1. Introduction
What is the status of the divisions that scientific theories purport to draw? Are there natural kinds? Some say yes: the world comes substantially divided up for us—a central task of science is to elucidate and characterize its structure. Some deny that the world has any kinds not beholden to our interests, limitations, or cognitive standpoints—classifications are inventions not discoveries. Often, however, objections against natural kinds monism seem to miss their mark. In what follows, I consider a way of responding to certain of these objections by taking up a minimal monist stance: monists might circumscribe their claim about kinds to the domains of, say, chemistry and physics. Here, the monist thesis about natural kinds looks more secure, as structural differences might plausibly be identified as the privileged boundaries between natural kinds. But even on this apparently monist-friendly ground, stumbling blocks remain. One tricky question I shall consider is: do enantiomers represent distinct kinds? Before turning to this question, let us sketch some of the theoretical terrain.
Let us say that a theory is ‘monist’ with respect to natural kinds if it admits of a unique natural way of “carving up” a portion of the world.Footnote 1 This well-known figure is both apt and misleading. First, it suggests activity where none should be implied—better to think of “plans” for (or “ways” of) carving that precede the wielding of the classificatory knife. Second, it suggests that “carvings” are disjoint, which monism does not require. When we carve the holiday turkey, for example, nonoverlapping pieces result. Monist classification schemes, however, may posit overlapping and hierarchically-ordered groupings: ‘gold’ may name a kind all of whose samples are also members of the kinds metal and element. A monist would be satisfied with a “cognitive carving” of the turkey that recognizes both the dark meat and the wing as distinct, but overlapping, bits of bird. The figure of carving is apt, though, for expressing monism's metaphysical neutrality. It provides no particular explanation for why only one carving is appropriate. Perhaps the world has natural “joints,” perhaps not—anyway, if our knife is sharp enough, perhaps we can cut through the bone. A monistic classification scheme might arise from our peculiar shared way of seeing the world. However, there seems less sense in applying the compliment ‘natural’ to the kinds of such a scheme.Footnote 2 Ordinarily, monism arises from a commitment both to objective measures of sameness and distinction for the objects in question and to certain theses about the relevance of these measures to questions of scientific importance.Footnote 3
Let us say that a theory is ‘pluralist’ with respect to natural kinds if it admits of more than one way of carving up the world. When asked to carve out the dark meat of our turkey, the pluralist replies that there are several cuts of meat picked out by that term—it depends on your theory about and interests in dark meat. There is not a single, “privileged” quantity of meat that is objectively the “dark meat.” Two people might harbor different equally respectable, useful “turkey-schemes” on which different quantities of meat are regarded as saliently dark. And if no turkey-scheme is privileged, no turkey carving is. Such a position, if not ruling out natural kinds altogether, at least minimizes their interest.Footnote 4
2. Structure and Kinds Realism in Chemistry
Perhaps natural kinds monism has received mixed reviews due to philosophers’ emphasis on different starting points for discussion. Those, like Dupré, who take their cue from biology usually find the project of discerning one unique set of natural kinds as forlorn. But the rejection of the long-standing synonymy of ‘kind’ and ‘species’ does not, of course, impel rejection of natural kinds monism tout court. For at its most modest, monistic realism requires only an existential claim: that the world possesses some unique structure fit to uniquely underpin a certain domain of scientific classifications. The claim that there is such a structure does nothing to delimit its extent or granularity. Has the world an objective ethical or aesthetic structure pulling our ethical or aesthetic judgments into convergence? The monist may of course comfortably admit that such judgments cut at no natural joints, that the class of beautiful things or permissible acts do not describe natural kinds. Perhaps this is false comfort for the monist, in that the properties in question lie outside the purview of science. So more to the point: would discovering that there were no privileged boundaries between species force us to give up monism in chemistry?
Clearly not. Giving up monism about species does not require abandoning monism altogether. Perhaps it would be a disappointment to learn that species have no real essence or even that they are in some sense ‘invented’. But this little more constitutes a devastating blow to monistic realism as a general thesis in the above sense than do analogous conclusions about ethics or aesthetics. It is, at best, and insofar as positions were taken to the contrary, a cause for concern. Recent retreats from monism about natural kinds to general pluralism or antirealism may thus look like overreactions. The natural kinds ontologies of chemistry and physics seem, by contrast, robust and well motivated. Those whose scientific tastes draw them to the small and orderly, then, seem more sanguine about monism's prospects. Quine, for example, though in general cautious about kinds and the allied notion of comparative similarity, believes that the latter notion is ready to hand in chemistry. He writes:
Comparative similarity of the sort that matters for chemistry can be stated outright in chemical terms, that is, in terms of chemical composition. Molecules will be said to match if they contain atoms of the same elements in the same topological combinations. … At any rate a lusty chemical similarity concept is assured. (Quine Reference Quine1969, 135)
Quine finds objectivism about chemical kinds lurking in chemical structure. If such properties are objective, then we have an objective notion of comparative similarity and hence a good candidate for monistic chemical kindness. Whereas biological species categories might fail to cut along appropriately “privileged” boundaries, monistic categorization of “chemical species” (a nominal irony) may proceed via structural investigation.
Justifying this structural account of “privilege,” however, requires saying why other modes of similarity don't count toward natural kindness. Here the monist might follow the essentialist in postulating the primacy of structure over function or disposition. Putnam and Kripke, free from Quine's antiessentialist scruples, suggest that the “real essence” of a kind (what it is to be a kind of thing) is having a certain structure. A kind's “nominal essence” (its commonly associated superficial or merely dispositional properties), on the other hand, serves only as an imperfect guide to kind identity. In peculiar circumstances (on Twin Earth, for example), water might be differently disposed (to ooze rather than pour, say)—just as stuff of a different structure might in those circumstances behave quite like water ordinarily does. In such a case, suggests Putnam, we go wrong calling the “watery stuff” on Twin Earth ‘water’. What it is to be water is not to have whatever dispositional characteristics water happens to have here on Earth, but to have the structure denoted by ‘H2O’—the structure, that is, which we discovered water to have.
We should note before going on how the division between nominal and real essence blurs on the structural account of kind membership as we descend the mereo-topological ladder. The structural account, as presented by Quine, is recursive. Molecules, for example, match (are of the same kind) if they have the same atoms in the same configuration; that is, if they have the same kind of atoms in the same configuration.Footnote 5 So in order to understand the notion of kindred molecules we must already possess the notion of kindred atoms. A parallel account will serve for atoms. They match just in case they have the same number of protons—a certain kind of particle itself characterized by its composition of certain kinds of quarks. Now perhaps it's “turtles all the way down” and this story continues ad infinitum. If, on the other hand, structure runs out, the essential properties of quarks (or strings or branes) will be there “on the surface,” so to speak. (At this point, it is an interesting question whether real essence could outrun nominal essence.) This is all just to point out that even on the structural account of kinds, structure might not be able to do all the work. Nonetheless, there is a strong intuition that the distinction is serviceable and important for a wide variety of circumstances as we sharpen folk classifications (at least in certain domains), discovering their underlying nature.
3. A Prima Facie Problem for Monism
The monist commitment to structure as somehow privileged is yet too vague. For structural similarity is a matter of respect—a point many kind-skeptics have taken very seriously. Individual atoms may be alike in nuclear constitution, but differ in electronic configuration. Dupré, for example, laments that though most of the atoms in his car have atomic number 26, “a large proportion of these are not iron atoms, but ferric ions, which obey quite different, and to me less desirable, laws” (Dupré Reference Dupré1983, 327). Here, it seems, a monist can make two claims: either all structural differences represent kind distinctions or only some do. Skeptics see this as amounting to a dilemma.Footnote 6 If it is only some, then decisions must be made about which ones are privileged. However, such classificatory decisions run contrary to the monist ethos: monistic-friendly kind distinctions are supposed to be discovered, rather than decided (Dupré Reference Dupré1993, 17). If, on the other hand, every structural difference were relevant, the number of natural kinds in the world would skyrocket out of control. Of the proposal, for example, that iron atoms, ferrous ions, and ferric ions belong to different natural kinds, Dupré writes:
First, there are good reasons for classifying objects in terms of properties that have some considerable stability over time. This proposal, on the other hand, would imply that objects were constantly changing from one natural kind to another. … But second, and more importantly, even this would not achieve the intended conformity to universal law. For even more transitory properties, such as energy levels of orbital electrons, will also affect the behavior of atoms. We would do better to accept that our fundamental classification should indeed be in terms of (generally) stable nuclear properties, while recognizing that these limit, rather than determine, the states of orbital electrons that are crucial in chemical interactions. (Dupré Reference Dupré1983, 327)
Recognizing any structural (or perhaps even any intrinsic) difference as demarcating a difference in kind seems to entail a problematic instability in our classifications. In many cases, this stance provides us with finer distinctions than are useful or manageable. But there are also tradeoffs for adopting coarser-grained taxonomies: they may not comport in the right way with physical laws. The monist's problem thus resembles Goldilocks’: finding the privileged measure for natural kinds that is just right.
How should such an investigation proceed? Take again the case of water. Putnam suggests that it is the essence of water to have the chemical structure represented by ‘H2O’ and that this was a discovery: after all, we cannot just look and see that this stuff is composed of molecules consisting of two hydrogen atoms and one oxygen atom in a particular configuration. Prior to this discovery, it is not entirely clear what is meant by ‘this stuff’. Even leaving aside Quinean worries about ostensive ambiguity, pointing to a certain sample and dubbing it ‘water’ does not suffice to pick out a certain kind of thing uniquely—for even on a monist view, that sample may be an instance of many overlapping kinds (H2O, liquids, hydrogen-containing compounds, polar solvents, etc.). Things are simplified somewhat if the stuff in question is pointed out and its salient superficial properties are mentioned. Otherwise, we can, like Putnam, specify that by ‘water’ we mean to refer to anything that bears the “same-liquid” (“sameL”) relation to this sample, where x bears the relation sameL to y just in case:
(1) x and y are both liquids, and (2) x and y agree in important physical properties. … Importance is an interest-relative notion. Normally the “important” properties of a liquid or solid, etc., are the ones that are structurally important: the ones that specify what the liquid or solid, etc., is ultimately made out of. (Putnam Reference Putnam1975, 238–239)
The Twin Earth thought experiment supposedly encourages the intuition that structure (and not disposition) is of primary importance.Footnote 7 But we must ask again: structure in which respect?
Consider the structural difference between H2O and D2O: they differ only in that D2O contains deuterium, a heavy isotope of hydrogen (specifically, hydrogen with an extra neutron: hence ‘heavy water’). Despite this, prevalent classificatory practice has it that D2O is of the same kind as H2O—both count as water. The monist might claim that this troubles neither the structural picture of kind essence nor the thesis that we discover the essence of kinds. For D2O and H2O, like other isotopic forms, are alike in structure in a very important sense. They have exactly the same electronic structure. Consequently, they are both polar molecules, excellent solvents, clear and potable in liquid phase, and so on; they resemble each other in nearly all “important respects.” Even if actual prescientific usages of ‘water’ never happened to involve actual samples of D2O—say, if D2O existed only in a remote and undiscovered lake—there is a clear sense that D2O would have been part of the extension of ‘water’ in virtue of its electronic structure. For that is the important structural feature of water.
But, of course, this presumes a standpoint about which are the important or characteristic superficial or dispositional properties arising from water's structural properties. Insofar as the structural difference between D2O and H2O does not result in any “important” dispositional differences, that structural difference might not be regarded as important. Just which superficial properties are taken to be important or characteristic, however, is a matter of decision (see Stanford and Kitcher Reference Stanford and Kitcher2000).
Joe LaPorte makes the point vivid by imagining that Earthlings travel to a world superficially like ours whose oceans and lakes are filled with D2O rather than H2O. The Earthling travelers (sensitive to past referential foibles on the Twin Earth jaunt) decide to test whether what the residents call ‘water’ really is water. And while superficial similarities abound, slight dissimilarities also crop up—the liquid boils and freezes at higher temperatures, is slightly denser, and so on. Further testing reveals a structural difference: the hydrogen atoms in the alien sample have neutrons and are dubbed ‘deuterium’. So some differences are slight. Others are remote: only D2O has the disposition can be used as a source of deuterium. However, some differences, though remote, are not at all slight. Only D2O can be used as source material for a weapon of massive destructive power: the hydrogen bomb. LaPorte asks:
In the light of this information the question is posed to the scientists: is … D2O water? Does it bear the [sameL] relation to the majority of what we call ‘water’? … The scientists conclude that D2O is not true water. … It diverges in behavior from the bulk of what we call ‘water’ on a number of accounts. (LaPorte Reference LaPorte1996, 118–119)
This conclusion, while at odds with that which our scientists evidently made, should seem perfectly plausible. It was considered by our imagined travelers an important manifest property of the alien liquid that it could be used as source material for a bomb—a property which, though not superficial in quite the same way as clarity or color, is rather obvious and important in certain circumstances. Granted, those circumstances could have been ignored, avoided, or never noticed. The point is that it was a classificatory decision, given the discovery of that disposition, that could have been made either way (or later revised). In fact, here on Earth, D2O was discovered several years before anyone suspected that it could be used as source material for Teller's hydrogen bomb, and so its explosive potential was not then considered as legislating for a classificatory distinction. D2O was considered as no more than an interesting subspecies of water. The subsequent discovery of the explosive potential of deuterium, raising this superficial difference to salience, might have impelled a shift of classificatory practice—but it did not. It appears plausible to read this referential stability as a decision to retain the previous classification in the face of an admittedly important discovery of difference—a decision that could have easily gone the other way. And if the extensions of kinds are artifacts of human classificatory conventions, then they are not at bottom objective, let alone monistic.
4. A Monist Rejoinder
The first complaint about the above criticism of natural kinds monism was that it was not sufficiently general. That there are no monistic biological natural kinds does not imply that there are no objective natural kinds: chemistry, for example, looks more hopeful on this front. But even in chemistry, the response goes, monism succumbs to the dilemma about which distinctions to regard as privileged.
Perhaps the monist has a natural line of response. Dupré mentions two desiderata for a robust kinds ontology (Reference Dupré1993, 17):
1. Distinctions should be sharp (in at least a serviceable way).
2. Classifications should be discovered rather than merely invented.
Notice that the second criterion pertains to classifications, rather than distinctions. Questions about classifications, however, don't even get started unless the distinctions are there to begin with. This gap between classification and difference lies at the heart of the dispute between monism and pluralism. Dupré is careful to point out that he does not deny that “a weak version of the doctrine of natural kinds is defensible. That there are objective divisions between some distinct kinds of things would be hard to deny.” What he denies is that “it is a generally appropriate and tractable question to ask, of an object, What is the natural kind to which it belongs?” (Dupré Reference Dupré1993, 5). And the monist may join him in this denial. What properties and dispositions we wish to regard as demarcating important lines of classification depends on our standpoint.
By placing emphasis on the objectivity of the distinctions themselves, rather than the classifications, the monist may adopt this “weak version” of monism and admit that classifications are ultimately decision-based. Classificatory decision making has no relevance to the monist's metaphysical claim. Even essentialists like Putnam leave open the possibility that decision- and interest-relativity can play a significant role in kind reference—such relativity allows that there are more kind distinctions than are dreamt of in one's classification system. The monist may thus grant that classification schemes can be as gerrymandered and parochial as you like (as need or interest dictates), so long as differences are sometimes ignored and sometimes regarded, rather than sometimes invented and sometimes discovered. Differences, the monist suggests, are discovered. Referential decisions may of course be made about which differences should be ignored in constructing a useful taxonomy of the world. Monism is a thesis about the actual kind distinctions (what natural kinds there are), not about the various ways we ignore those distinctions in cobbling together a workable system of classification. Classificatory pluralism does not entail metaphysical pluralism.Footnote 8
Hence the monist escapes the above dilemma by abdicating responsibility for giving an account of which boundaries are ontologically privileged: monists may regard every structural difference as marking off a distinction in kind.Footnote 9 Taxonomic privilege is another matter. Thus, worries about stability or manageability of kinds become irrelevant, as taxonomies may conveniently overlook what are regarded as minor differences. The monist I imagine, then, might have a developed ontology on which natural kinds form a structurally defined hierarchy, where each kind is intrinsically characterized by a real essence. On this picture, the most basic and natural kinds of things are the ‘infimic kinds’: those kinds whose members are intrinsically identical (Ellis Reference Ellis2001, 70). A hierarchy of higher-order kinds accrues from the different respects of intrinsic similarity. Wilkerson, for example, writes:
In insisting that the real essence must be a property that lends itself to scientific investigation and inductive projection, I have left open the possibility that it is some relational property. … However, I wish to argue that we must close off that possibility, and insist that the real essences of natural kinds are intrinsic properties, not relational properties. (Wilkerson Reference Wilkerson1995, 32)
Though such a monist may well join Dupré in denying sense to the question (asked from outside any particular theoretical standpoint) What is the natural kind to which this object belongs?, the weak monistic stance may afford it some sense. A taxonomy may be privileged, after all, in its being the most complete taxonomy—that is, the taxonomy in which no structural features of objects are ignored. Insisting on this notion of privilege may, however, be conversationally inappropriate. For a classification system to be privileged in such a way is not for it to be mandatory in a given discipline—it is not for it to be uniquely “correct.” Different classification systems may be correct so long as they comport with the objective distinctions.
It is not my aim here to describe in detail a monistic theory of natural kinds (for sustained defenses of this sort of monistic position, see e.g., Wilkerson Reference Wilkerson1995, Ellis Reference Ellis2001 especially section 2.3, and Ellis Reference Ellis2002). And no doubt detractors of monism will feel its apparent disinterest in the biological domain as a mark against it—or at least against its relevance to science generally.Footnote 10 I mean only to suggest that monism may plausibly be construed in such a way as to sidestep the general problems some pluralists ascribe to it—there are plausibly monist-friendly domains.
5. Enantiomers as Distinct Chemical Kinds
I have suggested that chemistry is a domain in which monism as an ontological thesis retains some plausibility. H2O and D2O may be treated as of the same kind if convenient, but they are in fact distinct items. For a little philosophic and scientific reflection, the monist supposes, reveals that structure is the chemical similarity concept of note—and actual practice bears this out. I want now to consider a case that I believe makes this foundational claim more problematic.
Enantiomorphs, ever a source of philosophical puzzlement, provide an interesting test case for such a theory of chemical kinds. Prima facie, enantiomorphs have different dispositions, but similar structures. How, then, should we regard them? May enantiomorphs be of distinct natural kinds? For any progress to be made, it seems, at least two other questions must be answered: Do enantiomorphs differ in internal structure? and Do enantiomorphs differ dispositionally? I shall suggest that neither question is very conclusively answerable and thus that neither is the first question of whether enantiomorphs may be of different kinds. Hence, monism is not so obvious a position even in chemistry.
5.1. Enantiomers and Structure
My right hand, ignoring subtle differences, is the mirror image of my left. Their respective parts, intrinsically identical, are related to each other in exactly the same ways. Yet they differ. As Kant put it in the Prolegomena:
What can be more similar in every respect and in every part more alike to my hand and to my ear than their images in a mirror? And yet I cannot put such a hand as is seen in the mirror in the place of its original. … There are no internal differences which our understanding could determine by thinking alone. Yet the differences are internal as the senses teach. (Kant [Reference Kant and Ellington1783] 1977, 286)Footnote 11
We want to say that they are duplicates, their only “difference” being that they cannot be brought to congruity (hence ‘incongruous counterparts’). My right hand, no matter which way I twist it, cannot come to occupy the same portion of space my left hand currently does (even if it was severed from my body). More mundanely, my right hand cannot wear my left hand's glove unless it is turned inside out.
The phenomenon of chiralityFootnote 12 was instrumental in the development of stereochemistry. Jean-Baptiste Biot discovered in 1815 that aqueous solutions of sugar or tartaric acid possessed “optical activity”: they rotated plane-polarized light to the right and left, respectively. Substances which rotated the light to the right were dubbed “dextrorotary” and those which rotated it to the left, “levorotary” (dispositions now enshrined in the ‘D-’ and ‘L-’ nomenclature of chiral molecules). However, the explanation of optical activity remained (as with many chemical dispositions) quite in the dark until correctly guessed by Biot's student Pasteur thirty years later. In his ingenious experiment, Pasteur produced an optically active solution from an optically inactive solution of sodium ammonium tartrate by manually separating and dissolving crystals of different hemihedral forms (like the quartz crystals of different handedness observed by Biot) obtained from the optically inactive solution. Pasteur observed that solutions made from lefty crystals rotated light in the opposite direction from those made from righty crystals, deducing that optical activity had to do with the spatial arrangement of atoms. Molecules could be “handed.” Such differently handed chemical species are referred to as ‘enantiomers’.
Since Pasteur, stereochemistry has taken an understandably central place in chemical investigations. Molecular structure, to a large extent, determines possibilities of molecular interaction (hence the monist's reliance on structure as demarcating chemical kinds). Water is disposed to dissociate sodium from chlorine ions in virtue of its charge distribution—a fact explained by its familiarly “bent” structure and the character of its constituents. Similarly, the structure of a protein (and the conformational changes it permits) often affords incredible reactive specificity. Target molecules fit tightly onto the active site of proteins—thus the “lock-and-key” model of enzyme-substrate interaction.
Does structure alone determine all possibilities of molecular interaction? Two enantiomorphic keys may not both open the same lock—but can we give a structural explanation of this fact? That evidently depends on what we mean by structure. The question of whether enantiomorphs differ structurally is notoriously controversial. Perhaps we ought to distinguish the concept of structure from that of internal structure, where the latter is to be understood as excluding spatial embedding or handedness.Footnote 13 Or we could insist on a more nuanced understanding of internal structure—call it ‘molecular structure’. The monist simply incorporates chirality into the notion of molecular structure, regarding it as a primitive internal feature. This, of course, represents a considerable elaboration of the “ordinary conception” of molecular structure (as, say, average internal angles between bonded atoms and orbitals, etc.)—but ordinary concepts often require revision in light of new discoveries. We might worry, however, about the metaphysical status of the kind distinctions affected by this revision. Upon making this philosophical change do the number of infimic kinds in the world suddenly increase or did we simply discover the correct notion of structure?
Alternatively, the monist might abstain from philosophical revision and hold fast to the ordinary conception of structure, insisting that enantiomorphs differ neither structurally nor in kind. That position, however, looks less plausible when we reflect on the homochirality of terrestrial life. Enantiomorphs can have systematically different dispositions. Let us, in what follows, take the controversial stance that enantiomorphs do not differ structurally and ask if that implies that they must not differ in kind.
5.2. The Dispositions of Enantiomers
It turns out that our amino acids are exclusively of the ‘L-’ configuration; the sugars we use exclusively of the ‘D-’ configuration.Footnote 14 Hence the biological effects of enantiomers may differ widely despite their sharing a great many other properties (e.g., density, boiling point, solubility in achiral media, etc.):
The inherent chirality of living systems dictates extraordinary specificity in the recognition of chiral molecules, so that a molecule and its mirror image, whether it is a pharmaceutical, an insecticide, a herbicide, a flavour or a fragrance, will almost always elicit different biological effects. (Raval Reference Raval2003, 463)
Consider, for example, a quite familiar molecule: glucose. We are so constituted as to be able to digest D-glucose, but not its enantiomer, L-glucose—the former but not the latter has the disposition is digestible by humans. So it seems that unlike many chemical dispositions, digestibility cannot be predicted on the basis of internal structure alone.
How do these considerations bear on the question of whether D- and L-glucose are different natural kinds? There is a prima facie inclination, I think, to abandon internal structure as a privileged dividing line and to regard their different biological dispositions as marking them off as distinct kinds. Only D-glucose has the (quite significant) disposition to be humanly digestible: a hearty diet of food rich in L- but not D-glucose would result in starvation. So what if they have the same structure? This just shows that structural sameness is insufficient ground for individuating distinct infimic natural kinds.
That is a bit too quick, of course. Similar arguments from the “modal properties” of statues and lumps of bronze have been taken to establish the distinctness of those items, despite their perfect spatial coincidence and sameness of parts—a conclusion taken by many as untoward. The question to ask in both cases is not whether modal or dispositional properties are legitimate substituends of Leibniz's law, but whether there really are (and how there could be) different dispositions afoot.Footnote 15 In the case of D-glucose, we must note that it is its relation to other chiral structures that explains its dispositional difference from L-glucose. A molecule's being humanly digestible depends on its ability to play certain roles in various multistep metabolic reactions. D-glucose, for example, possesses a host of chemical dispositions corresponding to the different steps of glycolysis. Consider the first step: hexokinase, a chiral protein, is disposed to add a phosphate ion to the sixth carbon of D-glucose—and only D-glucose is disposed to receive it. Its having that disposition depends on the particular chirality of hexokinase: L-glucose would possess the analogous disposition to react with the enantiomer of hexokinase (call it ‘L-hexokinase’).
Hence the defender of the distinctness of L- and D-glucose becomes faced with a circularity. Whether L- and D-glucose are of different kinds depends crucially on whether they have different dispositions, which in turn depends on the question of whether L- and D-hexokinase are of different kinds.Footnote 16 That, similarly, depends on whether L- and D-hexokinase have different dispositions, which depends (completing our circle) on whether L- and D-glucose are of different kinds. Perhaps this circle is not vicious (we must take care, remember, to distinguish pathological circularity from healthy holism)—anyway, the defender of the sameness in kind of L- and D-glucose is faced with the same circularity. The latter camp, however, may cite a neutral reason for refusing to regard objects with the same structure as possessing different dispositions. Dispositions, they might say, have intrinsic categorical (causal) bases.Footnote 17 On the Armstrongian picture of dispositions (Armstrong Reference Armstrong and Crane1996), for example, holding that enantiomers are intrinsic duplicates immediately shoulders out any possibility that they differ dispositionally, however natural it seems to claim so. Compare: two duplicate vases, one delicately placed on a shelf in a crowded nightclub, the other carefully packed away in a museum basement, may seem to differ in their fragility. Their “difference,” however, is better accounted for by reference to their respective contexts—its being more likely that the first vase will manifest its fragility does not imply that the vases are not both equally fragile.
Conditional analyses of dispositions—wherein dispositional ascriptions are reducible to subjunctive conditionals or functional properties of various kinds—plausibly solve this problem, but have been notoriously difficult to formulate. Dispositions can (apparently) be “finkish” in that paradigmatic triggering events can fail to elicit the disposition's particular manifestation (Martin Reference Martin1994; Lewis Reference Lewis1997; Bird Reference Bird1998; Gundersen Reference Gundersen2002). Defenders of conditional analyses face the further challenge of specifying an appropriate notion of modal “closeness.” In giving a counterfactual analysis of the dispositional difference between D- and L-glucose—something along the lines of ‘were D-glucose ingested by a normal human, it would be digested’—we must rule out presumably remote possibilities regarding the possibility of D-glucose “flipping” handedness. The claim that humans lacking the ability to digest L-glucose is accepted, in part, on certain assumptions about the topology of space. But perhaps there are possible worlds in which L-glucose might, by continuous rigid movements,Footnote 18 switch its orientation, becoming digestible. As Nerlich puts it:
Whether [an object] is enantiomorphic or homomorphic depends on the nature of space it is in. In particular, it depends on the dimensionality or the orientability, but in any case on some aspect of the over-all connectedness or topology of the space. (Nerlich Reference Nerlich1973, 345)
Supposing humans had stretch of nonorientable space between their mouths and stomachs (a sort of three-dimensional Möbius-esophagus), L-glucose would, in the course of its descent, flip handedness, becoming humanly digestible. Wherever located, the existence of nonorientable regions of space introduces a possibility that makes the difference between L- and D-glucose a bit shaky. If it is possible for L-glucose to be mirror-flipped, then it's possible for humans to digest it. Regardless of whether any particular molecule of L-glucose was ever (or will ever be) so flipped, the disposition of L-glucose to be indigestible would be lost. And in losing the dispositional difference between L- and D-glucose, we lose means and motivation for classifying them differently.
Perhaps instead of regarding travel through nonorientable space as bestowing new abilities on L-glucose, we ought to say simply that travel through certain regions might change a particular molecule's kind. It might be possible, that is, for a molecule of L-glucose to become a different kind of thing—in this case, the kind of thing that humans can digest. Putting matters this way only highlights other plausible reasons to resist deeming L- and D-glucose as distinct kinds: kind distinctions should be insensitive to extrinsic change. Consider a schematic example. Suppose that molecules of a certain structure have disposition D 1 in circumstances C 1, but that in circumstances C 2, they lose that disposition and acquire D 2 instead. Obvious cases of exactly this kind come to mind for circumstances of temperature, pressure, and the like. Classifying duplicate molecules differently in C 1 and C 2 raises at least two problems. It saddles the essentialist, for one, with a new, malleable sense of kind essence.Footnote 19 For the very same molecule might, in passing from C 1 to C 2 become a different kind of thing.Footnote 20 More generally, it weakens the relation of kinds and laws. Unless one restricts attention to a particular circumstance, classification of an item as a certain kind might not provide information about whether that thing possessed D 1 or D 2.
Permitting enantiomorphs to differ in kind acknowledges the possibility that real differences in kind might be fleeting. Consider the prospect of enantiomorphic biological species. Sorensen Reference Sorensen2002 examines the case of the “handed variants” of Perissodus microlepis, a species of scale-eating cichlid fish (discussed by Hori Reference Hori1993) whose variants differ in the orientation of their mouths. The fitness of righty fish depends on their relative frequency—being a minority morph confers a selective advantage. Hence, the ratio of lefties to righties oscillates—a presumptive instance of frequency-dependent selection. Though the variants of P. microlepis are not comprehensive enantiomorphs, we may wonder whether true enantiomorphs could differ in their fitness. Sorensen thinks not: relative to the same environment, “an organism cannot be more fit than its perfect duplicate” and enantiomorphs are perfect duplicates (Sorensen Reference Sorensen2002, 409).Footnote 21
Stability of kind weighs heavily on Sorensen. Kind instability should be ruled out: “An organism cannot change its biological type by sheer travel through the same environment” (Sorensen Reference Sorensen2002, 413). Sheer travel might have predictable repercussions—an organism might find itself in the wrong place at the wrong time—but natural selection (insofar as it relates to the fitness of kinds of individuals) will be blind to such predictions. Apparent differences in fitness between enantiomorphs must be accounted for outside of evolutionary theory (if at all).
But consider another try at the restriction sought earlier on what counts as a “close possibility” in our dispositional semantics. Supposing our space to be three-dimensional (at least, macroscopically) and orientable, the above considerations are irrelevant. Let ‘close’ in our account be circumscribed at the outermost by certain topological features of space.Footnote 22 Thus, in close possible worlds—worlds which do not differ too much from ours in their spatial topography—L-glucose cannot by continuous rigid transport become D-glucose just as lefty cichlids cannot become righties. Worries about our classifications losing a grip on circumstance-independent predictions are empirically unfounded.
Still, detractors may deny that the actual properties of space matter (Sorensen Reference Sorensen2002, 417; van Cleve Reference van Cleve1987, section IX). It does not matter that the lefty fish are “stuck” in their orientation (their spatial embedding).
This no more undermines their status as duplicates than if they were stuck in another way. Consider a field of dandelions. A botanist has determined that the dandelions are genetically uniform and so ideal for testing weed killer. He plans to apply the weed killer to the left half of the field and use the right half as a control group. … The botanist's scheme systematically darkens the prospects of the lefties. The lefties cannot escape by migrating to the right half of the field. … But the predictably brighter future of the righties does not make them more fit than the lefties. (Sorensen Reference Sorensen2002, 418)
But perhaps we can deny the force of the analogy. Part of our reluctance to classify the lefty and righty dandelions differently devolves from the irrelevance of their intrinsic properties to their predictably different fates. As with the two vases (one in the nightclub, the other in the museum basement), our reasonable prediction about their difference in longevity depends exclusively on their location (and certain facts about those locations). Classifying them differently on this basis doesn't reflect real difference. The vases could easily be switched, invalidating whatever predictive value associates with those classifications. And even if they were stuck (glued down, say), extrinsic factors might change. The nightclub might get closed down; the museum curators might throw a wild party. In the case of the dandelions, the botanist's intentions might change radically.Footnote 23 Herein lies the disanalogy: if the embedding of enantiomorphs cannot be nonviolently changed (supposing space to be stably three-dimensional), then it does not matter where they are taken or what plans botanists may have—their handedness and the biological dispositions that depend thereon will be insensitive to extrinsic change. On this picture, however, whether intrinsic duplicates may be plausibly understood to differ in kind depends on the nature of space (or perhaps even the remoteness of certain kinds of regions of space).
5.3. Indecision and Pluralism
So where does this leave us? I suggested that a committed monist might attempt to diffuse pluralistic concerns by making two points: First, that arguments drawing exclusively from domains like biology are too narrow to motivate a thoroughgoing pluralism. Second, that pluralism regarding classification schemes is a different business from an ontological pluralism about kinds. Dupré's suggestion that the existence of “real essences would imply that there is some unique, privileged scheme of classification” (Dupré Reference Dupré1993, 60) ignores his own pragmatic scruples about classification. And while the monist may recognize a “privileged scheme of classification” in one sense, this does not imply that other classification schemes may not be “privileged” in different respects. According to the monist, there may be many different, appropriate, eminently justifiable schemes of classification that sit atop a robust ontology of chemical and physical kinds, perhaps characterized by real essences. If the world has a unique, essential structure, nothing says that even if we came to appreciate this structure, it would impel a unique taxonomy. Perhaps the world has far more unique structure than we are interested in countenancing. A most natural (and monistic-metaphysics-friendly) pluralism concerns the various legitimate ways this structure can be ignored.
Enter enantiomers. I do not mean to argue for a view on which enantiomers differ in kind—there is yet much more to say about how such a view might possibly get worked out (especially with regard to the relevance of different philosophical theories of space, parity nonconservation, etc.).Footnote 24 But enantiomers as distinct natural kinds appears to me plausible. So, however, does the view on which their (apparent) difference outstrips the resources we might accord a theory of natural kinds: there is much we can say about the world without overt reference to natural kinds.
We might summarize the options so far surveyed in the following way. The three questions asked at the beginning of this section—do enantiomers differ (1) structurally, (2) dispositionally, (3) in kind?—may each be answered `yes' or `no', resulting in eight possible views. But, of course, the questions are not disconnected. Some views might be thrown out as more or less implausible or unmotivated.Footnote 25 The two views I find most plausible from a monist perspective—first, that enantiomers are structurally similar, but differ dispositionally (and hence in kind); second, that enantiomers are structurally and dispositionally similar (and thus of the same kind)—seem more or less on a dialectical par. Yet they disagree on what natural kinds there are. Adopting the first puts considerable pressure on the monist's thought that structure, at least in chemistry, can underlie a unique kinds ontology—but it does not seal monism's fate unless one can make sense of dispositional difference sitting atop categorical sameness.Footnote 26 Adopting the second view is more consonant with kinds monism but grates against actual scientific practice. The monist can make sense of different scientific disciplines ignoring differences between real kinds better than he can tell a story about differences being attributed ex nihilo or in light of extrinsic differences, for this would run contrary to the thought that the infimic kinds represented the most complete possible system of classification.Footnote 27
If the question of whether enantiomers like L- and D-glucose really have different dispositions is an instance of empty metaphysics, then plausibly, so is the question of whether they are of different natural kinds. Even if the question is substantive, the answer will no doubt depend on further philosophical or methodological commitments. Such commitments thus bear on a metaphysical as much as a classificatory issue insofar as they concern actual differences between kinds and not merely the different ways in which differences might be ignored. The surest road to pluralism—though not, perhaps, the easiest—thus proceeds by way of an evaluation of these relevant philosophical commitments.
6. Conclusions
The jury is still out on whether the world has a unique structure to be revealed by scientific investigation. Perhaps our primary goal in doing science, rather than seeking out this structure, is coming to an understanding of the workings of nature in respects we deem significant (Kitcher Reference Kitcher2001, chapter 6). I see no particular reason, however, to presuppose that significance will cut along ready-made intrinsic joints. If L- and D-glucose differ significantly in their reactivity, given the actual state of things (given their orientation in space, given the homochirality of terrestrial biology), then it had better be open to us not only to regard them as different as part of a convenient story about the world but to hold that they are different—if, as a consequence, essentialism must go by the wayside, then so be it. Similarly, it ought to be open to the essentialist, in honoring her philosophical commitments, to regard enantiomers merely as a telling example of the limited value of a kinds ontology to science. Perhaps the most coherent understanding of these facts is that the world “supports” either ontology—that is, that it is pluralistic (at least, to a degree).
