Primary and Secondary Qualities in Locke

In Book II chapter viii of the Essay Locke provides two related criteria by which primary qualities are to be distinguished from secondary qualities: first, primary qualities are inseparable from the bodies in which they inhere, whereas secondary qualities are not. Second, secondary qualities are merely the observable effects of primary qualities and not real in the same sense in which primary qualities are, indicating an ontological dependence of the former on the latter. Locke uses these two criteria of demarcation to argue to a conclusion about our ideas of primary and secondary qualities, namely that ideas of primary qualities resemble those qualities but our ideas of secondary qualities do not. I will argue here that Locke’s two criteria are flawed and that he is inconsistent in applying them to qualities, and that his mistakes in correctly drawing the ontological distinctions between primary and secondary qualities result in his misapplication of those distinctions to the epistemological realm.

The Secret Connexion: Causation, Realism, and David Hume by Galen Strawson
My rating: 4 of 5 stars

Very good book. Meticulously counters the traditional view of Hume as dogmatically denying the existence of causation. Strawson shows (by close reading of Hume's texts) that Hume's arguments about 'necessary connexion' being only in the mind are epistemological rather than ontological. Hume avoids the mistakes of the 20th century positivists who call themselves Humeans. 'Hume, then, is not a "Humean"' (p. 228).

The Weird World of Leibniz's God

One of the things that gets trotted out when someone argues that faith and reason are compatible is that many contributions to science or philosophy have been made by individuals who were religious. If faith and reason are incompatible, it is claimed, how can so many scientists and philosophers in history have been Christian (or Jewish, or Muslim, or...). But this quick argument--some philosophers are Christians therefore Christianity is reasonable, Q.E.D.--moves a little too quickly. Instead we need to ask questions about what role the idea of God plays in some particular thinker's philosophical system to assess whether his faith can be seen as truly consistent with his intellectual contribution: is the idea of God required for the system to be coherent? That is, is God a fundamental concept? Is the idea of God consistent with the rest of the system, but superfluous? Is the idea of God inconsistent with it? Because if God isn't required for, or consistent with, the wider system of thought, then it might be that all we really learn about the relationship of faith to reason from a religious philosopher is that people are capable of operating in spite of quite a lot of cognitive dissonance. One way to proceed with investigating the supposed compatibility of faith and reason would be to look at  thinkers who were personally religious and see what role God plays in their systems. What kind of world do we get if we start with the assumption that God exists and work from there? In this post, I'll look at the very bizarre but apparently internally consistent system that Leibniz built on some standard assumptions about God.

Meat Computers and Metaphysics

Jerry Coyne and Massimo Pigliucci have engaged in intellectual internet fisticuffs on the problem of free will, which is unsurprising given that they seem to disagree on nearly everything. But their particular disagreement is more interesting than "We have free will." "Nuh-uh." "Yeah-huh." This dispute has brought up some interesting questions about the nature of the free will debate, which I'd like to explore a little bit. What kind of question are we asking when we ask if we have free will? Is it a metaphysical problem? Scientific? Conceptual?

Freedom's Just Another Word for Nothing Left to Choose

The arguments for determinism are fairly straightforward, and unlike a lot of philosophical speculation can be formulated in relatively easy to understand terms. Sam Harris has a series of posts in which he outlines, in pretty accessible terms, the basic structure of the argument for determinism. In what follows here I'll discuss some of the interesting features that Harris brings out of the arguments for determinism, and take a careful look at some of his claims about the relation of the free will problem to moral philosophy.

Anything You Can Do I Can Do Or Not Do In Some Relevantly Similar Possible World

As I discussed in the first post on the free will topic, the idea that we are free seems obviously true. This is the intuitive position on the free will debate, or what we can refer to as the received view or the common sense view: we have free will. It is essentially the default position, one for which mountains of introspective evidence can be assembled.

But what is exactly meant when we say that we are free? According to just about any doctrine of free will, when I make a decision a few different things have to obtain in order for us to call the decision free. First, it must be possible for me to have done otherwise. At the moment of decision, it must be the case that given a list of possible choices, it is in a real sense undetermined which action I will select. Second, I (as an independent agent) am the cause of the action that I undertake. The decision I make is the result of my deliberation and my desires. Third, I am responsible for my decision. This follows from the conjunction of the two other, but is important enough to emphasize separately. In what follows I'll briefly sketch what is entailed by each of these components of free decisions, why we think it's important, and touch on some of the reasons for the belief that each of components really does obtain when we make decisions.

The Problem of Free Will (what's the problem again?)

Are we the authors of our own actions? Do we have free will? Whether this question can be answered in the affirmative has far reaching consequences for moral theory, political discourse, and even our interpersonal relationships.

At first glance it seems like a textbook example of a question so stupid only a philosopher could ask it. Of course we are free. We experience our free choice all the time, in big and small decisions. We decide what to have for breakfast, what to wear, what careers to pursue, who to marry, and a whole host of other things. What could be more obvious than that we have free will? It turns out, though, that when you look closely at the idea of free will, especially in light of our ideas about causation, the question isn't nearly as stupid as it looks.

Of Elephants and Tortoises: Locke on Substance

There is an apparent tension in Locke’s Essay regarding the idea of substance. On the one hand, Locke spends a great deal of time discussing the origin of this idea and detailing how it fits into his overall taxonomy of our ideas. On the other hand he seems at times to dismiss the idea as having little utility, even to the point of ridiculing it at times. I think the tension is the result of Locke’s attempting to reconcile the idea of substance with his "corpuscularian" view of how bodies are constructed, which leads him in the end to substitute the notion of real (or hidden inner) constitution and real essence for the substance/accident model of Scholasticism.

How seriously should we take Hume's skepticism?

The Treatise of Human Nature has given David Hume a reputation for being the arch-skeptic. After beginning his work with the goal of detailing a "science of man" he proceeds through the argument to undercut our confidence in the reliability of the senses, the objectivity of our notions of space and time, the reality of causal laws, and the rationality of our beliefs. By the end of Book I, he seems to have devastated human knowledge and beliefs, leaving us with the mere shell of skepticism. But then, paradoxically, he proceeds to write two more books in the Treatise as well as many other works in which these considerations appear to be largely ignored. So the question becomes: how seriously should we take Hume's skepticism?

McCarthy, Metaphysics, and the Moral Structure of the World

Cormac McCarthy is a philosophical novelist, even if he might not apply such a label to himself. He combines a subtle portrayal of the complexities of human nature and moral judgments with metaphysical ruminations on the deep structure of the world. He does this without explicitly invoking philosophers or having his characters engage in much lofty dialogue. His characters, with few exceptions, do not contemplate deep questions. And yet, those deep questions pervade his works and form the logical structure under which events unfold in his narratives.

Species and Essence in Locke’s Essay III.vi

In book three, chapter six of his Essay, Locke discusses the relationship between species and essence in the context of a broader discussion of the names of ideas of substances. In the sixth section of the chapter, Locke appears to claim that real essences are properties of species or sorts, rather than individuals. Interpreting the passage in this way poses two related problems for Locke’s theory. First, if Locke is claiming that the real essence, rather than the nominal essence, is closely related to the species, then this is a direct contradiction of what Locke says about which type of essence is possessed by species elsewhere in the essay. Second, given Locke’s criteria for the demarcation of real from nominal essences earlier in the essay, it is difficult to see how a claim that substantial species have real essences is even comprehensible. I want to consider these problems in more detail and offer an alternative interpretation of the passage in question that attempts to avoid attributing contradictory or unintelligible views to Locke.

The passage at issue is found in III.vi.6. Here Locke is discussing the difference between real and nominal essence in substances, and after defining real essence using his standard definition (the real constitution of a thing which is the foundation of its properties), Locke proceeds with the odd claim that essence “even in this sense, relates to a Sort, and supposes a Species: For being that real Constitution, on which the Properties depend, it necessarily supposes a sort of Things, Properties belonging only to Species, and not to Individuals.” An individual parcel of matter has no essential properties; it is just the arrangement of minute particles and their motion, which can change or be changed by its interactions with other bodies in the world. Therefore, according to Locke, any properties referred to as essential must be so only as part of a species concept, since it is only our general ideas which possess the requisite immutability to have essential properties. A “face value” interpretation of this passage leads us to the conclusion that Locke holds that species, not individuals, possess real essences.

The first problem with the face value interpretation is that the claim that species have real, instead of nominal essences, is in direct contradiction with things Locke says elsewhere in the Essay. For example, in III.iii.11-12 Locke concludes that “General and Universal do not belong to the real existence of Things” but are inventions of the mind for use in organizing perception and “concern signs only”. A general term signifies a class of a sort of thing by “being a sign of an abstract Idea in the mind”; under which particulars are classed which “agree” with the abstract idea. Species are on this account nothing but abstract ideas. Species do not exist in the world. To be a member of a species is nothing more than to have a right to be called the name of that species, a right which is derived from agreement with the abstract idea. This is the type of idea that Locke speaks about a few sections later in III.iii.15-16 when discussing nominal essence, which instead of referring to the real, hidden constitution of things is applied to the “artificial Constitution of Genus and Species.” This is highly suggestive that Locke does not consider species of substances to have real essences (i.e., real constitutions) but rather artificial constitutions and nominal essences.

In addition to these passages suggesting that species have only nominal, and not real, essences, there are passages that indicate that only the nominal essence is used in the formation of species concepts. III.vi.7-8 indicates that the nominal essence is at work in the classification of species. It is impossible to arrange or classify substances by reference to their real essences (which are unknown), so the only means by which to erect species boundaries is by reference to their nominal essences. We class as the same species different particulars whose properties agree with the abstract idea that is signified by the name of the species. Locke also cites as additional evidence that species classifications are based on nominal essences rather than real essences the fact that two objects of the same species can differ from each other in their inessential characters as much as they might differ from another object from another species, without disturbing that classification, even though all properties flow from the same real constitution. A similar point is made by Locke at III.iii.17. So if we are to keep Locke from contradicting himself and also interpret our original passage as indicating that Locke holds that species, not individuals possess real essences, we will owe an explanation as to why species are listed as just the types of things which have nominal essences instead of real essences in the case of substances as well as why only the nominal essence would be useful in species classification.

In addition to the problem of keeping Locke from contradicting himself, there is the problem of making sense of what a real essence of species could possibly be, given Locke’s discussion of the differences between real and nominal essences. For Locke, real essences are related to the real constitutions of things, which are unknown in substances (III.iii.15). In substances, those essences are material; they are the physical arrangements of the parts of substances, from which their properties flow. The conjunction of properties which makes up our idea of a species (i.e., its nominal essence) is distinct from the real essence in substances. Locke claims that the essence of a species is nothing but the abstract idea to which its name is annexed (III.iii.12; III.vi.2) and that no one would assert that the abstract, complex idea of a species is the real essence and source of the properties we observe (III.vi.3).

But if a species is an abstract idea and possesses a nominal essence, what would it mean for the species to have a real essence? In other words, how can an abstract idea have a hidden internal constitution from which the properties of an object flow? Since our complex, abstract ideas are constructs of the mind, Locke believes that we have full access to them—nothing is hidden from us. Also, nothing in Locke indicates he believes species to be material things, and his discussion of the real essences of substances indicates they are material in nature. To attribute real essence to species we would need to resolve this difficulty. In addition, in his discussion of real essences, Locke makes use of causal language to indicate that the real essence is the source of the properties by which our species concepts are formed. Since the species concept (the abstract idea we have of a species) is the nominal essence of a species, by making the species the seat of the real essence Locke would then ensure that the nominal essence would be identical to the real essence in substances, something which Locke explicitly denies (III.iii.18). These considerations ought to make us consider that perhaps reading our original passage as attributing real essences to species is incorrect.

But if Locke is not claiming that species have real essences, what is going on in our original passage? I believe it is possible to interpret III.vi.6 in such a way as to minimize the complications sketched above and still remain faithful to the text and its context. The sections immediately preceding and following the passage in question are occupied with Locke’s attempt to demonstrate that individual parcels of matter do not have essential properties. For example, a piece of iron must possess the ability to be pulled by a magnet if it is to be iron, but there is no necessity for any parcel of matter to be magnetic; in other words, Locke is claiming that it is possible for bodies to have properties other than the ones they have (III.vi.4-5). The properties that individual objects have are then used by the mind to classify objects into kinds and assign nominal essences. The real essences of objects are not made use of in the assignment of species designations to individuals (III.vi.7-8). Essentially, the context in which Locke links species with real essences is an epistemic, rather than a metaphysical one. He is not here claiming that there are real species which exist independently of things that fall under them or of the minds which construct them, but rather that the properties which are referred to as essential to a particular thing are only intelligibly called essences in terms of the sort or species of thing they are taken to be an instance of. He is claiming that the only intelligible way to speak of properties as essential is to do so in the context of species concepts.

I do still want to affirm, however, that Locke intends to tie real essences to our species concepts in this passage. Locke believes that properties which manifest themselves when we observe a substance are caused by the real essence of the substance. The ideas we have of these properties are then combined into the complex idea that forms the nominal essence of the species of the substance; this is the link whereby the real essence is connected to the species. Locke says as much at the end of the section: “as to the real Essences of Substances, we only suppose their Being, without precisely knowing what they are: But that which annexes them still to the Species, is the nominal Essence, of which they are the supposed foundation and cause” (III.vi.6). Basically, species have real essences, but only mediately, through their possession of the nominal essence which is composed of ideas which are caused by the real essences.

This reading of Locke in III.vi.6 allows us to affirm that Locke does tie real essence to species in a deliberate way, as a face value reading indicates. However, by interpreting the claim as one which ties real essence to species through our epistemic or classification practices rather than as positing a real essence directly possessed by a species, we can read Locke without making him contradict himself or claim something incomprehensible.

Mental Causation II: Nonreductive Physicalism

I. A Version of Nonreductive Physicalism
The exclusion argument works because if token physicalism is true and microdeterminism is true, any causal properties possessed by mental events are possessed in virtue of their physical properties, not in virtue of their mental properties. For physicalists such as Jaegwon Kim (see also 1984, 1993a) there are ontological as well as epistemological consequences to the exclusion argument. The ontological consequence is that mental properties have no independent causal powers of their own. The epistemological consequence is that invoking mental properties in explanation is superfluous.

Robert Van Gulick (1993) draws two main conclusions from his discussion of the exclusion argument. First, he provides an argument to the effect that causal relevance is distinct from causal potency (basically, the idea is that the ontological and epistemological consequences of the exclusion argument are distinct). He argues that if the causal relevance of a mental property is determined by having a useful role in causal explanation, then it is not clear that we should count mental properties as causally irrelevant simply on the basis of their lack of independent causal powers (242-5). Discussion of Van Gulick’s view of causation will follow, but suffice here to note that this conclusion regards epistemology, and is distinct from any ontological consequences Van Gulick draws.

Second, he argues that although the events and objects picked out by the special sciences are composites of physical constituents, it is not the case that their causal powers are determined solely by the physical properties of the constituents of those organized processes; rather, the organization of those constituents into patterns also plays a role in the determination of causal powers. Those organizational patterns turn out to be the referents of the predicates of the special sciences (250). It is unclear exactly what these patterns are supposed to be, and in what sense they should be considered “higher order” than physical properties. However that is resolved, Van Gulick appears to be moving from his thesis about the relevance of higher order properties to causal explanations to a thesis about the causal efficacy of those higher order properties; he is drawing substantive ontological conclusions from his epistemological considerations. He argues for the reality and causal potency of the organizational patterns based on the following:

(1) Organizational patterns are recurrent and stable features of the world.
(2) Many patterns are stable despite variation in or exchange of constituents.
(3) Many patterns are self-sustaining despite physical forces that might disturb them.
(4) Patterns may affect which causal powers of their constituents are likely to be activated. The constituent may have many different causal powers, but only a subset will be active in a given situation.
(5) The selective activation in (4) may contribute to the maintenance of the pattern itself.

On this view, higher order properties act by selective activation of physical powers, not by their alteration (250-2).

Van Gulick argues that even the properties of the lowest levels of the physical organization of things are in fact “stable self-sustaining recurrent states of the quantum flux of an irreducibly probabilistic and statistical reality.” He gives as examples of these recurrent states the property of being a proton or the property of being an electron with ½ positive spin. The fact that the interactions of objects at the lowest level approximate deterministic regularities seduces us into believing that these objects and interactions have a privileged role in determining the organization of the world. Van Gulick sees the exclusion argument as weakened by these considerations in the following ways, then. The complete physical description of the world will have to include specifications of boundary conditions, since the higher order properties and organizations play a role in which causal properties of the constituents of higher order properties are active. But looking exclusively at the lower level constituents and their properties will not reveal the higher level organization. There are no complete translation functions from one level to the other. In addition, special sciences are able to refer to these higher order property instantiations and elucidate the way in which the temporal sequence of events is determined by the interaction of higher level properties (254-5). For Van Gulick this inability of lower level explanations to account for higher level phenomena does not appear to be merely an epistemic limitation; it is blocked in principle by the real causal powers of higher level phenomena.

II. The Problem with Van Gulick’s Response to the Problem

Part of Van Gulick’s objection is that the properties picked out by physics as the basic entities are every bit as abstracted as the higher order patterns and properties picked out by the special sciences. At bottom, how is the property of ‘being a proton’ categorically different than the property of ‘being a belief’? Both are abstractions, so ultimately they have the same ontological status on Van Gulick’s analysis. I am inclined to agree with him. But I think that where Van Gulick goes wrong is in his assertion that properties like ‘being a proton’ constitute actually constitute the lowest level of reality. This I think is incorrect. It is not the properties of protons, electrons, and whatever other entities make up the domain of quantum physics that comprise the basic level of reality (and therefore analysis) on the assumptions of the exclusion argument, but rather the protons, electrons, and other basic elements themselves that do. It is useful to talk about the properties of quantum entities in descriptions of their interactions, but those properties are abstractions, as Van Gulick points out. But the ultimate constituents of higher level properties are not themselves abstractions, but rather entities. If there is a distinction between the entities that occupy the lowest level and the properties that are abstracted over them, then Van Gulick cannot slide so easily from epistemology to ontology.

Without a parallel between physics and the special sciences, Van Gulick loses the analogy that provides his model for how higher order properties can have causal powers independent of their constituents. His argument hinged on being able to move from the epistemic viability of higher order properties to their viability as independent causal agents. If the objects of the special sciences are not in fact at the same level of abstraction as the objects of physics, then Van Gulick’s claim that the organized patterns of high level properties are independent of their constituents cannot gain purchase from physics being in the same boat.

It might be useful to look at an example of the sort of higher level process that is supposed to be causally independent and capable of downward causation. Van Gulick’s article is light on examples, but I think the sort of thing he has in mind is something like the developmental processes studied in biology. Developmental systems theorists often point to the interaction and complexity of developmental systems as an indication that genetic reductionism is a flawed perspective from which to analyze what’s going on in the production of phenotypes or other biological processes. Developmental systems are composed of a multiplicity of entities and processes and their interactions, including codons, noncoding DNA sequences, epigenetic and regulatory machinery, etc., each of which plays a role in the production of phenotypic effects. Some biologists argue that it is not practically possible to functionally decompose the developmental system into discrete parts and predict the behavior of the system on the basis of the properties of the individual constituents. Developmental systems theorists might also talk about the role developmental systems play in higher level processes; many evolutionary biologists consider the changes that take place in lineages of developmental systems to be a key (or even “the” key) unit in phylogenetic change over evolutionary time (Sterelny and Griffiths 1999: 94-100).

These then would seem to be prime candidates to serve as examples for Van Gulick. However, none of this implies that there are higher level system properties that exercise downward causal force to determine what is going on in the biochemical reactions that comprise the system. For each element of the system there is a mechanism that is, at least in principle, specifiable in terms of the basic constituents of the system and their interactions; there is no top-down pressure required to explain what is going within the system. There is broad agreement within biology about metaphysical reduction and the idea that higher level systematic properties and processes are determined by lower level, physical ones. Debate in biology about reduction is about the epistemological aspects of theory reduction and explanation (Rosenberg 2007: 120-121). Ultimately, it may not be possible to have a useful developmental theory that refers only to atoms and molecules, but that does not indicate that the pragmatics of explanation dictate the positing of higher level phenomena with independent causal powers. The pragmatic requirements of explanation do not dictate the ontological commitments of biology.

In the same way, nothing Van Gulick has given us should prompt us to abandon the metaphysical consequences of the exclusion argument. His failure to recognize a distinction between the constituents of higher level properties and the abstracted properties of those constituents has forced him to blur the boundaries between metaphysics and epistemology, and he has taken his epistemological conclusions to have substantive consequences for ontology. I suspect that despite explicitly recognizing that pragmatics should drive our explanatory frameworks, he fails to notice that the coarseness of the grain of our explanations does not have ontological consequences. What Van Gulick is seeing is the variety and complexity in the lower level systems which produce higher level properties that can be productively grouped into multiply realizeable kinds for the purpose of analysis and that the details of the physical systems that exhibit that complexity do not directly contribute to special science explanations and takes from that a real gap in the actual causal properties of the higher and lower levels. The differences he notices that seem to indicate an incongruity between the causal properties of the low level processes and entities that constitute higher level properties and those of the macrolevel patterns that are analyzed by the special sciences are in reality epistemological distinctions that need not indicate a real metaphysical divide. What Van Gulick has done is mistake a mismatch in grain of different types of explanations of one event or process for a true ontological divide that requires the positing of downward causation to make the world intelligible.

References
Kim, J. 1984. “Epiphenomenal and Supervenient Causation.” In Kim 1993, pp. 92-108.
Kim, J. 1993. Supervenience and Mind: Selected Philosophical Essays. Cambridge: Cambridge Press.
Kim, J. 1993a. “The nonreductivist’s troubles with mental causation.” In Kim 1993, pp. 336-57.
Rosenberg, A. 2007. “Reductionism (and Antireductionism) in Biology.” In D. Hull and M. Ruse (eds.) The Cambridge Companion to the Philosophy of Biology. Cambridge: Cambridge University Press
Sterelny K. and P. Griffiths. 1999. Sex and Death: An Introduction to the Philosophy of Biology. Chicago: University of Chicago Press.Van Gulick, R. 1993. “Who’s in Charge Here? And Who’s Doing All the Work?” in J. Heil and A. Mele (eds.) Mental Causation. Oxford: Clarendon Press, pp. 233-56.

Mental Causation I: The Exclusion Problem and Supervenience

Our commonsense intuitions and folk psychology point to the efficacy of mental properties. It seems that nothing could be more obvious than that things like beliefs and desires cause things to happen in the world. Analysis of the causal role of beliefs, desires, and other mental properties leads quickly to questions about this commonsense picture. Are beliefs and desires physical entities? Are they identical to neurological states? If they are not identical to neurological states, which should be considered the true cause of the effects claimed for mental causes, mental properties or brain properties? These sorts of considerations are at the heart of the exclusion argument put forward by Jaegwon Kim in a variety of versions over the last 30 years.

The basic issue that the exclusion argument attempts to elucidate is the following: if we are realists about mental causation and physicalism, then any mental event m that causes physical event p does so in virtue of m and p falling under mental kind M and physical kind P respectively. Does p have a physical cause as well as a mental cause? If p does not have a physical cause, in other words if m is a strictly nonphysical property, then m’s having causal efficacy is a violation of causal closure of the physical, which is at the heart of physicalism. Allowing nonphysical causes moves the discourse into Cartesian dualism with all the unsolved mysteries surrounding the relation of physical to nonphysical properties that accompany it. If we add another restriction to our interpretation of the causal sequence, that if p is a physical event it must have a physical cause, then the question becomes, if p has a physical cause, then what role would the purported mental cause play? The physical cause seems to exclude the mental cause (Kim 1998: 37-8).

So, what sort of options are there to deal with this exclusion?

I see four options for dealing with the problem, which I plan to deal with in later posts: reductionism, eliminitivism, nonreductive physicalism, and epistemic compatibilism. Here, though, I'll briefly follow Kim's argument showing why an appeal to supervenience is a nonstarter as a solution to the problem.

The claim that mental states supervene on physical states means that for any mental event m that instantiates mental property M at time t, there is some physical property P such that m has P at t and anything that has P has M. If supervenience fails, then the intelligibility of mental causation fails as well. The conjunction of supervenience with the causal closure of the physical provides the possibility of explaining mental causation within the general account of physical causation, but only if mental causation can be reconciled with causal closure and supervenience. Supervenience brings mental causation under the purview of physical causation. So mental causation without supervenience is unintelligible from the perspective of physicalism (Kim 1998: 37-41).

Unfortunately, mental causation is unintelligible if supervenience holds as well. Consider some mental property M that is purported to cause another mental property M*. In this case if supervenience holds, then M and M* each supervene on physical properties, say P and P* respectively. If this is the case then there are two potential explanations for why M* obtains, M and P*. This can be alleviated by claiming that M causes P*, but that solution only provides a new difficulty, namely, how to explain the downward causal powers of M. Or alternatively, we could claim instead that P causes P* granting causal power to M in a derivative or dependent sense. This however, brings us right back to the exclusion argument. Kim contends that the most natural way of viewing the relation of M to M* in this case is to say instead that P causes P*, M supervenes on P and M* supervenes on P*. On this view the mental properties depend for their existence on the physical subvenient properties, and all of the causal activity takes place on the physical level. The regularity between the mental properties is not an accidental one, though; there is a real causal regularity that explains the connection between these properties, and is able to do so without violating the causal closure of the physical or leading to overdetermination. But on the supervenience model the real causal action takes place at the physical level, not the mental level, and thus supervenience doesn't really offer an alternative that avoids the exclusion problem (Kim 1998: 41-7).

Kim, J. 1998. Mind in a Physical World. Cambridge: MIT Press.

I Put the "Super" in Supervenience

So, I'm taking a course in contemporary metaphysics this term, and I feel a bit like I've wandered into a foreign land of nonsense and speculation. For the most part, I am interested in the history of philosophy or philosophy of science and so my study of philosophy has to this point been fairly bereft of bizarre arguments about possible worlds, multiple realization, and some of the other arcane theoretical entities and relations that stalk the pages of philosophy journals.

I think I prefer it that way. We're into week 3 on "supervenience" and I still have no idea what that term is supposed to mean. The slogan version is simple enough: if property A "supervenes" on property B, then any two entities that are exactly the same in respect to their B-properties will be identical in respect of their A-properties. If mental properties supervene on neurophysiological properties, then any two identical brains will have identical thoughts.

Ok. So far, so good. The problem comes when you try to say anything further. Saying that mental properties supervene on neurophysiological properties doesn't tell us much at all about either set of properties, and it tells us absolutely nothing about the relation between the two sets of properties.

The way I read it, a supervenience claim is either an uninformative claim that the two sets of properties show covariance or an incomplete claim that one set of properties depends on the other. Either way, just saying that A supervenes on B is not saying much.

Maybe I don't really understand the philosophical import of this notion. Let's hope so, because otherwise it's going to be a long quarter.

Universal, not Universal

In an effort to start my blog by jumping right into a hot topic everybody is talking about, I’ve decided to write about two conflicting twenty year old solutions to a logical difficulty in a 2400 year old book.

The problem is this: in Aristotle’s Metaphysics vii, Aristotle says things that lead one to the logical conclusion that substance is a universal, then turns around and claims in vii.13 that substance is not a universal.

Daniel Graham (1987) and Frank Lewis (1991) offer different solutions to this problem, which I think are interesting in a couple of regards.

Graham’s identification of the problem is part of his thesis that the ontological framework of the early work of Aristotle in the Organon is “incommensurable” with the ontology of the later works (he calls the two “systems” S1 and S2 respectively).

Here’s how Graham sees the problem. After arguing that species is essence which is form which is substance, and despite the fact that his conception of form and essence require that those be universals, Aristotle argues in vii.13 that universals cannot be substance (Graham 1987: 252). He applies the substance criteria to the concept of the universal, and determines that the universal is ruled out as potential candidate for substance (Graham 1987: 253). Aristotle recognizes this difficulty at 1039a14-23 (Graham 1987: 254).

Graham argues that at this juncture in Aristotle’s argument it would make sense for the philosopher to jettison the requirement that substance cannot be a universal, but instead Aristotle argues the opposite (Graham 1987: 255). Graham’s diagnosis of the source of the contradiction of substance as universal and not universal is that the requirement that substance not be universal is an out of place holdover from S1­, which has no grounding in the S2 ontology (Graham 1987: 261). Graham takes this as support for his Two Systems Hypothesis.

Lewis seems to express broad agreement with Graham on the nature of the consistency problems of Metaphysics vii, but views the problem as resolvable internally, using the conceptual resources of the Metaphysics itself.

The problem as Lewis sees it is the following (310-11):

For primary substance, Aristotle is committed to the thesis that for any x, if x is a substance then x is not a universal, or, formally:

(x)(Substance (x) → ~Universal (x))

But Aristotle is also committed to:

(x)(Form (x) →Species (x))
(x)(Species(x) → Universal (x))

And therefore he must hold that (x)(Form (x) →Universal (x)) by transitivity and modus ponens.

But given that Aristotle also holds that (x)(Substance (x) → Form (x)), he must hold that (x)(Substance (x) → Universal (x)).

So we have a contradiction. A logical consequence of his views on form, species and universals requires that substance be a universal, but his explicit view is that substance cannot be universal.

Graham contends that this contradiction is irresolvable, and that Aristotle’s problem is generated by carrying over the “non-universal” criterion for primary substance from the S1 ontology to the S2 ontology, and that the solution would be to discard this assumption. Lewis’s solution to this contradiction, by contrast, is one that offers Aristotle a way out by modification of the way substance is predicated. Instead of being a one-place predicate, whereby some x is a substance simpliciter, in Lewis’s view the terms “substance” and “universal” come to stand for two-place, relations in the Metaphysics ontology. Substance is “substance-of” and universal is “universal-to”. Something can be universally predicated of another only if it is not the substance-of that very same thing.

Formally this is expressible as

(x)(Universal-to(x, y) → ~Substance-of (x, y))

Thus a form can be universal-to a particular bit of matter, so that, Universal-to(F, m); in this case the form would not be the substance of the matter, but would instead be the substance of a compound particular, c so that Substance-of (F, c). In this way, Aristotle can avoid the contradiction by saying the non-universal requirement is simply modified as relations of forms to different entities. Since the compound is not identical to the matter, the form Man could be the substance of Socrates, and predicated of (universal to) Socrates’ matter.

The solution is appealing, since it would allow us to avoid attributing what would appear to be an egregious and obvious contradiction to Aristotle, but it does have the consequence of eliminating the prospect of being able to fully reconcile the ontological framework of the early metaphysics of the Categories with the later ontology of the Metaphysics. In the earlier work, Aristotle allowed as primary substances only those entities that become hylomorphic compounds in the later work; it is particular men or horses that have primary claim to substance-hood, but in the later work, those entities have something which is more primary predicated of them as the “substance-of” them. So it seems that even Lewis’ solution is at least potentially supportive of Graham’s hypothesis.

One curious aspect of this solution, however, is that if the substance-of relation is to be considered as a distinct metaphysical relation, then if F is the substance-of something, then it is the substance-of something of which F is itself already a component. I have difficulty making sense of this. Normally we think of relations as between two distinct things (with the exception of the identity relation) so I’m unclear how this relation should be conceived.