Haugeland on hylomorphism

In his essay “Ontological Supervenience” (in his anthology Having Thought: Essays in the Metaphysics of Mind), John Haugeland puts forward an unusual criticism of hylomorphism, essentially accusing it of being too parsimonious.  The standard objection to hylomorphism is that it posits more distinctions and entities than are necessary.  Haugeland suggests that it posits too few, thereby failing to capture all of reality.

Let’s take a look, starting with the key passage:

The fatal flaw in hylomorphism is that it leaves too little room for distinctions: being concrete, individual, temporal (contingent), and material are all lumped together.  That is, all and only material entities are particular, temporal concreta – everything else is an abstract, eternal kind.  Moreover, there is only one possible relation between the two sides: inhesion (instantiation…).  And, finally, the modes of composition are asymmetrical: you can go horizontally or up, but not down.  That is, you can take an arbitrary bunch of material individuals and fuse them into a new one; or, you can abstract away from – rise above – all their materiality to get a pure form...  What you can't do is go the other way, and make matter out of forms – once eternal, always eternal.  The result is a rigid hierarchy, with all temporal individuals exactly on a par at the bottom. (p. 121)

The first thing to say is that Haugeland appears to have Aristotle’s version of hylomorphism in mind, rather than the emended versions developed by Aquinas and other Scholastics.  For those later versions do indeed recognize further possibilities beyond those to which Haugeland says hylomorphism is limited. 

Consider the account Aquinas gives of angelic intellects.  Each such intellect is a concrete particular, not an abstract kind.   And since each has a nature, each can be said to have a form.  But an angel is not a material substance, and thus its form is not instantiated in matter.  Since an angel is immaterial, it is also not in time, though it is also not strictly eternal.  It has an intermediate kind of existence which Scholastics called aeviternity.  Moreover, though there is a sense in which it exists in a necessary way, there is also a sense in which it is contingent.  It is necessary in the sense that once it exists, it cannot be made to go out of existence by anything in the created order, either in its own nature or in other created things.  But it is contingent in the sense that, like anything else in creation, it could not exist at all if it were not caused to exist by God, and it could be annihilated if God ceased conserving it in being.

No doubt Haugeland wouldn’t acknowledge the existence of angelic intellects.  He might also object to the metaphysical apparatus Aquinas deploys to make sense of immaterial substances, which includes notions such as the real distinction between essence and existence.  But that is not to the point.  What matters is that the key notions of hylomorphism in fact can be and have been systematically elaborated upon and supplemented in a way that allows it to accommodate more kinds of reality than Haugeland thinks it can.

But why would Haugeland suppose in the first place that there really are any entities that hylomorphism cannot capture?  The answer is that he offers a couple of specific examples that he thinks don’t fit comfortably into hylomorphism’s ontology.  He asks us, first, to consider a story and its relationship to the particular material entities that convey it (such as a collection of ink marks on the pages of a book).  Deploying the type-token distinction, Haugeland says that the story itself is a type and the different sets of ink marks that convey it (in different copies of the same book, say) are tokens of this type.  He claims that “in some sense, a story-type is composed or ‘made up’ of its tokens: it has its being in and through them – without them it wouldn’t exist at all” (p. 121).

But exactly what, Haugeland asks, is the relationship between these tokens and the type?  Should we think of it as a part-whole relationship?  That can’t be right, for that would make copies of a story parts of it in just the way that chapters in a story are parts of it, which they obviously are not.  Moreover, if there were only one copy of a story, the distinction between type and toke would collapse.  Should we think instead, asks Haugeland, of a story-type as a timeless kind?  But a story is temporal and contingent, coming into being at some point.  And timeless kinds are not like that.  Furthermore, any given particular story is not really itself a kind, but rather an instance of a kind – of the mystery story kind, or the romance kind, or whatever.

It’s not clear to me exactly how this is supposed to be a problem for hylomorphism.  For one thing, Haugeland’s suggestion that “a story-type is composed or ‘made up’ of its tokens” seems to me just wrong.  The word-type “cat” is not somehow made up of all its many tokens (all the particular individual instances of the word written in pencil, ink, or chalk, the various verbal utterances of it, etc.) as is evident from the fact that all of those could go out of existence, but the word “cat” would not thereby go out of existence.  Word-types are abstract objects of a sort, and story-types seem to be too.  But abstract objects are not “made up” of anything.

Perhaps Haugeland merely means to suggest that the hylomorphist must think of a story-type as made up of its tokens?  The idea here, perhaps, is that since hylomorphism takes things to be made up of form and matter, it must regard a story-type as a kind of form and its tokens as a kind of matter.  But in that case (Haugeland might then be objecting) this proposal is open to the difficulties he identifies.

But if this is what Haugeland means, the problem is that I don’t know of any hylomorphist who would conceive of story-types in this fashion.  Nor, as far as I know, would any hylomorphist say that everything, without qualification, is made up of form and matter.  The immediate application of the form-matter analysis is to physical substances, specifically.  A stone, a tree, or a dog is composed of form and matter – more precisely, of substantial form and prime matter – but there are lots of other things that are not.  I’ve already given one example, namely angelic intellects, which are immaterial substances.  But there are lots of other things that are not made up of form and matter.  For example, substances have attributes and bear relations to one another.  And attributes and relations are not made up of form and matter (even if the substances that bear the attributes and relations are made up of form and matter).

And the ontology of the typical Scholastic hylomorphist goes well beyond this.  For example, there are what Scholastics call “beings of reason” – things that exist as objects of thought.  Now, this is how to understand abstract objects.  They are natures, properties, patterns, and the like considered by the intellect in abstraction from the concrete circumstances in which they might be instantiated.  And this, I would say, is also how what Haugeland calls “story-types” should be understood.  They are “beings of reason,” not physical objects or even immaterial substances.  Hence it is a mistake to try in the first place to give them a form-matter analysis, so that the difficulties in doing so identified by Haugeland are moot.  Once again, Haugeland sees a difficulty for hylomorphism only because his conception of hylomorphist ontology is simplistic and neglects what later Aristotelians added to the picture.

The same can be said of Haugeland’s other example.  He asks us to consider a club devoted to some hobby, which has twelve members.  He says that “in some sense, a club is composed or ‘made up’ of its members” (p. 122).  But he also takes it to be obvious that “a club is identical neither to the set of its members nor to the fusion of their bodies.” (ibid.).  And we can readily agree, given that, for example, a club can persist despite a complete change in membership.  But then (Haugeland seems to think) it’s not clear what hylomorphism would say is the relationship between the club and its members.

The problem here is that, like many critics of hylomorphism, Haugeland neglects the distinction between a substantial form (which marks a true substance) and an accidental form (which is what mere aggregates and artifacts have).  The latter have looser identity conditions than the former, identity conditions that can depend on human custom or convention.  One mistake critics of hylomorphism make is to take an example of some aggregate or artifact, note that there is a difficulty with giving its identity conditions (which is not surprising given that these sorts of entities inherit all the messiness of human purposes), and then fallaciously conclude that the hylomorphist account of true substances is therefore problematic.  That seems to be what Haugeland is doing here.  A club is a kind of artifact, and thus inherits all the messiness that artefactual kinds tend to exhibit given the vagueness, contradictions, etc. of human purposes.  But this tells us nothing about the plausibility of the hylomorphist analysis of natural kinds (stone, water, lead, gold, trees, dogs, etc.).

In fairness to Haugeland, it should be noted that his primary target in the article in question is not hylomorphism itself, but a metaphysical position developed by Geoffrey Hellman and Frank Thompson which Haugeland thinks is in certain ways similar to hylomorphism.  Hence he criticizes hylomorphism as a way of indicating what he thinks is wrong with their position.  It may be that the deficiencies in his objections reflect an inadvertent assimilation of the one to the other – that what may (or may not) be good objections to the Hellman/Thompson view are simply non-starters when applied to hylomorphism as the Scholastic tradition developed it.

For a detailed exposition and defense of hylomorphism (or “hylemorphism,” a spelling which is less common but which I prefer), see my book Scholastic Metaphysics, especially chapter 3.