Eliminativism: the elusiveness of the ordinary

There is a deeper puzzle about eliminativism.

(E) There are no A's. There are only B's.
(R) There are A's but A's are only B's.

Clearly (E) and (R) do not disagree about the basic ontology. They agree that there are only B's. But they fundamentally disagree about the definition of 'A'. The eliminativist (E) claims that an A, as the term 'A' is correctly and properly understood, cannot exist, because its definition would include features inconsistent with being a B. The reductivist (R) is saying that, as the term 'A' is correctly and properly understood, it is entirely consistent that an A can be a B, indeed that every A is a B.

But if it is merely a quarrel over definitions, why is there any disagreement at all? There is no disputing over definitions. Perhaps the answer lies in the difficulty that surrounds all philosophically interesting notions. The SEP says "Like many philosophically interesting notions, existence is at once familiar and rather elusive. Although we have no more trouble with using the verb ‘exists’ than with the two-times table, there is more than a little difficulty in saying just what existence is". That is, there are certain terms which we all understand and have no difficulty using with a standard sense in everyday life, but which we find terribly difficult to define. Hence there may be profound disagreement over which features are essential to the term, and hence profound disagreement between (E) and (R). Both agree that in using the term 'A' they are talking about the same kind of thing, and using the term in the same sense. But they disagree about what are the fundamental features of an A. The eliminativist believes that there is some feature of A's, correctly understood, that makes it inconsistent with an A being B. And since he believes there are only B's, he holds that there are no A's. The reductivist agrees that this feature is inconsistent with being B, but regards it is non-essential, and so it is possible - indeed true - that no A actually has the feature.

Considering the example of truth - which is as philosophically interesting as any - we have

(E) There is no truth. There is only warranted assertibility.
(R) There is truth, but truth is only warranted assertibility.

The disagreement here does not involve equivocation (as I previously thought). Both (E) and (R) both think they are talking unequivocally about the same thing: truth, an idea that we have no more trouble in employing than in using the two-times table. Both agree on the 'ontology': there is warranted assertibility and nothing more. Where they disagree is that (E) thinks that truth involves more than warranted assertibility, and is inconsistent with the ontologyl.(R) by contrast thinks that truth involves no more than that, and so its existence is consistent with the ontology.

So the disagreement is not about ontology or about which entities/features are to be eliminated. Both agree that 'truth involving more than just warranted assertibility' is to be eliminated. But (E) eliminates it by eliminating truth itself. (R) eliminates it by holding onto truth, but eliminating anything more than just warranted assertibility. The disagreement lies in the analysis of the everyday notion of truth, and not 'ontology'.

The deeper puzzle is how people can agree on the meaning of a term, and yet disagree about its definition. How can we agree on the meaning of 'existence' - a term which is no more difficult to use than a times-table, and yet disagree profoundly on how to define it? Why are there similar puzzles and disagreements over the nature of truth, individuation, identity, reference and all the rest? This problem goes back at least to Plato, and we appear to be no closer to solving it.

Reducing Hesperus

Does a reductivist or 'identity' theory have to be eliminativist? Perhaps not. The theory could simply be asserting an identity. Take an early scientific discovery: the heavenly body that appears in the morning (Phosphorus) is identical with the heavenly body that appears in the evening (Hesperus). Is the theory that Hesperus = Phosphorus a reductivist theory? Nothing has been eliminated. Hesperus has not been eliminated because Hesperus is identical with Phosphorus, and Phosphorus has not been eliminated. By equal reasoning, Phosphorus has not been eliminated. So nothing has been eliminated. But is the theory reductivist? It seems to be lacking the asymmetry that we expect of a properly reductivist theory. It seems reasonable to assert that mental states are really brain states. But if this were merely an identity statement like Hesperus = Phosphorus, it would be just as reasonable to assert that brain states are 'really' mental states. Which seems odd.

On artefactual identity

Following some severe strictures by Brightly on my earlier posts I offer reformulations of the 'aporetic set' of propositions, as follows.

A. If an artefact X1 has a replacement component, and X contained the replaced component, and the other components of X1 have been components of X since the replacement, and if the arrangement of components is the same in X1 as in X, then X1 = X.
B. If the components of Y are now the components of X, similarly arranged, then X=Y
C. If Fx and x=y, then Fy
D. If it is not the case that components of X = the components of Y, then it is not the case that X = Y.
E. Artifacts exist.

The only notion that may need elucidation is 'plural identity'. I propose the following definition: the x's = the y's if every one of the x's = some y, and every one of the y's = some x, which uses singular identity. Note I have made a significant change to C. That is because substitutivity is more fundamental than transitivity, and because (I think) we can derive transitivity. For let x = y, and y = z. The identity x = y allows us to substitute 'x' for 'y' in 'y = z'. Then x = z, which is transitivity.

We can then prove inconsistency as follows.

1. Suppose there are an X1 and an X such that A is true. Then X1 = X.
2. Suppose next that there is an X2 such that A is true regarding X2 and X1. Then X2 = X1, and from C above, X2 = X.
3. Repeat this process, ensuring that each replaced component is always one of the original components (and not a replacement component). Then, repeating this n times for the n original components, we have Xn such that Xn = X. Note that while all of the components of Xn are components of X (because of the identity, and because of C above), it is nonetheless true that at one time none of the components of Xn was a component of X. Indeed, it is true (again because of C above) that none of the components of X was at one time a component of X.
4. Take all the original components of Xn/X and reassemble them so that they are arranged exactly as they were when they were part of Xn/X, to make Y. Then (from B above) X = Y.
5. But from D above, since the components of Xn/X are currently different from those of Y, it is not the case that X = Y.
6. Thus X = Y and not X=Y. Contradiction.

Is Hume an eliminativist about objects?

Hume's philosophy is more radical than we imagine. He did not hold, as Locke held, a representative theory of perception such that our sense impressions are fleeting and perishable, but are representative of objects which are relatively permanent and durable. He held that, when we analyse and examine the world carefully, we find that the things we call tables, chairs and houses really cease to exist when we are not perceiving them. For such things are really identical with our perceptions or sense impressions. He says* "it is impossible for us distinctly to conceive, objects to be in their nature any thing but exactly the same with perceptions". But "it is a gross illusion to suppose, that our resembling perceptions are numerically the same; and it is this illusion, which leads us into the opinion, that these perceptions are uninterrupted, and are still existent, even when they are not present to the senses. " Only philosophical analysis can penetrate and expose this illusion. Even that analysis can fail. The illusion is so strong that as soon as philosophers have discovered it - namely discovered, like Locke and the seventeenth century philosophers of perception, that our sense-impressions are fleeting and impermanent - that they immediately invent a new set of impressions that have the required permanence. Thus the representative theory at once refutes and re-establishes the illusion.

[...] it is liable to the same difficulties; and is over-and-above loaded with this absurdity, that it at once denies and establishes the vulgar supposition. Philosophers deny our resembling perceptions to be identically the same, and uninterrupted; and yet have so great a propensity to believe them such, that they arbitrarily invent a new set of perceptions, to which they attribute these qualities.

Does this make Hume an eliminativist, or a reductionist? I have argued elsewhere that the distinction is arbitrary, and I shall argue that this applies to Hume's position also. If we define 'material object' as something which is mind-independent and permanent, then it is clear Hume is denying the existence of any such things. The only objects we are aware of, he says, are these fleeting and perishable sense-impressions, which have no continued and uninterrupted existence. So he is an eliminativist regarding material objects defined in this way. But as I have argued, we don't have adopt this definition. If we define 'material object' as something identical with our sense-impression, then uninterrupted existence turns into a mere accidental feature of objects. An accidental feature that, as Hume argues, may not apply to any object at all, just as 'carried by the ether' does not apply to light, as people once thought.

In summary: whether Hume is an eliminativist or reductionist about the term 'material object' depends entirely on how you choose to define the term. There are the observable phenomena - the sense impressions - and there is whatever unobservable X explains or underlies these phenomena. And that X can have any features you like. There are no 'essential features' of things that are essentially unobservable.

* A Treatise of Human Nature, Book I. 4. ii - "Of scepticism with regard to the senses". This section is essential reading for any understanding of Hume. People often don't read it because it occurs towards the end of the first book, and because there is a lot of focus on the causation stuff in Part III. Part IV, particularly sections 2-4, are by far the most interesting and enjoyable and indeed strange parts of the work.

Radical versus conservative ontology

Before I move on to discuss the other assumptions that generate the conclusion that there are no artefacts, I realise there is something else to discuss. This is the distinction between what Steven Savitt calls 'ontologically conservative' or 'retentive' theory change, or reductionism, and 'ontologically radical' or 'eliminitavist' theory change, or eliminativism. This is explained well here. Ontological conservatism is when we revise our view of the entities posited by the original theory, without eliminating them. For example, our conception of light was transformed by the discovery that it is electro-magnetic radiation, and older explanations of it had to be rejected as incorrect or incomplete or misleading. But this discovery did not lead to abandoning the existence of light. Ontological eliminativism is when we drop completely the view that certain entities exist. For example, there is nothing in modern psychology that justifies the existence of malevolent spirits or demons. So we dropped demons altogether from our modern scientific ontology.

This is closely connected with what I discussed here about Inwagen's paraphrase argument. If his paraphrase of a story about ships is intended merely to change our view about ships (e.g. that they are not identical with their component parts, and that there is something singular over and above the parts corresponding to the grammatically singular 'a ship') then the paraphrase is ontologically conservative. It still affirms the existence of ships, just as the electro-magnetic theory affirms the existence of light, but it changes our view of what ships are, just as the electro-magnetic story changes our view about light. By contrast, if the paraphrase really 'loses ships', then it is ontologically radical. It challenges our common-sense ontology in a way that modern psychology challenged demons.

It is also connected with the 'aporetic' set of propositions I discussed earlier. (I have modified proposition B to meet an objections made by earlier comments).

A. An artifact remains numerically the same if one of its components is replaced, without replacing the others.
B. If the components of X, are now the components of Y, arranged in the same way, then X=Y
C. Identity is transitive (if X=Y and Y=Z, then X=Z)
D. If artifact a and artifact b have numerically different components at the same time, a and b are numerically different.
E. Artifacts exist.

If the four propositions A-D are true, artefacts do not exist, which is the ontogically radical position. Otherwise, one of more of these four propositions is false. This would revise our view about the fundamental properties of artefacts, which is ontologically conservative. This is by way of background. I will return to the original plan tomorrow, when I will discuss whether modifying our account of numerical identity could resolve the logical inconsistency between the five propositions above.