I’ve been stalled for some time now in my attempt to write a review of Thomas Nagel’s Mind and Cosmos. My primary stumbling block has been his reliance in one section on Sharon Street’s “A Darwinian Dilemma for Realist Theories of Value”, which attempts to show that natural selection (in its current form) is not compatible with moral realism. Where Street takes this incompatibility as a reason to reject moral realism, Nagel takes it as a reason to reject natural selection (unmodified by teleology).
I don’t think that Street’s argument is as strong as Nagel thinks (despite the fact that I find the constructivist alternative to which she refers more attractive than moral realism). In what follows I say why – first outlining Street’s argument and then my objections. This piece is longer and more technical than the typical PEL post (and also excruciatingly analytical), but I thought it was worth airing my objections now that I’ve thought them through. (I haven’t conducted research to see if others have made similar objections in academic publications – although we can assume that this is probably the case).
Street argues that the content of our evaluative judgments (“x is good”) has been shaped by the “relentless selective pressure” on the basic evaluative tendencies of our pre-linguistic ancestors. These ancestors could not, without language, make explicit evaluative judgments, but they had unreflective experiences of one thing as “counting in favor of another” that form the basis of our more sophisticated evaluative judgments. This is not to deny that there are other factors influencing the content of evaluative judgments (including culture and evolutionary factors other than natural selection, such as genetic drift). It’s just that “our system of evaluative judgments is thoroughly saturated with evolutionary influence.”
Natural selection must be viewed as having a “purely distorting” influence on evaluative judgments, in the sense that it selects the underlying evaluative tendencies based on whether they are adaptive, not on whether their corresponding value judgments are true or false. This differs from the case of basic perceptual judgments, for which accuracy is clearly adaptive. For instance, the accuracy of my judgment that there is a fire nearby might be important to my survival, and so adaptive. But not so the accuracy of my judgment that caring for offspring is good. The tendency to take care of offspring itself is obviously adaptive, but not the accuracy of any judgment related to this behavior. If we attempt to defend the claim that the ability to accurately recognize the goodness of this behavior is adaptive, we merely add an unnecessary layer of explanation with little explanatory power: we cannot say why accuracy of this judgment is adaptive, whereas it is clear why the tendency itself is. So for instance, if it were actually bad to take care of one’s offspring, the inaccuracy of our judgment that it is good would have no adaptive implications; whereas the disappearance of the tendency to take care of offspring would clearly have such implications. In light of all this, it would be an incredible coincidence if all our adaptive tendencies had been fashioned by natural selection to form the basis of accurate judgments of real goodness and badness.
One might object that it could be the case that the ability to grasp evaluative truths is a non-adaptive byproduct of a capacity that is explained by natural selection, in the same way that our ability to do astrophysics is a non-adaptive byproduct of our broader, cognitive capacities. It is important that this more basic capacity itself be non-evaluative: otherwise it falls prey to the problems outlined above, and infects its byproduct with them. If the ability to do astrophysics were a byproduct of basic tendencies to take care of one’s offspring and evaluate this behavior as “good,” it could not borrow any claim to truth from basic non-evaluative judgments in which adaptiveness and truth are non-coincidentally aligned.
But if this more basic ability is non-evaluative, then it is functionally unrelated to our higher level evaluative abilities, and so according to Street only coincidentally related to them. This is untenable, because our evaluative abilities are “highly specialized” and must be “specifically attuned” to certain evaluative truths. Similarly, the human eye is too complex and functionally specified to have been a byproduct of some unrelated capacity (even if the evolutionary rudiments of the eye were in fact byproducts).
Street concludes that moral realism is incompatible with natural selection. This is not to say that we cannot show certain evaluative claims to be true or false: it’s just that their truth or falsity must be functions of our actual evaluative attitudes and cognitive constitution. This anti-realist view can be cached out in various ways, including a constructivist account in which the truth of an evaluative judgment depends on whether one would make it in a state of reflective equilibrium (in which all of one’s basic beliefs have been adjusted to the point of maximum reconciliation). Evaluative truth would then be a function of “how all the evaluative judgments that selective pressures (along with all kinds of other causes) have imparted to us stand up to scrutiny in terms of each other ....”
While I find a constructivist account of morality is attractive, I think Street has failed to show that moral realism is inconsistent with natural selection. And that failure is related to her attempt to answer the objection that our ability to make evaluative judgments is a byproduct of a more basic, non-evaluative capacity. That’s because it’s not clear that our cognitive and evaluative abilities are – as Street suggests – as functionally specialized as the human eye. For instance, it is very likely not the case, contrary to the speculative suggestions of some evolutionary psychologists, that there is a specific cognitive trait amounting to a tendency to rape, for which there are selective pressures. Such notions are related to the modularity of the mind thesis, which argues that the mind consists at least in part of modules that have specialized functions determined by natural selection; and massive modularity, which argues that all mental functions, even higher level cognitive abilities such as planning and problem solving, are modular. The modularity thesis is not uncontroversial, and has its fair share of critics; and massive modularity is even more problematic.
Arguably, the brain has evolved to be an organ that is itself highly adaptive to its environment, and as such to provide us with capacities that are very un-specialized and “domain-general.” These include capacities for planning and problem-solving, for which our higher level abilities to reason and engage in activities like mathematics and philosophy seem to be byproducts. It seems possible that the ability to make evaluative judgments is also in part a byproduct of such general cognitive capacities (and one intimately related to the ability to reason). The moral realist claims that evaluative judgments can be true or false. As such, a capacity that is generally sensitive to truth and falsity might allow us to gauge the truth or falsity of evaluative judgments – or any domain whatsoever where truth and falsity are relevant concepts – regardless of whether the capacity has been specifically tailored to such judgments by evolutionary pressures. The sciences reliably take us into such novel domains all the time. Street argues that this is so because truth and adaptiveness line up with each other when it comes to basic judgments that allow us to “model” the world, effectively acknowledging that our cognitive capacities are not functionally specified to the finest level of detail. But for evaluative judgments, she demands an arbitrary level of functional specificity; and draws very narrow criteria for the relatedness of the ability to make true evaluative judgments and any underlying capacity for which it is a byproduct (if the underlying capacity is non-evaluative, then according to Street the capacities are not sufficiently related and the coincidence objection kicks in). But I don’t think the demand for this level of functional specificity is consistent with what we know about the domain-general capacities of the mind. Street acknowledges that it is common mistake of evolutionary theorizing to assume that “every observable trait (whether cognitive or physical) is an adaptation resulting from natural selection, as opposed to the result of any number of other complex (non-selective or only partially selective) processes that could have produced it.” But I think she falls into a similar pitfall in this paper by assuming that cognitive traits must have a certain level of functional specificity.
If there is a capacity to make veridical evaluative judgments that is a byproduct of more general cognitive capacities, it’s nevertheless clear that in many cases these judgments will also be related to the selection-pressured evaluative tendencies that Street describes. The point is that it’s possible that there is an additional ability to reflect upon such tendencies that goes beyond reconciling them in a state of reflective equilibrium, and allows us to critique them in terms of what is really “good” and “bad.”
This is not to say that natural selection does not pose a challenge to moral realism. Street’s coincidence objection will kick in again unless the moral realist can either a) show there are at least some evaluative judgments which are not simply the result of more basic evaluative tendencies that have been shaped by evolutionary pressures (or better, are inconsistent with an evaluative judgment under reflective equilibrium that takes into account all tendencies but falls short by virtue of some form of moral reasoning that only the realist can supply); or b) show why tendencies that are clearly the result of evolutionary pressures so neatly line up with the results of a capacity for evaluative judgment that is supposed to be unrelated to such tendencies (what Street calls “tracking”). For (a) to be the true, it cannot be the case that our system of values cannot be as thoroughly “saturated” with the influence of natural selection as Street thinks it is. One option for (b) is to argue that adaptiveness and what is “good” are systematically related in such a way that selective pressures will tend to produce a tendency to true evaluative judgments. After all, what is adaptive is arguably a species of the good (although it’s possible that this line of thought leads us back to a constructivist account by relativizing the good to the constitutions or organisms).
While it’s unclear whether the moral realist can accomplish (a) or (b), I don’t think Street has shown that they cannot. And while I can’t say I find moral realism more attractive than the constructivist, anti-realist account, I don’t believe Street has successfully shown that it is inconsistent with natural selection.
-- Wes Alwan