5. METAETHICAL ASPIRATIONS

     In his attempt to shed light on the origins and nature of human normative practices, we need to ask, did Gibbard succeed? Did he accomplish what he had hoped? Was his goal sufficiently illuminating to be helpful if he attained it? Did he identify and address the most important concerns? My assessment finds his project lacking on all these counts.

     Early on, I asked what we might hope to get from a metaethical investigation such as Gibbard undertakes. My wish list may not have been definitive, and readers might propose additions to or deletions from it. Clearly, not all my goals were also Gibbard's. I hope it will nonetheless be interesting to contrast my idealized ambitions for metaethical inquiry with what Gibbard has to offer us.

5.1 Are There Normative Facts?

     I wished for a way to explain away the naturalistic fallacy, showing how ought could be reduced to a complex is (apparently undeniable putative) fact about the nature and origins of the circumstances in question. Gibbard, however, rejects normative facts. We need to examine whether that rejection is warranted.

     In response to critics of his book, Gibbard (1993, p. 315) explains a key tenet of norm expressivism. "It's a chief aim of my book to argue that if we didn't believe in normative facts, we could coherently go on talking and thinking, normatively, pretty much as we do talk and think." I need to point out that if our ordinary normative thought and discussion are inadequate for modern human circumstances (which is quite likely the case), it's time we understood that and found better methods. We might begin by reconsidering normative facts.

     Gibbard provides a very useful and telling comparison between the ways he views beliefs about facts and beliefs that are normative (122):

     "There is a realm of atomic facts, and the beliefs of atomic physicists artificially represent facts in this realm. We cannot fully explain why atomic physicists have the beliefs they do without citing facts about atoms. No corresponding realm of normative facts need appear in a full explanation of why we have the normative beliefs we do."

     Gibbard thinks the development of normative discussion and the resulting adoption of normative governance belong to a realm of biological facts which explain our normative beliefs without any need to cite normative facts. I disagree. Not only is such an explanation incomplete because of its failure to consider evolved mechanisms for evaluation and judging, but Gibbard crossed his fact/value boundary (one I don't acknowledge, but one he's committed to) at the outset, immediately upon attempting an evolutionary explanation for our normative practices. After all, adaptive value is an indispensable element of such an explanation. New traits, or changes in old ones, do not persist or spread in a population unless they are good for those who possess them.

     I realize there is some conceptual distance between good for x and the good. I will attempt to bridge the chasm as we proceed. But I don't ever expect to encounter a sensible claim that adaptive value isn't both the sort of biological fact Gibbard mentions and yet normative in nature. Even by Gibbard's norm-expressivist lights, concern for adaptive value has to 'make sense'! That is why I contend that Gibbard, and everyone else who cares about Galilean explanation, should acknowledge that there are normative facts: apparently undeniable (though putative) facts about what I call 'objective values' that exist as values only because they are good for some x. I will continue to develop this idea as we proceed.

5.2 Premature Dismissal of Natural Representation

     Gibbard doesn't expect that normative facts will turn out to be complex material facts (the view I will present here), so he views them as inappropriate considerata for Galilean explanation. This is reasonable enough, but one might wonder why a Galilean wouldn't expect normative facts to be a variety of physical fact; after all, the potential theoretical reduction to physical facts is the default assumption of the Galilean stance. Indeed, no evolutionary explanation can fail to rely upon at least one sort of fact that is both material and normative in nature: adaptability.

     The primary reason Gibbard rejects normative facts is that if they existed (he claims) they would be the sort of thing represented (either naturally or artificially) by the content of our normative judgments. He can't find an evolutionary account that makes sense of this view.

     There are two immediate problems with this. First, it unreasonably demands success of an evolutionary attempt to develop a method of representing normative facts. This strikes me as something like denying that light could exist if no creature had ever evolved a mechanism for visual representation.

     The second difficulty lies with the relation between judgment content and fact. Gibbard's real point here, one might think, is not that we must insist on the content of judgments representing facts in the world (charitably reconstructed as 'representing substantive facts expressed by propositions identical to the content of those judgments' -- a reconstruction I'll continue to presume implicitly for economy's sake), but that we explain the origin of such judgments with respect to their content. In fact, the reason that the content of judgments representing material facts might seem to be important is the explanatory power of our theories about how psychic representations evolved for our ancestors (and how they function in us).

     After all, our perceptual judgments result from an evolved response strategy we might call homomorphic representation -- the creation of images in perceptual space, whose features map onto features of material objects, and to which we respond as though they were material objects (and perceptual space were physical space). It should be no surprise that the content of our perceptual judgments naturally represents material facts, or that the same proposition can at once assert both a material (substantive) fact and a perceptual judgment (a putative fact). We reach these judgments by way of senses that are transparent⁶: they reproduce, approximate, or mimic with considerable fidelity the very properties they represent.

     That isn't Gibbard's point, however. He rejects (186-188) analyses treating normative judgments as being like sensory judgments of some secondary quality. "All I seriously claim when I go from 'It looks blue' to 'It is blue' is that my color vision is normal and conditions are normal." However, "I cannot explain trust in my normal feelings [e.g., of anger or guilt] by a straight appeal to meaning, by saying that all my judgments claim for those feelings is that the feelings are normal."

     Not all our judgments are produced by so transparent a mechanism as vision. Do judgments about our appetites -- say, about hunger or lust -- naturally represent their content in such a way that we respond directly to it? No. When we see a tree in our path, we walk around it. But when we feel hunger, we ignore it as best we can and look for food. Appetitive judgments like this are opaque, hiding the real object of significance (in this case, not even the food we seek, but rather the survival it promotes) behind the focus of our experience. Indeed, we might say that these opaque secondary qualities (such as hunger) function much like judgments about tertiary qualities (such as a nutrition deficit). Clearly, a Galilean account of a judgment that I'm hungry will explain it as naturally representing a nutrition deficit (or better, a general fact about the evolutionary history of dealing with nutrition deficits), and not as naturally representing its content.

     To be sure, 'I'm hungry' does express a judgment that represents its content -- but how? Perhaps we should say it semantically or expressively represents its content, being a straight-forward expression of it. As Galileans, though, should we prefer to treat appetites as judgments about secondary qualities -- qualia we experience -- because they then represent their content as Gibbard would like, or should we prefer to treat them as misdirected judgments about tertiary qualities -- judgments that seem to us to be about secondary qualities, but naturally represent significant biological facts? The latter, I think.

     Might normative judgments, then, like appetites, represent something more complex and opaque than immediately discernible qualities of some material state of affairs? Gibbard briefly acknowledges such a possibility (119):

Normative judgments may naturally represent something, but if there is any class of fact that they naturally represent in virtue of being normative, it is a complex kind of fact about the observer and his community. These facts, though, are never the content of a normative judgment.

     Yes, it turns out that the content of a normative judgment is the crucial issue for Gibbard. It's how we tell factual from normative beliefs (100):

[P]sychologically, normative beliefs are much like any other beliefs. The fact-norm distinction is not between two quite disparate kinds of psychological states but between the kinds of content that can reasonably be ascribed to them.

     This statement flies in the face of settled neuropsychology. It's a fine example of the monolithic conception of judgment I have mentioned, presuming that the same functional mechanism can produce judgments that are evaluative, judgments that are intuitive or heuristic, and judgments that are methodical and deliberative.

     Indeed, it's not uncommon for a scholar who is intimate with the theory of evolution to nonetheless treat all forms of judgment as varieties of inference. Stephen Stich (1994, p. 347) provides an example: "Natural selection might perfectly well opt for an inferential strategy which produces false beliefs more often than true ones." This is true enough -- importantly so; but Stich has been writing about illogical aversions to food experienced by rats. I have to say that rats very likely do not infer (at least not in the ordinary sense that is taken to be distinct from simple association -- though we should avoid the essentialist assumption that inference differs in kind from simple association, for I think the former has developed in stages from the latter), but they surely evaluate -- and an evaluational strategy (instead of an inferential one) explains nicely the way they respond to food associated with physical discomfort.

     I've been referring often to large-scale functional divisions of the brain, each of which is specialized for a different variety of judgment, and in just a moment (as we consider the origins of normative governance) I'll provide some authoritative support. For now, I want to make a different point.

     Gibbard thinks a psycho-social explanation of human normative practices is what we need. In response to his critics (1993, pp. 317-318) he takes a state of mind ("accepting a normative directive that applies to oneself right now") as explanatorily basic. But that state of mind would appear to require additional explanation in Galilean terms.

     There's an even more telling passage in his book (121): "even if there were a special realm of normative facts, it would be gratuitous to think that normative beliefs represent facts in this realm, even artificially. If the speculative evolutionary story I have told is correct, then we have another explanation of these beliefs."

     Gibbard's preoccupation with content -- with the mental -- has, it seems, pre-empted his investigation into what complex biosocial facts might be naturally represented by normative judgments, and thus explain them in Galilean terms. The unwieldiness entailed by such complexity might explain (though it can't excuse) the avoidance of more detailed Galilean explanations.

     Philosophy will be best served by extending, everywhere it can be applied, Pierre Duhem's famous observation that huge chunks of theory are smuggled into seemingly pedestrian observations. The implicit beliefs about the efficacy and function of Science's methods and apparatus are merely the tip of the proverbial iceberg -- aspects of our culture that we hadn't always appreciated as influencing our processes of belief formation. Add to that the entire evolutionary history of perceptual and cognitive mechanisms, and we finally have under consideration the pertinent types of theoretical influence ... for epistemology. Metaethics will require a deeper investigation yet.

5.3 Justifying Claims of Accuracy and Objectivity

     The second thing I wished of metaethical inquiry is some way to justify claims that certain moral judgments are accurate. This goes hand-in-hand with the universal claim (as Gibbard rightly notes that humans tend to make) of normative objectivity.

     Despite Gibbard's efforts, this issue remains mysterious. If our belief in normative objectivity persists because it has been adaptive for our ancestors, it can't be because it responded to standards set by nature; that would consist of normative facts, which Gibbard has rejected. Fortunately, he doesn't claim that normative "truth" is attainable by a "committee of the whole", via some exercise of communal judgment. Unfortunately, he offers no explanation of how normative discussion responds to anything (other than the normative facts he rejects) in developing and justifying normative consensus.

     Galileans should not be satisfied with things the way Gibbard has left them. If normative discussion and normative governance are adaptive, we want to know how. The benefits of social harmony and cooperative effort, in themselves, are too close to the surface to be the ultimate benefits they confer on us. The default assumption for Galileans is that they are adaptive for the same reason anything else is: because they promote inclusive fitness (reproductive success over many generations), the coin in which evolutionary gambles pay off.

     From whence might a solution come? As we'll see later on, there's some hope that embracing normative facts, for what I think are good and compelling reasons, will provide exactly the justification we need here.

5.4 Explaining the Conceptual Relations Among Ethical Systems

     How, I next wondered, might a metaethical story make sense of the fact that differing systems of ethics have emerged, each focusing on a foundational concept: e.g., 'value', 'utility', 'virtue', or 'duty'. What evolutionary roles have the qualities represented by these concepts played in our ancestors' development? What, if any, are the relations connecting these concepts in a way that might identify one of them as fundamental and the others as derivative?

     Gibbard didn't leave this question entirely unexplored. He mused about rights and social justification (272-273), about utilitarianism (285), and about how utilitarianism and intuitionism correspond roughly to his two prongs of normative discourse (292). But the sort of explanation I hoped for is lacking in these excerpts; Gibbard's purpose in introducing them was apparently to make contact with competing views and show how the broadly moral concerns and narrower moral sentiments of norm-expressivism might explain the attraction of those views.

     A Galilean might quickly see a way to explain the relations among these concepts. In the smallest nutshell, if there is objective, intrinsic value in nature (say, Life), then utility might be instrumental value promoting nature's intrinsic value; virtue might be an evolved disposition to act in some way that promotes intrinsic value; character might be a settled collection of dispositions (including virtues, we hope); and duty might be an obligation to act in ways one expects will promote intrinsic value.

5.5 You Can't Get Here from There

     An important benefit I wanted from a Galilean inquiry into metaethics is to find a possible and defensibly likely path through Design Space for our normative propensities. It won't do if the alleged history of these propensities includes impossible ancestors or unbridgeable evolutionary gaps. I also noted that a telling sign of anti-Galilean influence is the presence of surprising (given the author's outstanding acumen) errors of evolutionary theory.

     There is some evidence that Gibbard's account of our evolutionary history suffers from common misapprehensions. He mentions that little genetic change is thought to have occurred in humans over the last several hundred generations (27). Perhaps he was merely being conservative, but genetically modern humans (Homo sapiens ) have been around for much longer than that. If a generation is taken to be 20 years, he is talking about a period corresponding roughly to recorded human history -- not the period nearly two orders of magnitude greater in duration for which no significant anatomical difference from today's humans is noted in the fossils of our ancestors⁷.

     Now, it is possible that some important new, genetically determined behaviors have emerged in the more recent generations Gibbard mentions. The problem with any claim that they did is explaining how that could happen without generating noticeable changes in brain structure (as revealed by the interior features of fossil skulls) or, say, vocal tract structure (revealed by the relative position and articulation of cervical vertebrae and skull and points of attachment for the relevant musculature). The lack of such evidence indicates that humans have been fully capable of language (and emergent verbally developed and transmitted culture) for several hundred thousand years.

     It is also quite possible that cultural evolution over recent generations accounts for the development of a "linguistically infused normative control system" for humans -- but cultural evolution doesn't respond to the same selection pressures as does genetic evolution. A brain with general purpose cognitive and linguistic capacities can harbor any number of cultural systems, and selection operates at the level of cognitive acceptance of the system -- not at the level of genetic adaptation, the level Gibbard's theory depends upon (for genetic change is what provides a heritable difference of behavior for subsequent generations -- genetic adaptation).

     It appears Gibbard has relied upon effects of genetic evolution that simply have not (and probably could not have, according to our best theories about hominid environments) occurred in anything like several hundred generations. His theory relies on the notion that the benefits of coordinated behavior, thanks to linguistically infused norms, could have produced genetic effects in comparatively few generations. The theory fails for lack of evolutionary time to produce the results claimed -- for lack of a plausible path through Design Space.

     Gibbard's claim, however, that the benefits of coordination afforded by normative discussion account for the development of our present normative control system, might be rescued by appealing to cultural rather than genetic evolution as the driving mechanism. Gibbard's main point is that our normative practices were adaptive -- not necessarily that the adaptivity was genetic rather than cultural in nature.

     Perhaps Gibbard would be safe in claiming that our moral normative practices are something like our epistemic normative practices, which developed with language and persist because of the advantages they bring us: effective and reliable thought, inference, and communication.

     Not even that adjustment will help. It is doubtful that the advent of language among primates has much altered the broader benefits of coordination -- at least with respect to questions of morality (why and how to coordinate with others; clearly the range of opportunities for coordination has increased). More likely, humans inherited from their ancestors a well established normative control system that had governed primates for some millions of years, and language provided new avenues for this system to operate. We will examine this possibility now as we consider the origins of normative governance.