In a lucid and engaging piece, Tim Lewens (this issue) argues that debates over the enhancement of human nature are often predicated on a fundamental “essentialistic” error—namely, the obsolescent Aristotelian notion that biological species are abstract classes with essences that explain their patterns of populational variation and the intrinsic properties of their individual organisms. The anti-essentialist consensus in biology and the philosophy thereof is that particular species are not atemporal natural kind classes with organisms as their members, but rather spatiotemporally restricted, weakly cohesive, evolving individuals (metapopulations) with organisms as their constituent parts.Footnote 1 Lewens is right to suggest that the ethical defense of human nature need not subscribe to a pre-Darwinian conception of species and speciel natures: asking if it is ethical to enhance human nature may not be as metaphysically misguided as asking whether it is ethical to enhance unicorns or other nonexistent entities. Yet, as his discussion of the neo-Aristotelian positions of Thompson, Kass, and others brings out, it is difficult to square many mainstream ethical positions vis-à-vis species' natures, infused as they often are with morally loaded conceptions of “the natural,” with the nonnormative “population thinking” that is characteristic of the modern evolutionary biological worldview.

Any serious ethical discussion of the enhancement of human nature must begin with a reasonably accurate picture of the causal-historical structure of the living world. In this Comment, I show that even biologically sophisticated ethical discussions of the biomedical enhancement of species and speciel natures are susceptible to the kind of essentialistic thinking that Lewens cautions against. Furthermore, I argue that the same evolutionary and developmental considerations that compel Lewens to reject more plausible conceptions of human nature pose equally serious problems for some prominent critiques of biomedical enhancement that presuppose the existence of a “given” biological potential that can be distorted by agentic cultural influences.

1 The Psychological Pull of Species Essentialism

The psychological pull of species essentialism is not limited to ethicists with little exposure to the theoretical and empirical foundations of the life sciences. Even among more biologically savvy and thoroughly naturalistic discussions of the biomedical alteration of human nature, authors have a tendency to gravitate toward essentialistic conceptions of species. To take a recent example, Nick Agar in Humanity's End: Why We Should Reject Radical Enhancement (2010) argues that radical biomedical enhancements, or modifications that alter human capacities beyond the existing normal range, are likely to result in beings who, by virtue of these modifications, “exit [the human] species and either become a member of a different biological species or a member of no species at all” (p. 32). Elsewhere in the same work, Agar argues that transgenesis carried out in the laboratory would not blur natural species boundaries any more than the intentional physical transformation of one atomic element (e.g., uranium) into another (e.g., plutonium) would blur natural atomic boundaries (p. 22). Though Agar may be correct with respect to the latter conclusion, analogies between the genetic modification of organisms and the physical modification of atomic elements are misleading because the former are parts of an individual lineage whereas the latter are tokens of a natural kind class. As Lewens notes, biological species are precisely not like elements of the periodic table: they are not tokens of classes whose members are determined by a set of necessary and sufficient intrinsic properties (or in the case of a cluster class, a sufficient number of defining properties) which, once lost or modified, render an organism a member of a different species.

Simply put, species are not categories that can be entered into and exited as the properties of organisms change. The philosophical and scientific consensus is that biological species are lineages: separately evolving, spatiotemporally delimited historical individuals, and qua individuals, they have a single beginning and a single end. If a new lineage arose having (against all odds) independently evolved the same phenotypic properties as the only living species of giraffe (Giraffa camelopardalis), including the relational capacity to interbreed with the same, it would still not belong to or be a constituent member of G. camelopardalis. To think otherwise would be to conceive of organisms as akin to instantiations of atomic elements—tokens of a type that can (in principle) be transmuted in real time into tokens of another type through the alteration of their nongenealogical properties. Whereas tokens of an ahistorical class like “hydrogen” can be transformed into a token of “helium,” an organism cannot analogously “exit” one species and “enter” another through alterations of its genotypic or phenotypic properties, including its capacity to interbreed.

Agar's foray into species essentialism is presumably not intentional. In his discussion of the implications of radical enhancement for persistence of the human species, Agar appears to adopt a version of the standard biological species concept, usually attributed to Paterson (1985), according to which two organisms are conspecifics just in case they share a common fertilization or mate recognition system. What Agar misses, however, is that by referring to a “common” mate recognition system, biologists mean a “homologous” mate recognition system, or a mate recognition system with similar properties due to common descent. Consider two lineages (A and B) that independently evolve mate recognition systems with similar properties. These mate recognition systems would not be homologous between A and B, since the similarity would not be due to common ancestry (i.e., a continuous causal process of evolutionary descent)—and therefore A and B would belong to separate species despite possessing similar mate recognition systems. Likewise, if a chimpanzee embryo were genetically modified such that the resulting hominin could successfully mate with a human, this would not, even on Paterson's view, render it a member of Homo sapiens, since the functionally compatible mate recognition systems would not have a single, common evolutionary origin (cf. Wagner 2007). To fail to conceive of species as wholly genealogical entities is to cling to an obsolescent Aristotelian interpretation of species that is flatly inconsistent with modern evolutionary biology, since it cannot accommodate or make sense of the evolution of species.

Some biologists have attempted to define species a-historically (but nonessentialistically) according to synchronic, hierarchically ordered patterns of similarity, without relying on assumptions or hypotheses of common ancestry (Patterson 1982). Such an account could, in theory, allow for the “exiting” of a species via transgenetic manipulation. However, “pattern cladism” (as this approach is called) has been widely rejected in biology and the philosophy of biology for being motivated by an unwarranted skepticism of evolutionary claims, and for failing to account for (and explain) the origins and persistence of species as well as patterns in the fossil record and biogeography (for an excellent discussion, see Beatty 1982). Taxonomic classification as practiced by contemporary biologists is a thoroughly historical endeavor.

Because the human species is a genealogical entity, the preservation of language, bipedalism, relatively hairless skin, specialized sweat glands, an enlarged frontal lobe, or any other derived human trait is not a necessary condition for its preservation (although some of these may be necessary for the preservation of “humanity” functionally defined, or of species-normal states that we find morally desirable). Phenotypic traits may serve as evidential markers that enable us to distinguish species operationally, but they are not definitional of species. Qua historical reproductive lineages, species can evolve with respect to any of their distinguishing features and nonetheless remain the same species. It would be correct to say that altering the biological traits of individual organisms of a population could create conditions that increase the probability of a speciation event, but whether speciation in fact occurs can only be determined in retrospect, as this will depend on whether the potentially speciating population continues to diverge on an independent evolutionary path over time without going extinct or being reabsorbed by its parent lineage (Mishler and Brandon 1987).

Therefore, changes in particular organismic properties, such as genetics, morphology, physiology, cognition, or even the capacity to interbreed (see, e.g., hybridization in distantly related plant species), are not necessary or sufficient conditions for speciation events (de Queiroz 2007). Widespread human-initiated transgenesis in animals could affect our ability to reconstruct phylogenies for certain animal groups, much as “natural” lateral gene transfer has done in the case of prokaryotes (e.g., bacteria); it does not, however, permit one organism to “exit” one species and “enter” another in essentialistic fashion.

The upshot is that even scientifically informed ethical discussions of the genetic alteration of species, including those touting their consonance with evolutionary biology, have a tendency to get tripped up over unicorns.

2 Human Nature, Evolution, and Development

Leaving essentialistic tendencies behind, I agree with Lewens that more sophisticated attempts (e.g., Machery 2008) to conceptualize human nature so as to maintain its consistency with modern evolutionary theory either (1) fail to include traits that are traditionally associated with human nature due to their partial developmental dependence on cultural resources or (2) result in an overly permissive definition of human nature which, if not an outright reductio of the concept, undercuts any theoretical or programmatic utility it might have had. I will briefly describe this aspect of Lewens's critique and then highlight some of its broader implications for the ethics of biomedical enhancement.

Lewens's argument rests in large part on the fact that development is a highly, if not “irreducibly,” interactive process in which genes are merely one subset of developmental resources that bias phenotypic outcomes. In light of the interactionist nature of development, there is no clear theoretical motivation to exclude from the set of properties that comprise human nature traits that rely heavily on culture for their development, just as there is no clear reason to exclude traits that rely heavily on noncultural environmental factors for their development. In cultural species like humans, some stable phenotypic outcomes that are universally distributed among healthy individuals and canonically linked to “biological” human nature—such as bipedalism and language—rely not only on genes and nongenetic environmental factors (such as nutrition), but also on significant cultural scaffolding and socially mediated learning to ensure their reliable acquisition. It is rare to find surviving human children who have been completely deprived of early social learning, but there are some documented examples in the literature. “Feral” children, who have been severely isolated from human contact at a very young age, fail to develop normal walking posture and basic language abilities, among other canonical human traits and capacities (McNeil et al. 1984).

The recognition that many “defining” biological features of the human species are heavily dependent on culture for their development compels evolutionary psychologists like Machery to incorporate into a description of human nature traits that rely at least partially on genes for their development. But herein lies the rub. The flip side of the developmental equation, as Lewens astutely points out, is that quintessentially cultural traits, such as Catholicism and cricket, rely on biological structures for their acquisition and transmission, including the neurocognitive subsystems dedicated to imitation, motor control, language acquisition, theory of mind, agency detection, and so on. It is noteworthy that autistic children, who exhibit reduced functioning in certain of these cognitive capacities, were confused for feral children in the early scientific literature. Genetic and cultural deficits can thus lead to similar pathological outcomes, suggesting that both are developmental causes, both are integral parts of the developmental explanation, and both can be the difference-making basis of pathological variation.

3 Respect for the Evolutionarily Given

The same theoretical considerations that compel Lewens to reject the most sophisticated evolutionary psychological conceptions of human nature also present serious problems for prominent critiques of biomedical enhancement, such as that due to Michael Sandel.

Sandel argues that the biomedical enhancement project expresses a bad character trait: namely, a Promethean desire to master nature, including human nature (2007; pp. 26, 47). This attitude toward the “natural,” “given” parts of human nature is not conducive to human flourishing, he contends, for it leaves us “nothing to affirm but our own will.” He contrasts the desire for mastery with a moral value he refers to as “respect for the given” or (following theologian William May) an “openness to the unbidden.”

In his brief consideration of Sandel's work, Lewens maintains, contra Sandel, that the mere fact that a trait is a “given” part of human nature says nothing about the ethics of modifying that trait (see also Buchanan 2009). Elsewhere, Lewens (2009) has suggested that Sandel's reliance on the distinction between natural and unnatural capacities is problematic, since it is not clear that genetic interventions interfere with natural capacities any more so than widely accepted nongenetic interventions (such as medically indicated diets). I will expand on this latter argument by showing that Sandel's ethical objection to genetic enhancement rests on an erroneous picture of human nature, development, and evolution.

We shall take Sandel at his word that his argument is a secular, not theological, one. Accordingly, we shall assume that his senses of “given” and “giftedness” in relation to the natural world do not advert to pre-Darwinian notions of final causation or purpose, but rather to what Kahane (2011) calls “theism-neutral existential values”: values we hold toward realms of nature that remain unconquered by human agency. Contemplating biomedical enhancement, Sandel asks us to affirm (in a normative sense that implies positive acceptance) that much of the universe, including and especially our evolved biological capacities, remain “unbidden” or unaffected by our powers of rational manipulation. That a character trait or capacity is the outcome of an unbidden evolutionary process seems for Sandel to constitute a prima facie reason not to alter, manipulate, or otherwise cause it to bend to our will. G.A. Cohen recently defended a similar conservative attitude, arguing that accepting the given as such is an important moral value to be contrasted with the “repugnant” attitude of universal mastery (2011, p. 207).

Faced with the palpable objection that we attempt to control all sorts of “given” biomedical states of affairs that we deem undesirable, such as uncontroversial disease states (p. 101), Sandel is forced to draw a distinction between biomedical interventions that allow our natural talents and capacities to flourish, which he believes express a proper respect for our “given” (read: evolved) biological natures, and modifications that alter these natural talents themselves and thereby express a “bid for domination” over the unbidden. Sandel reasons that although many diseases are unbidden, they are not the sort of given biological states of affairs that command respect, since they tend to undermine rather than ensure the “natural” unfolding of human potential.Footnote 2 “Healing a sick or injured child,” Sandel argues, “does not override her natural capacities but permits them to flourish” (2007, p. 46). It is only in light of this natural biological potential that enhancement constitutes an expression of the desire for mastery whereas the treatment of disease does not.

Biomedical enhancement is thus troubling for Sandel because unlike standard medical treatment, “it distorts and overrides [our] natural gifts” (p. 31). It stands to reason that biotechnological interventions that facilitate the development of our natural biological potential are consistent with Sandel's conservative value of respect for the given, whereas interventions that modify this potential in either direction of the Gaussian distribution (resulting in pathology in one case and enhanced functioning in the other) are inconsistent with it.

This sets the stage for thinking of cultural forces as interfering with, rather than as an integral part of, a natural process of biological development—thus assuming a relation between biology and culture that rests on the same fallacy that motivates many gene-centric conceptions of human nature. Although Sandel's critique is sufficiently broad to encompass many types of cultural agency, he apparently finds genetic intervention particularly worrisome because he views it as proposing to interfere at the most fundamental level with our evolutionarily given potential.

Sandel's normative claim is thus premised on a prior metaphysical one—namely that there exists a biological potential that the agential forces of culture (especially biotechnology) can interfere with, alter, or distort. It is here that unicorns make their appearance, and Lewens's reasoning is once again instructive.

4 Natural Biological Potential: a Unicorn in Sheep's Clothing

Whence this “natural” biological potential with which the “unnatural” agential forces of culture might interfere? Sandel is silent on the matter, but we might guess at the biological assumptions motivating this intuition. One plausible candidate is the notion that biological outcomes are somehow “coded” in our genome, which serves as a “blueprint” for design of the adult organism. More sophisticated versions of this idea allude to metaphors like “genetic information” or “genetic programs” that specify the features of complex biological traits. Even if Sandel and other critics of genetic engineering do not regard genes as essences that cause the manifest properties of organisms, they likely see genes as containing a set of encoded instructions for building the phenotype, presenting a natural (and perhaps normative) “given” trajectory that can be thrown off-track by agentic influences.

This intuition may be appealing, but it is empirically unsupported. Biologists and philosophers of science have had a notoriously difficult time defending the claim that genes play an ontologically privileged role in development as compared to the multitude of nongenetic developmental factors (for a discussion, see Griffiths and Stotz 2006). They have had an even more difficult time defending an account of genetic information that goes beyond superficial metaphor and extends to the phenotype proper.Footnote 3 The “genetic blueprint” metaphor is misleading for reasons that Powell, Kahane, and Savulescu discuss elsewhere in this Special Issue and which I will not rehearse here. The bottom line is that viewing any trait as the normative culmination of some preexisting genetic potential, pace Sandel, is not a metaphysically viable position given our contemporary scientific understanding of biological development. There is no “biological potential” contained in genes that preexists the causally distributed process of development (cf. Robert 2004); ergo, there is no “natural” biological potential with which agential forces can interfere.

Lewens's critique of human nature shows that there is no clean developmental distinction between traits that rely on a bidden process like culture for their development and those that are evolutionarily unbidden, since bits of the bidden and bits of the unbidden are involved in the development of many complex human traits. Sandel's argument is untenable because it assumes, incorrectly, that complex human traits have a biological potential that exists independently of their wider developmental context.

A further metaphysical problem with Sandel's argument is that it assumes cultural traits are not part of our “given” biological or evolutionary heritage. Sandel is right that the human capacity for technological industry is the most powerful tool at our disposal for mastering nature; what he overlooks, however, is that our technological (and other cumulative cultural) capacities are themselves evolutionarily given traits. The unique flexibility of the human genus to adjust to climatic fluctuations via technological innovation helped it to perdure environmental upheavals throughout the Pleistocene, whereas other less techno-savvy hominin groups, such as the Australopithecines, went extinct. This pushes Sandel to a conclusion the irony of which I am not the first author to report (see also Kahane 2011): When we attempt to restrain our evolutionarily given capacity for technological development, say by prohibiting certain biotechnologies, we thereby express a bid for domination over one aspect of the unbidden. In effect, we are trading respect for one evolutionarily given trait for another, without any guiding principles to navigate such conflicts. The deeper trouble, however, is that Sandel's view of the relation between biology and culture blinds him to the conflict between these “givens” in the first place.

5 Combating the Reverse Naturalistic Fallacy

I will conclude by way of a sociological speculation. The pervasiveness of biological essentialism among bioconservatives is not due solely to a lack of familiarity with contemporary biological science. Many bioethicists are motivated to embrace premodern biological theory in an effort to avoid the moral “vertigo” that is perceived to flow from a thoroughly mechanistic, nonnormative biological worldview. For instance, Sandel counsels that if we wish to avoid cultivating an attitude of mastery, we will have to “find… a way beyond mechanism to the re-enchantment of nature” (2002).

Unfortunately for the conservative gut, we cannot rationally reject the best available scientific description of the living universe simply because we find its apparent normative implications unpalatable. Enchanting or not, this is the metaphysical foundation on which our ethical project must be built.