. Rottschaefer, The Biology and Psychology of Moral Agency (New York: Cambridge University Press, 1998). Cf. Boyd’s homeostatic property clusters, in Richard N. Boyd, “How to Be a Moral Realist” in Essays on Moral Realism (ed. Geoffrey Sayre-McCord, Ithaca: Cornell University Press, 1988), pp. 181-228.
[15] Charles Taliaferro mentions “work on the notion of an ideal observer by Hume, Schopenhauer, Smith, Fith, and others.” See Charles Taliaferro, “The Divine Command Theory of Ethics and the Ideal Observer,” Sophia 22 (1983), p. 3.
[16] E.g., Roderick Firth, “Ethical Absolutism and the Ideal Observer” Philosophy and Phenomenological Research 12 (1952): 317-345; Taliaferro 1983, pp. 3-9; Paul Copan, “Can Michael Martin Be a Moral Realist? Sic et Non,” Philosophia Christi, Series 2, 1/2 (1999), p. 51, n. 25; Peter Williams, “Is There a Personal God?” Philosophers’ Magazine 8 (Autumn 1999): 22-23.
[17] This may indicate a distinction between Martin’s version of the IOT and Roderick Firth’s. Whereas Martin says the Ideal Observer does not exist, Roderick Firth maintains that the Ideal Observer need not exist. In Firth’s words, ethical “statements are not intended to imply either that there exists, nor that there does not exist, a being who satisfies the description of an ideal observer; they are intended to imply, on the contrary, that the existence or non-existence of such a being is irrelevant to the truth of the statement.” However, it isn’t clear from Martin’s remarks whether he considers the nonexistence of the IO essential to his version of the IOT. Perhaps Martin agrees with Firth that the IOT is silent on whether the IO exists and holds, on the basis of other evidence, that there is no IO. However, if that is the case, then Martin is silent on why he believes the IO does not exist. See Firth 1952, p. 323.
[18] Cf. Taliaferro’s defense of IOT against Adams’s divine command theory of ethics, where Taliaferro modifies Firthian IOT by “adding that the ideal observer is loving and replac[ing] the notion of being dispassionate and disinterested with the stipulation that the observer is impartial.” See Taliaferro 1983, p. 3.
[19] Richard Garner, Beyond Morality (Philadelphia: Temple University Press, 1994), p. 62.
[20] I am grateful to Glenn Branch for extensive feedback that made this review much better than it would have been otherwise. I also thank Evan Fales for providing constructive criticism of a previous draft of this review.