The following link is to a dialogue I wrote up on Lockean substratum. You have to click on the link to see the PdF file. I wrote it out long hand in a cafe since I have been experiencing technical difficulties with my laptop.
Opinion on the notion of Substratum
Opinion: better alternatives than Aristotelian substratum may be found to ground the existence of the attributes of substances. One such is presented here, which is put forward as logically consistent and one that satisfies both the demands of Aristotelians and their Empiricist opponents.
Before Tina Turner sang, “We don’t need another Hero” Locke wrote in a similar way about the non-necessity of Aristotelian substratum. Why did people think that everyday objects needed a substratum to be sufficiently groundedontologically? Well, the reasoning can be taken as proceeding like this:
When we look at everyday objects, we can see that they have certain attributes, such as a certain height or color or weight.
These attributes can be said to have a certain order of dependence one upon another: for example, an object has a color only if it first has extension in space. Only if an object is extended in space, having at least two dimensions, it might be said, can it thereby have a color. This is true even in one’s imagination. Necessarily, if an object has a color it has a height and width, if not a depth and weight at a certain time.
But, the advocate of substrata would claim, extension is not enough to establish an object as a real entity by itself. Being a mere attribute of an object it cannot exist on its own.
The justification for #3 is related to #2 in the sense that it involves ontological dependence. This time the dependence is spoken of in terms of “inherence”. The argument goes as follows: (a) the qualities of an object such as its colors, which are a type of accident (see #1) cannot exist on their own without belonging to or inhering in a substance; (b) in the same way, the quantitative accidental attributes of a substance, such as its height or weight or length cannot exist apart from an object. Moreover, (c) the qualities and quantitative attributes of an object make up the sum total of an object’s internal attributes (i.e. those that are not relative and depend upon its internal attributes). But (d) all such attributes must inhere in something since they cannot exist on their own (see a-c). Hence (e) all such attributes must depend upon a further something that might ground their existence. That further something is what the Aristotelian would call a substratum.
The claim that this argument can be taken as a synopsis of the Aristotelian view on “Substance” perhaps needs some defense. This can be presented briefly, but I think conclusively, as follows.
The word Aristotle uses that is usually translated as “substrate” in the Categories is to hypokeimenon, or literally, “the underlying (something)”, which is often translated as “subject” (as in the grammatical subject of predicates). It should be seen at the outset that the term does double duty in the Categories as a term for a grammatical subject (as a subject for predicates) and for a subject in the sense of substance: e.g. Aristotle writes in Chapter 2 of the Categories, “By being ‘present in a subject’ I do not mean present as parts are present in a whole, but being incapable of existence apart from the said subject.” By introducing the notion of accidents as existing in a substance (the Greek word here is hypokeimenon as above) Aristotle is in effect shifting the ontological ground of his discussion from the consideration of the role of a subject as a bit of grammar to subject as a countable entity.
The effect is such that the discussion of the hypokeimenon-subject of the second chapter of the Categories naturally bridges over into Aristotle’s further discussion of ousia-subject in the fifth chapter. It is significant that Aristotle has an entitative sense of “subject” in mind in the second chapter because it lays the foundation for a dualism between substance and accident that follows from his claim (quoted above) that accidents cannot exist on their own (chapter 2), although a substance can (implicit in chapter 2 where he identifies subjects as unitary things such as individual horses and men in contrast to things that are present in a subject, and explicit in chapter 5 where Aristotle says that the ability to subsist is one of the primary differentiating characteristics of a substance).
The non-reducibility of accident to substance follows from this duality: one subsists while the other merely “exists in” (see 4 a and b above). But this in turn yields the further result that no collection of accidents can yield a substance. This reasoning is, quite arguably, implicit in Aristotle’s distinction between accidental and substantial change in the first Chapter of his lectures on Physics. It is this non-reducibility that provides the basis for the Lockean critique of Aristotle’s notion of substance.
It is, of course, fairly well established that Locke, in the end, decides to keep substrata in his inventory of “the things that are”. But he does so only after presenting a very compelling empiricist counterargument that has left many wondering whether he could actually be serious about maintaining that there are such things as substrata in the universe. His counterargument can be related to the dualism between substance and accident mentioned above.
Essentially, Locke’s argument works on the irreducibility of accident to substance that underlies that dualism. In Book II, ch. xxiii of his “Essay” Locke points out that if it were asked wherein qualities (secondary qualities) inhere in a substance one could answer that they inhere in the quantitative features of an object (in Locke’s terminology, their primary qualities). But if one were to press further and ask wherein the primary qualities of an object inhere, one could only answer that it must be an unknown something-something the senses do not perceive. It can be seen here that Locke is attempting to be a consistent empiricist in claiming that the notion of substratum is unfounded: since it cannot be sensed, it violates the basic criterion for what any good empiricist would accept into his or her ontology. This type of argument might be called the “Indian argument” for short, since it has been referred to as such in the secondary literature.
Thus, the issue substratum theory poses comes to this: one can either say with Aristotle that a substratum is necessary as a ground or support for things that cannot exist on their own, or one can side with Locke’s empiricist argument against substrata (the indian argument) in saying that no such entity can be admitted into one’s ontology on empiricist grounds. Moreover, to admit substrata into one’s ontology is implicitly to move from a monistic position (materialism or phenomenalism) to a dualistic one (materialism plus an immaterial entity): since only “qualities” in Locke’s terminology, or “accidents” in Aristotle’s can by their very nature, be perceived, substrata, or substances without attributes must be taken to be imperceptible by their very nature. Hence it follows that for the substratum theorist, there must be at least two basic kinds of things: those that are perceptible and those that are imperceptible by their very nature. In admitting substrata into his ontology, Locke basically affirmed his commitment to a kind of dualism that in many respects mirrored Cartesian dualism, and for similar reasons (e.g. non-reducibility), has engendered dissatisfaction ever since.
The argument I would like to advance in response to this issue is that it is unnecessary to suppose substrata as a ground for accidents/qualities. This position prompts two lines of inquiry: firstly, can the quantitative features of an object exist on their own or do they require something wherein they may exist? I.e., how does one overcome the common sense view that says things like a particular height or weight must inhere in an object in order to have any reality? Does this not inescapably prompt the need to suppose the existence of a substrate wherein their existence may be grounded? Secondly, if the quantitative features of an object can somehow be taken to be the fundamental ground of being of an object and this in turn implies either a materialistic or phenomenalistic monism, can such a monism do all the metaphysical work the alternative dualism with its supposed substratum was able to do? An affirmative answer to the latter question can be justified through resolving the problems posed by the first line of inquiry: basically, a satisfactory monism will be able to do all the metaphysical work that the alternative dualism was able to do without supposing a fundamentally imperceptible substratum. Since substrata have this characteristic, they would seem to deserve the title of “occult entity” as much as any others of the same ilk and their elimination would seem to be desirable from both an epistemic and an empirical point of view if one happens to be committed to empirical principles in drawing up the inventory of one’s ontology. Let’s proceed then to see what can be done to eliminate substrata.
Simply put, the key to eliminating substrata is to let the term “substance” indicate an entity that essentially includes all its attributes (or qualities or accidents) in its concept. This might be taken to be a Leibnizian view of substance. The next move is to see that, taking on the idea of the attributes of a substance as fundamentally dependent beings, they imply a substance. Thus, we are left with the following set of logical relationships between substances and their attributes: necessarily, if there is an attribute, then there is a substance; but, moreover, only if there is a substance can there be an attribute. If one wants to know wherein an attribute exists, it may be answered that it exists in or belongs to a substance; but only if there is a substance may there be an attribute at all. The latter justifies the idea that substance can function as the ground of existence for attributes, while the former affirms the basic desire to say that attributes must exist in something other than themselves.
This view of things fits perfectly well with any empiricist/phenomenalist view of perception that begins with particular percepts as a basic epistemic starting point. Consider an object; let’s say, a book, for example: the book may be considered a substance, while its color or height may be considered as belonging to it as a substance. The substance in this case is an independent entity but not one that is devoid of attributes. Indeed, as above, the existence of substance can be taken to pre-suppose the existence of attributes in a non-circular manner. Here again, the perceptible phenomenal attributes belong to the book as to a substance that grounds their existence; but that substance in turn sufficiently implies the existence of at least one attribute.
This may seem rough and ready given the long standing persistence of this metaphysical issue. But it should be enough merely to ask the reader to consider the logic that prompts the series of moves presented above. Consider that in posing the definition of substance presented here (seemingly out of thin air?) as an alternative to substratum theory, what has been shown is the non-necessity of substratum theory, provided that the version of the relationship between substances and their attributes is not self contradictory and can be said to be basically sound. The two paragraphs immediately above can be taken as answering to both of those requirements.
The desire to reify substance, to make it into a separate entity apart from its attributes, seems to follow from a mistake in thinking that whatever substance is, it must be something independent not only from other substances but also from its own attributes. It might be thought, for example, that since a substance may undergo a change in its attributes (for example changing from short to tall) it must be independent from its attributes. But this assumption is non-necessary from a logical standpoint: it may be said, alternatively, that substance sufficiently implies the existence of at least one attribute while the existence of an attribute necessarily implies a substance. Moreover, the irreducibility arguments above for an independent substance need only lead to the conclusion that a substance should be distinguished from its attributes from a conceptual standpoint: clearly, attributes must be fundamentally different from substances in some way, but that need not imply a total independence in re from one another. The above way of construing their inter-relationship allows for conceptual dependence on a logical basis with greater fidelity to the epistemological basis for that dependence than substratum theory offers.
In conclusion, the definition of substance presented above allows for a way to conceive of substance that satisfies the demand for both irreducibility required by Aristotelians and other substance dualists and the epistemological grounding in perception sought by phenomenalists or materialists. It makes the two compatible by simply distinguishing conceptual independence from interdependence in re in a way that is logically valid and, arguably, sound. As a viable alternative, it argues against the need to construe substances as substrata.