Hylomorphism, the metaphysics of form and matter, is a theory both ancient and modern. It is, as is well known, the basis of Aristotle’s metaphysics, a basis which carried over into much of medieval philosophy. In that medieval form, it was thought to have been vanquished with the development of modern philosophy and science. But it is making a comeback and today there are quite a number of prominent philosophers who advocate one or another variety of Hylomorphism, often (though not always) describing it as such and acknowledging the influence of Aristotle. In this essay, I want to contrast two different varieties of contemporary Hylomorphism, each of which picks up on a different element in Aristotle’s own version of the approach. One of these contemporary varieties is my own view but this essay is not so much concerned with advocacy of that view (though I will try to indicate why I favor it over the other variety of contemporary Hylomorphism that I look at) as it is with understanding how these modern varieties of Hylomorphism relate to each other, and to Aristotle, with whom I begin.1
Aristotle’s use of the concepts of form and matter extends over many kinds of cases: substances such as biological organisms; ersatz substances such as artifacts; the soul and its relation to the body; chemical theories of stuffs; geometric figures; and no doubt more besides. Here I want to focus on those cases in which the use of the concepts of form and matter most clearly intersects with Aristotle’s explanatory framework of the four types of causes: formal, efficient, final, and material. This intersection between Aristotle’s use of the concepts of form and matter and his theory of the four types of causes is on display in his multiple treatments of biological organisms, which are his paradigm cases of hylomorphically complex substances, and it is this intersection, rather than merely the use of the concepts of form and matter, that I shall call Aristotle’s Hylomorphism. Aristotle’s Hylomorphism is also on display in his treatments of artifacts which, on the one hand, fail to qualify as bona fide substances but, on the other, serve as his primary means of introducing readers to his Hylomorphism.
The intersection occurs in two stages. First, and most obviously, form and matter are themselves identified as two of the four causes – the formal and material causes. The formal cause of something, the “account of what the being would be,” is its form; the material cause, “that out of which as a constituent a thing comes to be,” is its matter (or perhaps its original matter).2 That still leaves the efficient cause, “the primary source of the change or the staying unchanged,” and the final cause, “the end… what something is for” (Physics II,3 194b23-35). The second stage of the intersection between the concepts of form and matter and the theory of the four types of causes is accomplished when Aristotle writes that the formal, efficient, and final causes “often coincide. What a thing is [the formal cause], and what it is for [the final cause], are one and the same, and that from which the change originates [the efficient cause] is the same in form as these” (II,7 197a25). So the efficient and final causes are integrated into the framework of form and matter through their coincidence with the formal cause.
I will briefly give two illustrations of how the formal, final, and efficient causes coincide, one of a biological organism, the other of an artifact. Here is a highly schematic version of Aristotle’s account of sexual reproduction in humans. An individual human is a composite of matter and the form human (which is therefore the individual’s formal cause).3 The efficient cause of the coming to be of this individual is the male parent. “Men come to be from men,” as Aristotle says (II,1 193b8). The parent is able to generate the offspring, a composite of form and matter, because he himself has that form (he is himself a human). So besides identifying the efficient cause with the male parent, Aristotle also identifies it with the form, acting in or through the parent. Thus, the formal cause of our individual is also its efficient cause. But what is the mechanism whereby this generation takes place? The semen of the male parent carries the form in question to the matter provided by the female parent - blood in the uterus. This union creates a composite of matter and form that either is, or becomes, the individual offspring. But clearly, at the point of union, the resulting composite is a long way from being a fully developed exemplar of the human species. It is the form, once again, now informing the matter of the generated individual, that guides the development of the individual through the various stages it must traverse until it becomes a developed, adult human, a completed and exemplary manifestation of the form. Thus the final cause, the end state towards which development is directed, is the form, and hence coincident with the formal and efficient causes of the individual.4
There is one further crucial observation to be made about this whole process concerning the transmission of the form from the male parent to the offspring. This form, I said, is carried in the semen; but the semen itself is not a human being at all, even in an undeveloped state (and not the matter of a human being). Montgomery Furth comments:
A striking aspect of Aristotle’s account is the certainty and clarity of its appreciation of the fact that the biological phenomena require there to be two different ways in which specific form occurs: one the way in which it is exemplified by specimens of the species, and a different way that figures in the copying process from forebear to offspring… [T]he recognition that the second way must indeed be different is perhaps Aristotle’s most remarkable single insight, biological or otherwise. (113, emphasis in the original)5
Furth is surely not wrong to stress the importance of this view for understanding Aristotle since it underlies and renders unmysterious the teleological character of his outlook. The final cause is explanatory of the process that leads up to its attainment not because it reaches back spookily into the past but because it is already present in the past: as efficient cause in the parent, in one way, and in the semen, in the other way.
As an artifactual example, consider a bicycle. The bicycle has matter – steel, rubber, chrome, etc. (or perhaps wheels, frame, tires, etc.); these are what the bicycle is made out of, they are that “out of which as [constituents a bicycle] comes to be” (II,3 194b23). This matter is informed by something, the form of a bicycle, which makes the bicycle what it is. This is its formal cause. What is the bicycle’s efficient cause? Aristotle sometimes implies it is the maker and sometimes the art of bicycle making. What this adds up to is that the maker is able to make the bicycle because of the art of bicycle making that is ‘in’ her in some way, the art of bicycle making which includes the form of a bicycle. Furthermore, it is in virtue of having the form of the bicycle in it that the bicycle functions as a bicycle, that its wheels turn as the pedals do, that changing the gears can make it easier to go uphill, and so on. Thus we have the coincidence of the efficient, formal, and final causes of the bicycle. We also see, once again, Aristotle’s striking insight at work: the teleological explanation here too, in this artifactual case, depends on the possibility of the form of the finished product’s being in things in two different ways. It can be in the bicycle and it can be in the mind of the maker in a way that does not make an actual bicycle out of her mind but enables it to play a role in bringing bicycles into existence.6
What these examples both exhibit, in virtue of the coincidence of the formal, efficient, and final causes, is a tight internal connection between what a thing is, how it comes to exist, and what it does or is for. The origin of something, its essence, and its function or characteristic behavior are not adventitiously connected in such entities. An adequate account of them demands some unity of, or interconnection between, origin, essence, and telos. For the cases to which this theoretical framework applies, ontology is itself essentially both historical and teleological. This powerful philosophical vision – that there are internal connections relating essence, origin, and telos – is distinctively Aristotelian, but it also transcends the particular way in which Aristotle develops it. Aristotle’s version of it depends, as we have seen in the two examples above, on positing the existence of special entities, forms, that are causally active in distinctive ways. They are capable of being in things in several different ways, they can be transferred from one thing to another, and they guide the activity and development of the things they are the forms of. This reliance on the existence of a certain kind of entity (forms) makes Aristotle’s way of developing an ontology that is both historical and teleological what I call “realist.” Modern science finds no place for any causally active entities that resemble forms in being able to enter into historical and teleological explanations of something. Hence the rejection of Aristotelianism by modern science.
In the following, I will discuss two responses that both count as neo-Aristotelian but which pick up on different elements in Aristotle’s account. Crudely put, one response keeps Aristotle’s realism but jettisons the overarching vision in which origin, essence, and telos come as a package. The other, my own view, keeps this vision but relinquishes the realism.
2. Contemporary Realist Hylomorphism
One kind of neo-Aristotelianism is found among philosophers who embrace Aristotle’s realism but adopt a ‘thin’ conception of form according to which it has no essential connection to the broader vision of an historical and teleological ontology. They thereby avoid the conflict with modern science that afflicted Aristotelian Hylomorphism. I call this approach Contemporary Realist Hylomorphism (CRH) and it numbers among its advocates Kit Fine, Mark Johnston, and Kathrin Koslicki.7 CRH takes from Aristotle the idea that certain entities are composites of things that play the role of form, and things that play the role of matter but they look to the abstracta of contemporary ontology – such things as properties, relations, structures, and (mathematical) functions – to play the role of Aristotelian forms.8 Such thin substitutes for full-blooded Aristotelian forms cannot, or can only incidentally, integrate an account of what a thing is with accounts of its origins and telos. In order not to have to work with three different theoretical frameworks, I shall confine my attempt to make good on this claim almost entirely to a discussion of Kit Fine’s work.
Fine postulates a kind of entity he calls embodiments. Embodiments can be rigid or variable. Variable embodiments would include all biological organisms and most (if not all) artifacts, in other words, the paradigm examples of Aristotle’s historical, teleological ontology. I will approach the theory of variable embodiments through the theory of rigid embodiments, starting with a special kind of rigid embodiment that Fine calls qua objects. For any property P that an object O has, Fine suggests there is another entity, O insofar as it is P, or, as he canonically puts it, O qua P. Thus, Socrates was, among other things, a teacher of Plato and snub-nosed. So Fine thinks we should also recognize the existence of Socrates qua teacher of Plato and Socrates qua snub-nosed. These are distinct from each other and distinct from Socrates. As Fine explicitly notes in various places, Socrates is like the matter of each of these objects, and the properties of being a teacher of Plato and being snub-nosed are like the forms. Fine uses the terminology of basis and gloss respectively for the things that play the role of matter and form in qua objects. Fine postulates three axioms governing qua objects:
F1) O qua P exists at a time t if and only if O has the property P at t;
F2) a) O qua P and O’ qua P’ are identical if and only if O=O’ and P=P’; and b) O qua P is not identical to O;
F3) O qua P inherits all of O’s normal properties.9
F1 tells us the existence conditions for qua objects. The axiom places no restrictions on O and P; for any object and any property, if the object has that property at a time, a corresponding qua object exists at that time. Since qua objects may themselves be the bases for other qua objects, and given the conditions on identity of qua objects given by F2, qua objects are likely to be very numerous indeed, even supposing, as Fine does not, a sparse conception of properties. F3 tells us what (some of) the properties of qua objects are. If Socrates is snub-nosed, and being snub-nosed is a normal property, then Socrates qua sitting is snub-nosed too. Non-normal properties include such things as existing, being identical to x, and not being a qua object. Since it will play no role in my discussion of Fine, I leave the distinction between normal and non-normal properties at an intuitive level.
The theory of qua objects naturally suggests an extension of itself to objects that have more than one basis and appropriate n-ary relations in place of properties as their glosses. Fine calls such objects rigid embodiments and qua objects are simply monadic rigid embodiments.10 Canonical notation for such objects is x,y,z…/R, where x, y, z… are the bases of the rigid embodiment so designated, and R, a relation, its gloss. Fine gives as an example of a rigid embodiment a particular suit, S, that has a jacket, J, and a pair of trousers, T, as its bases and the relation of being made for one another as its gloss: J,T/being made for one another. A number of postulates are given for rigid embodiments, mostly mirroring those for qua objects but also describing explicitly the part-whole relations that characterize rigid embodiments.
It will be evident that rigid embodiments, as their name suggests, do not allow for the change of matter over time that is characteristic of hylomorphically complex objects like bicycles or biological organisms. F2(a), and its analogue for the general case of rigid embodiments, make it impossible that Socrates qua snub-nosed should continue to exist while coming to have Callias, rather than Socrates, as its basis or that S should continue to exist while coming to have a different jacket, J’, as one of its bases.11 Rigid embodiments do not have ‘metabolisms,’ literally or figuratively. To give a hylomorphic account of entities which can change their matter over time, Fine develops the theory of what he calls variable embodiments. Despite the common language of embodiment, however, and despite the connections between the theories of rigid and variable embodiments, variable embodiments turn out, from many points of view, to be quite different kinds of things from rigid embodiments: not a simple extension of the latter somehow to accommodate change of matter or parts, but a completely different kind of entity.
The rough idea is that a variable embodiment is like a succession of rigid embodiments tied together by some abstract string. Associated with a variable embodiment (e.g. with a particular organism or artifact), canonically represented as /F/, is a function F (called the embodiment’s principle) that takes times into objects. The object determined by F at t is called /F/’s manifestation at t. We can therefore say that a variable embodiment exists at a time just in case it has a manifestation at that time. And /F/ is identical to /G/ just in case F=G. Variable embodiments can change their matter or parts because the functions that are their principles may map different times onto different objects. For example, let B be the function corresponding to my bicylce=/B/. (We may suppose it agreed that a given bicycle can continue to exist even as some of its parts are changed.) B may map t1 onto a rigid embodiment BRt1, which itself is analyzed as (with a lot of simplification about what a bicycle is like) W1,W2,F/Bi, where the Ws are two wheels, F is the bicycle frame, and Bi is the relation that obtains when two wheels are attached to a frame in the correct ‘bicycle-ish’ way. Between t1 and t2, I change W2 for a new wheel W3. My bicycle, i.e. /B/, persists because its function B associates t2, after the wheel change, not with BRt1=W1,W2,F/Bi (which of course itself does not exist at t2) but with BRt2=W1,W3,F/Bi. It should be evident from this example that the combined theories of rigid and variable embodiments enable us to describe objects with highly complex hylomorphic structures. Here, the manifestations of /B/, a variable embodiment, are the rigid embodiments BRt1 and BRt2. These rigid embodiments have as their bases wheels and frames, which themselves may be variable embodiments. Those variable embodiments will be manifested at different times by different rigid embodiments so that one and the same wheel or frame may persist through changes in its parts (spokes, handles, etc.).
Fine treats the bases and gloss of a rigid embodiment as parts of that embodiment. Both Socrates and being snub-nosed are parts of Socrates qua snub-nosed. Since the parts of a rigid embodiment are not subject to change, Fine takes the parthood relation in question to be one of timeless parthood. With variable embodiments, we encounter the notion of temporary parthood. Fine takes a manifestation of a variable embodiment at a time to be a part of that embodiment at that time, hence a temporary part of it. The principle of a variable embodiment is a timeless part of it. Further axioms relate the two notions of parthood. For example, W1 is a timeless part of the rigid embodiment BRt1; BRt1is a manifestation, and hence a temporary part, of /B/ at t1; so Fine’s axioms imply that W1 itself is a temporary part of /B/ at t1 as well. All of the parts of a variable embodiment that come to it via its manifestations (i.e. all parts of it other than its principle) are therefore temporary parts of it and, at least as far as Fine’s mereology applies to embodiments, all temporary parts are parts either by being manifestations, or by being parts of a manifestation, of something. The relation of manifestation, therefore, is the crux for understanding how things can change their matter or parts over time. Variable embodiments are, at any time at which they exist, initially analyzable into two parts, a timeless part which is their principle and a temporary part which is their manifestation at that time. These, as Fine himself notes, are the counterparts of the form and matter of Aristotelian Hylomorphism.
I claimed above that Contemporary Realist Hylomorphism, here exemplified by Fine, while preserving Aristotle’s realism cannot, or can only incidentally, capture the richness of Aristotle’s Hylomorphism with respect to its integration of the origins, essences, and teloi of the objects in its domain. Let me now attempt to explain why. For the sake of simplicity, I shall develop my comments around monadic rigid embodiments - qua objects - and so pretend that the examples I deal with do not, and cannot, change their matter over time.12 Take a simple example, a flint arrowhead. By analogy with Fine’s (self-confessedly simplified) treatment of a bronze statue as the bronze qua statue-shaped, one might think to identify the arrowhead with the qua object the flint qua arrowhead-shaped.13 Though it is indicated partially by means of the concept arrowhead, the property of being arrowhead-shaped is itself a purely natural one. Any old piece of flint can have it, by design or by accident. As an artifact, however, it is essential to the existence of an arrowhead, rather than merely an arrowhead-shaped piece of flint, that it be the product of a certain kind of intentional making and that it have a certain telos or function. Its function is to bury the arrow in what it is shot at, and its origin requires that it be made with the intention of having this function, or the intention that it be an arrowhead. (In Aristotle’s terms, the form of the arrowhead must come to it from the fletcher and it is that form that it gets from the fletcher in virtue of which it has the function that it does.) These features of an arrowhead are not guaranteed to the flint qua arrowhead-shaped. Thus, prima facie, Fine’s theory fails to offer a unified account of the essence, origin, and telos of the arrowhead.
I say ‘prima facie’ because it might be thought that I have simply picked the wrong qua object with which to identify the arrowhead, and that Fine’s theory implies the existence of some other qua object which, if identified with the arrowhead, will ensure the right facts about its origin and telos. There is some truth to this objection, but I shall argue that it will not fully close the gap between Fine’s approach and Aristotle’s in the relevant respects. First, we must ask which qua object is a better candidate to identify the arrowhead with. One might think there is a qua object the flint qua arrowhead and that this is what the arrowhead is. Being an arrowhead, unlike being arrowhead-shaped, is not the kind of property that something can have by accident. Indeed, my claims in the previous paragraph were precisely that nothing could have the property of being an arrowhead (nothing could be an arrowhead) unless it had the right kind of origin and telos. Hence the flint qua arrowhead seems like an object that necessarily has the right kind of origin and telos to be the arrowhead. Unfortunately, however, on Fine’s theory, there is no such qua object. The flint and the arrowhead (assuming the arrowhead is a qua object with the flint as its basis) must be distinct (by axiom F2(b)). So the flint is not identical to the arrowhead and thus cannot have the property of being one. The object the flint qua arrowhead, however, only exists at such times as the flint does have the property of being an arrowhead. Hence, it never exists.
One relevant property that the flint might have is not that of being an arrowhead, but that of being the matter of one. In fact, if the treatment of things like arrowheads as qua objects is on the right lines, the flint must have that property.14 So, there must exist an object the flint qua matter of an arrowhead. Might this qua object itself be the arrowhead? There are reasons to be cautious about such an identification. The flint will only have the property of being the matter of an arrowhead if there is some qua object, of which it is the basis, which is the arrowhead. To take that object itself to be the flint qua matter of an arrowhead (i.e. the flint qua being the basis of a qua object that is an arrowhead) seems to invite worries about well-foundedness.15 Exactly how to make precise these worries is beyond the scope of this paper and it is conceivable that they might be finessed in such a way as not to disallow the identification of the arrowhead with the very object, the flint qua matter of an arrowhead. But perhaps a superior candidate for identification with the arrowhead would be the flint qua having been intentionally worked on with the aim that it should be the matter of an arrowhead. (I shall abbreviate this property, or others of similar form, as Int.) Since being the matter of an arrowhead is not here taken as a property of the flint, but occurs only in the content of an intention, problems about well-foundedness would likely be avoided. So, let us take this qua object (the flint qua Int) to be the best candidate for identification with the arrowhead that Fine’s theory can offer. Still, the theory comes up short, I shall argue, in its attempt to combine realism with an account on which essence, origin and telos are integrated in the manner of Aristotle’s theory.