The thesis that species are individuals (SAI) was originally motivated in large part by the so-called Species Problem. As standardly described, the problem is how to formulate a satisfying response to the long-standing dispute among biologists about how to characterize species (Wilkins 2009). One can interpret this challenge in two different ways: first, there is the challenge of actually resolving the ground-level dispute by, say, showing conclusively (or satisfactorily) that a particular conception of species was correct; second, there is the challenge of squaring the prevalent sense that species are in some sense real, objective features of the world with the fact that even after decades (if not centuries), biologists cannot agree on how they should be defined. If species are real, what sort of features of the world are they?
SAI was originally proposed by the evolutionary biologist Michael Ghiselin as a solution to the Species Problem construed in this second sense.11 Here is the basic thought: rather than thinking of species as natural kinds characterized by 
intrinsic essences, we should regard them as spatiotemporally extended concrete particulars—persisting composite objects made up of their organisms. Unlike paradigmatic objects such as rocks or tables, these species objects would typically be scattered at a time, but they would nevertheless be connected over time by the sorts of lineage-forming relations (such as interbreeding and gene-exchange) that biologists often employ for demarcating species in the first place. The objectivity of the part- hood relation’s obtaining would secure the objectivity of species against the threat posed by the species problem. Whatever disagreement there was about how to conceptualize species, there would be some fact of the matter about how organisms were united into larger, persisting individuals, rendering (so the thought went) some concepts objectively better than others at mirroring this structure.
But why accept SAI? That it would solve a problem (if true) does not, by itself, show that it is true. Without surveying all of the relevant arguments (see Slater [2013, §4.2] for more), it is worth mentioning a representative few. In the first place, concerns about essentialist models of natural kinds (alternatives to which were not clearly seen when SAI was first proposed) loomed large. In a two-category ontology—of universals and particulars (or abstracta and concreta), say—a simple eliminative inference suggested that, if they are not abstract types, species must be concrete particulars. This move cohered nicely with Ernst Mayr’s campaign to embrace “population thinking” as an alternative to “typological thinking.” On this line of thought, species are certain sorts of groups of populations (or better: metapopulations), and these are not like properties that things can have; they are things with properties themselves, with a certain spatiotemporal extent—in other words: they are individuals.
A second, and related, style of argument focused on pressing for an analogy between the way biologists think of organisms and how they think of species—for example, in how much disruption or change they can endure or in the level of “cohesion” among their parts (Hull 1978, 347; 1989, 84; 1999).
A third argument strategy highlighted the explanatory power of the SAI thesis. Why are species spatiotemporally restricted (Ghiselin 1987, 128)? Why are there no natural laws about particular biological species (Hull 1977, 150)? Why are biologists hesitant to speak of organisms as “instances” of species—of a particular organism exemplifying the property of “tigerhood,” say (Ghiselin 1974, 536)? All of these questions, the argument goes, would receive satisfying and unifying answers on the thesis that species are individuals. Thus we should tentatively conclude that SAI is probably right.
A certain naturalistic orientation is clear in each of these broad styles of argument. The first is premised on empirical considerations coming directly out of biology. It is an extension of the apparently successful naturalist critique of essentialism (setting the aforementioned complexities aside). The second stems from the attitudes of biologists in how they think about and refer to species. Responding to philosophers’ objections to the argument-by-analogy strategy (e.g., Kitcher 1989), Richards notes,
Those who work in the biological sciences typically don’t see such a distinctive and important disanalogy between individual organisms and individual species taxa. There are, I believe, several reasons based on familiarity with biodiversity, tradition and disciplinary practices... . A full consideration of biodiversity reveals the bias in philosophers’ commonsense notions of individuals, and its focus on vertebrates and humans. (2010, 163-64; see also de Queiroz 1999, 67)
So perhaps biologists’ training and experience in their science disabuses them of the need to see individual objects as spatially cohesive—rather like experience in physics disabuses us from naive conception of objects as solid and continuous. The final argument strategy is transparently naturalistic, using explanatory patterns of inductive inference familiar from the sciences and drawing on purported facts that while familiar to biologists might seem surprising to the uninitiated layperson. After all, this gap between expert and lay beliefs about species could be cited as a main factor behind philosophers’ essentialist foibles when it came to species.
Relatedly, some SAI advocates have argued in a Quinean spirit that the way that biologists refer to species strongly legislates for the individualistic interpretation. Unlike the lay who might think of, for example, being a tiger as a property that things can possess, these SAI advocates will point out that the practice of demarcating species historically indicates that “tiger” functions instead as a singular term. Here is Coleman and Wiley’s version of the argument:
Biological theory is replete with generalizations that seem to be about particular things called “species.” The generalizations “There are species” and “Species are variable” appear to be statements that are truly or falsely said about species and not about the organisms that comprise a species. . Thus, one way to interpret discourse about species is to understand at least some of the expressions used to talk about particular species taxa as genuine singular terms (i.e., as terms referring to particular things rather than designating kinds of things). This manner of interpreting talk about species comprises what we call an “objectual account of species.” (2001, 499-500)
This argument from ontological commitment would of course be stronger if it could be shown that biological discourse could only be captured by treating species names as singular terms—something that Coleman and Wiley attempt later in their article.
Even if none of these arguments are conclusive, it would not be unreasonable to feel swayed by their collective weight. For they each appear to be grounded, in one way or another, in our best biological science. Add to this the consensus among philosophers of biology and biologists attending to the matter, and it can start even to seem foolish to suppose that one could argue against SAI from a metaphysical perspective—particularly if one felt any naturalist sympathies. Would this not be tantamount to a non-naturalistic stance ? I will say no; but vindicating this answer requires articulating a different vision of naturalized metaphysics. While I will not be in a position of completing this task in any detail in this chapter, I want to set off in this direction by considering a way in which (apparently) purely metaphysical considerations tell against this popular naturalistic metaphysics of species and asking how we should think about this debate from a naturalistic perspective.
The considerations in question emerge from attention to the metaphysics of vagueness and indeterminacy over the past few decades, specifically as they concern the possibility of indeterminate parthood, identity, and existence. A good starting point is Gareth Evans’s much discussed (1978) paper purporting to demonstrate the incoherence of ontic indeterminate identity—that is, the incoherence of identity statements that are indeterminate in truth-value not due to any semantic defect or imperfect knowledge. Like many later commentators, Evans took his argument as addressing the coherence of the idea that “the world itself might be vague. Rather than vagueness being a deficiency in our mode of describing the world, it would then be a necessary feature of any true description of it” (208). The argument takes the form of a reductio. Suppose that some thing(s) a and b are indeterminately identical (and that this is a metaphysical, not a semantic fact due to referential ambiguity). Another way of putting this is to say that individual a has the property of being indeterminately identical to individual b. However, b does not have that property: it is definitely itself! So there is some property—being indeterminately identical to b— that a has but b lacks. Since, by Leibniz’s Law, something cannot have different properties than it has, it must be that a and b are in fact distinct; and if they are provably distinct, they are not indeterminately identical after all—contra our supposition.
Evans’s result has been extended to contest the possibility of indeterminately existing objects. In Ted Sider’s argument, we begin again from a supposition
(for reductio) that there are indeterminately existing objects. From this it follows that there are numerical sentences expressing how many objects there are—sentences of the form “3x 3y 3z ... (x ^y & x ^ z &y ^ z ...)”—that are indeterminate in truth- value. But there is nowhere for that indeterminacy to lie, as the vocabulary of such sentences contains only sentential connectives, unrestricted quantifiers, and identity, none of which can admit of indeterminacy. So the sentences are not indeterminate in truth-value after all; reductio achieved (Sider 2001, 125-30).
What does the prohibition against ontic indeterminate existence and identity have to do with SAI? The problem is that biological discourse concerning species concepts exhibits little interest in obeying the strictures against ontic indeterminate existence. Worse, it displays tendencies in the opposite direction. Every species concept of which I am aware allows some indeterminacy in its application. Yet this discourse is what tells us how organisms are united into species—including whether some organisms comprise one species or multiple species. Thus if we interpret SAI as a metaphysically robust account of what species are—that they are real, left-alone objects—our best biology foists upon us an apparently intolerable conclusion: that some objects only indeterminately exist (or are indeterminately each other).
Call this the Indeterminacy Problem for SAI. I think that it provides us with strong reason for rejecting SAI, despite that view’s strong naturalistic credentials. Now of course there are a number of ways in which an SAI sympathizer might respond. Let us set aside responses that significantly weaken the metaphysical import of the thesis—say, by construing it as a sort offagon deparler or a semantic thesis of some kind. These may be worth considering, but they are of little interest in the context of a discussion of naturalized metaphysics. One might attempt to contain the problem for SAI without engaging directly with the metaphysics of indeterminacy by reinterpreting the SAI thesis in a non-mereological way (Haber 2013, 2015). At best, I think this postpones answering key questions: How are biologicalparthood and existence (or biological existence ?) conceptually linked? What is the metaphysics of biological parthood such that it evades the indeterminacy problem? Does it do so, for example, by taking a deflationary approach to the metaphysics of species ? If not, then so long as SAI involves ontological claims about what exists and what is identical with what, it is not clear to me how the precise formulation of a theory of biological parthood will be able to address this problem (Slater 2013, 89).
A more direct way of responding, of course, engages directly with the Evans/ Sider arguments. Might a naturalist simply “recalcitrate” (as Quine put it) and take the existence of indeterminately existing (or indeterminately identical) species for granted and use it as evidence that the Evans/Sider arguments are unsound? Of course; but doing so is unsatisfying. Where specifically does the argument go wrong (cf. Dorr 2010)? Suppose that the SAI-naturalist replies as follows: I do not need to tell you where these arguments go wrong. I have the falsity of the conclusion—that indeterminately-existing things are possible—directly from some very well-confirmed science (evolutionary biology and systematics). All you have is your a priori argument that this is impossible. But since such empirical considerations trump your a priori conclusions, we can be confident that these arguments must go wrong somewhere.
This sort of response will not do as is, however. For one, it is obviously false that the empirical should always trump the a priori. Imagine an investigator coming to a conclusion at odds with some piece of pure mathematics; she would scarcely consider thinking “well, so much for arithmetic!” Rather, she would double-check her work and look for a mistaken assumption or botched procedure, explaining to colleagues that the numbers just did not add up. Now, one might object at this point that mathematics constitutes a principled exception to the above ranking of the empirical above the a priori and that our imagined naturalist’s reply should be interpreted as committed to the thesis that the empirical always trumps a priori metaphysics. But having made one exception, it seems reasonable to wonder how we might recognize others. Moreover, one might doubt that relative epistemic weight is exhausted by the epistemic modality by which we come to believe something.
Relatedly, and more to our present concerns, E. J. Lowe (1994) argued that certain physical scenarios involving quantum entanglement reveal the falsity of Evans’s conclusion about the impossibility of ontic vague identity. Without descending too far into the detail, he considers a scenario in which an electron is absorbed by an atom and another electron is later emitted. Is the emitted electron the same as the absorbed electron? Lowe claims that there is no saying one way or the other (due to quantum entanglement); but not because we do not know enough or because our names are fuzzy. No, the indeterminacy is ontic. So there we go: ontic indeterminate identity, contra Evans, is possible. And this is shown by one of the best-confirmed scientific theories in the history of science.
The case is not so simple, however (for discussion, see Noonan 1995; Hawley 1998; French and Krause 2003). Nor is the comparison obviously helpful for supporters of SAI. For one, and to Lowe’s credit, he did not simply invert a modus ponens into a modus tollens; he offered a diagnosis of the Evans argument. More significantly, as the critical discussion of the paper made clear, a good deal of interpretation of the robust experimental and theoretical results of quantum mechanics is necessary in order for Lowe’s case to do the job it was meant to do. One has to do some metaphysics of science in order for the experimental results to bear on further metaphysical matters. This appears to be a case in Chakravartty’s general point that “even on the most metaphysically austere, contemporary conception of the sciences ... there is a metaphysical dimension to [scientific] inquiry” (2010, 70).
This illustrates a second reason why we should find the recalcitrating response to the Indeterminacy Problem unsatisfying: the conflict is not simply between a metaphysical thesis about ontic indeterminate existence and a robust empirical result; in fact, both theses have metaphysical content. Suppose that the SAI-advocate argues on naturalistic grounds that the closer proximity of their metaphysical thesis to empirical science breaks any imagined epistemic tie: on the one hand, we have a metaphysical thesis that is continuous with current science; on the other, we have speculative a priori metaphysics at great remove from empirical science. We should not be happy with this characterization of the dialectical situation, however. In the first place, the asymmetric labeling of these sources as “speculative” begs important questions; and anyway, scientific conclusions and observations can often themselves be highly speculative. But set this concern aside and grant that the relevant empirical results in biology are as epistemically secure as you like. This is just one part of the story. While there is a certain temptation to gauge the relative epistemic priority of the competing theses on the degree to which they involve metaphysical or empirical premises, the contest should also involve the strength and character of the inferences from these premises.
Here matters become shakier and dependent on further metaphysical assumptions, especially when the inferences are eliminative (e.g., when one argues that species must be individuals because they are not natural kinds), as these inferences assume that we have seen clearly all of the relevant alternatives—something that history shows is often unreasonable to assume (Stanford 2006). Consider Haber’s claim that “species are spatiotemporally located, concrete entities, with a beginning and an end, and have parts, not members, is a function of how lineages are generated, and of the facts of reproduction” (2015, 15). But it is not at all obvious how the biology of reproduction entails these further characterizations without the intervention of further metaphysical theories. Biology can tell us how lineages in fact form, but it does not without significant additional interpretation tell us how they must form or how we should conceptualize them metaphysically.
-  Indeed, the 1974 paper in which SAI received its fullest initial presentation was titled “A Radical Solution tothe Species Problem.”
-  Again, I am not granting these points here. My present focus is descriptive and interpretive. Barker and Wilson(2010) have argued compellingly that the kind of cohesion enjoyed by the members of a species is not especiallycompelling as an integrating force.
-  Lots more could be—and has been—said about the Evans argument and the issues surrounding ontic vagueness; for helpful discussion, I would see in particular Lewis 1988; van Inwagen 1988; Heck 1998; Moore 2008;and Barnes 2010.
-  As with Evans’s argument, Sider’s has not been without its critics—to whom Sider has responded (see Koslicki2003; Sider 2003; Liebesman and Eklund 2007; Sider 2009; Barnes 2010). And while there is certainly roomto press further, my sense is that its conclusion is extremely plausible. Even Barnes’s (2010) critical discussionof Sider (2003) allows that “[i]ndeterminate existence is deeply problematic when interpreted de re. The claimthat there is some thing such that it’s indeterminate whether that thing exists is hard, if not impossible, tomake sense of” (960). David Lewis expressed this same thought when he asked (rhetorically) of a putativelyindeterminately existing object, “What is this thing such that it sort of is so, and sort of isn’t, that there is anysuch thing?” (1986, 212).
-  I document this in some detail in Slater (2013, §4.3.2), but to get the flavor, we can consider a popular speciesconcept like Mayr’s “Biological Species Concept” (BSC), according to which a species is a group of “interbreeding natural populations that are reproductively isolated from other such groups” (Mayr 1963, 17). However,reproductive isolation—at least, as it is used in this species concept—is not generally a sharp notion. Whethertwo populations count as members of one species or two sometimes depends on the degree to which theirmembers can exchange genes (Sterelny and Maclaurin 2008, 28). One might argue that it is possible to sharpenthis criterion and avoid the threat of indeterminacy. But while there are certainly some pairs of organisms forwhich such isolation is very much an all-or-nothing matter, a general insistence on sharpness leads to someimplausible consequences that biologists are not at all eager to embrace. Indeed, many explicitly acknowledge acertain level of vagueness in their account (see, e.g., van Valen 1976, 234).
-  One might hope for an epistemic or semantic resolution of the Indeterminacy Problem. Alternatively, onemight attempt to downplay the significance of ontic indeterminacy—for example, by suggesting that it is infact everywhere. I discuss why I believe that these strategies will fail in Slater (2013, 92-95).
-  This is not, of course, to say that the reverse tie-breaking principle should be employed. As noted above, we haveseen purportedly a priori axioms of Euclidean Geometry overturned by results from physics. And lest we chalkthat episode up as a curious historical anomaly to be reclassified somehow, note that more recent thinkers havefelt pressed, if not moved, to reconsider other a priori claims on the basis of empirical discoveries. For example,Quine (1981) entertained the possibility that certain interpretations of quantum mechanics might even compelus to drop such logically sacrosanct principles as Bivalence.
-  For some suggestions of previously unconceived alternative on this front, see Slater (2013, chap. 5).
-  I am indebted to Celso Neto for comments that helped clarify my arguments in this section.