1. ASSESSING GALILEAN EXPLANATIONS

     Let's begin by establishing a context for my criticism of Gibbard's norm-expressivism. There are three items to dispose of at the outset. First, we need to to consider what we might want from a specific sort of Galilean project: an investigation into metaethics. Second, we will examine some standards by which Galilean explanations in general might be judged. Third, in the section following this one, we must explore a cultural source of error that is unfortunately endemic among self-proclaimed Galileans. All this may seem to be excessive preamble, but I believe circumstances require it.

1.1 Explanatory Goals Concerning Metaethics

     Though Gibbard would like his analysis to "help us say what sorts of things really are rational or irrational, right or wrong" (9), he notes that his explanation illuminates our normative practices, but perhaps not the merit of our norms (249): "in strict logic, the facts of our normative lives leave unsettled the normative questions themselves." The famous is-ought gap remains (275): "Does the fact that we do feel some way support norms that say to feel that way?"

     The most valuable goal for a Galilean reduction in metaethics would be to show that the naturalistic fallacy, which holds as mistaken any attempt to derive ought from is , is not a fallacy after all. The fact that we may tend to be fair to others indeed has no bearing on whether we should be fair to others; but there may be facts about the nature of human beings that do directly imply that we should be fair. Instead of attempting to justify moral claims concerning immediate circumstances by appealing to surface facts about those immediate circumstances, a Galilean will need to look much deeper into the causal matrix that made current circumstances what they are.

     Daniel Dennett (1995, p. 468) explains that the problem with the naturalistic fallacy is not that it is an exercise in reductionism, but rather that it is "greedy reductionism" -- a simple-minded attempt to rush from facts to values, rather than a circumspect attempt to "unify our world-view so that our ethical principles don't clash irrationally with the way the world is." We must go further. Our ethical principles require justification, and the causal order of the universe is the only place I have hope of finding it.

     I say the resolution of the naturalistic fallacy is the most important Galilean goal because, if good and bad turn out to be matters of fact, then the way is open to what I call the 'Unholy Grail' of philosophy (because it seems to be regarded more as a threat than a prize), a naturalized ethics.

     A good Galilean reduction of metaethics would also explain and justify our apparently accurate moral judgments. We would understand how our normative practices evolved, and how moral judgments count as correct or mistaken.

     It would also be very helpful to explain, in Galilean terms, the concepts value, virtue, duty, and utility -- along with how they relate to one another and why each has been a focus for systems of ethics. We might discover that one of these concepts represents some fundamental property, defined in purely physical terms, while the others correspond to very common cognitive strategies (in the biological sense of 'evolved systematic responses' rather than 'intentional plans') humans and their ancestors have developed for dealing aptly with that fundamental property.

     Our Galilean reduction in metaethics will need to show a possible and defensibly likely evolutionary path through Design Space for our normative propensities. Daniel Dennett (1995, p. 111) introduces what he calls the Library of Mendel, each of whose volumes is a description of a possible sequence of nucleotides, and which contains all possible volumes of this kind. Most of the volumes will be nonsense, but some will represent the DNA of organisms that could actually live (given the right circumstances). Design Space is a vast, multidimensional theoretical space representing all traits that organisms could, even fancifully speaking, possess. Living creatures, past, present, and future, are part of the Tree of Life that extends through design space, branching this way and that, but leaving nearly all of Design Space as a barren landscape of unfulfilled designs.

     Dennett spends chapters discussing how there may or may not be any path through Design Space from an actual creature to some later creature we imagine could have descended from it. For a Galilean explanation of our normative propensities to work, it must locate somewhere along a path of real possibility (and real probability) in Design Space each of the essential traits upon which those propensities are supposed to depend, possessed by viable organisms at all stages along that path.

1.2 General Standards

     Galilean explanations are expected to cohere with the core of Galilean metaphysical beliefs about the world and how it works. They also entail predictions about our observations of the phenomena they purport to explain, and these predictions are expected to correspond to the observations we actually make. A Galilean explanation is successful, then, if it meets the requirements of coherence and correspondence. It may be unsuccessful, yet correct, if problems with coherence or correspondence are the fault of other parts of the web of one's beliefs -- and may later become successful if an explanation of this failure corrects the flaw and restores coherence and correspondence.

     Another important concern is whether an explanation is fully Galilean: whether it explains the phenomena in question without leaving gaps -- theoretical steps for which no physical mechanism is identified and described, or, at the very least, specifically posited.

     Now, the Galilean core has plenty of gaps -- in the first place, because not all the questions of physics itself have yet been settled. Beyond that, there are honest differences of opinion as to how physical explanations of complex phenomena should go, despite agreement that such explanations are available. Such gaps need not be fatal for the Galilean core, provided the existence of a physical mechanism to close the gap is both expected and plausible. We might provisionally accept a gap-ridden theory under these conditions, until either the gap is closed or a better explanation surfaces. But such incomplete explanations should be carefully and explicitly treated as provisional, lest we forget about important gaps in our theoretical system.

     The reader may have noticed that Galilean explanations have strong parts and weak parts: a good, functional explanation of how something came about is the goal, but sometimes we must settle for the weaker explanation that something arose from physical processes, without much elaboration as to what those processes were. "That" explanations just serve to bridge, rather than close, gaps in a Galilean theory. Only "how" explanations do truly explanatory work; "that" explanations are promissory notes that "how" explanations will ultimately be available. Because of this, we need to assess (once we get, at long last, to a synopsis of Gibbard's book) how frequently Gibbard offers "that" explanations instead of "how" explanations.

     Most importantly, for my purposes: if Gibbard relies on anything besides physical explanations at any point (and I will contend he does), his project is not Galilean. His conclusions could be correct, but he would be wrong concerning why they are correct. If he misunderstands his success as being due to Galilean principles when it is not, his argument is flawed. And if his argument is flawed, then we need other reasons beyond those he provides to accept his conclusions.

     In assessing a theory like Gibbard's, another applicable concern arises when a competing explanation is proffered: Which explains the most? The most comprehensive contender (it is generally accepted), all other things being equal, will be the best choice.

     I contend that Gibbard's account of human normative practices fails to satisfy any of these criteria very well. Specifically, the speculative accounts of psychology and evolutionary development Gibbard relies upon have coherence and correspondence problems -- and both are eclipsed by superior explanations from neuropsychologists, evolutionary psychologists, and biologists. Furthermore, his account has serious explanatory gaps that go unacknowledged. The psychological mechanisms purported to span those gaps are presented without any discussion of whether they can be explained in terms of physical processes (or whether they are explainable at all).

     Finally, Gibbard's explanation doesn't elucidate nearly as much as does my competing view; it offers, at best, a partial account of a somewhat proximate cause of human normative behavior, when an evolutionary account of the ultimate cause is readily available (and much more illuminating).