The Worst Argument in the World: A Defence

David Stove was wrong to dismiss idealism as legitimate philosophy

The Australian philosopher David Stove was a real prick, at least in writing. Evidence: check out his article, “The Intellectual Capacity of Women”.

Since, as tradition teaches, evil is the mere absence of good, and nothing is entirely evil, there must be some good in Stove’s prickhood; I think the good lies, first of all, in providing a useful illustration of what a closed mind looks like, and also in the entertainment provided by the snarling and overconfident attacks he liked to launch.

One of the most entertaining is found in his two essays “Idealism: A Victorian Horror-story (Parts One and Two)”. These are published in a book called The Plato Cult, which is largely a collection of ad hominem attacks. Rather than spending much time examining the proofs or disproofs of various philosophical positions, Stove engages in amateur psychology to pathologise people who held views “of which [his] opinion is low”. He justifies this ad hominem approach in the Preface on the grounds that the philosophy in these cases is so bad that the only possible explanation for its espousal must lie, not in any rational justification, but only in the personal failings of the espousers.

In the case of idealism, Stove finds the personal defect is a longing for the comforts of religion, long after Darwin and other scientists had (in his opinion) cut away any reasonable basis for it. Idealism holds that there is no reality independent of mind or spirit, thus implying that the entire universe must be suffused with spirit or overseen by mind, or even that the ordinary things of the world are in fact the mere appearances of some Great Mind called the Absolute.

Stove might be onto something here; in the seventh edition of his work Elements of Metaphysics, the British Idealist A.E. Taylor admitted the following:

The one thing of all others I have had it long on my conscience to say, is that I have always wished my book to be understood in a definitely theistic, indeed in a definitely Christian sense. I have never disguised it from myself that when I speak of the “Absolute” I mean by the word precisely that simple, absolutely transcendent, source of all things which the great Christian scholastics call God. I would add that when, following the tradition of my own teachers, I speak of the “creatures” as “appearances” of the Absolute, I mean by this precisely what St Thomas, for example, meant by the doctrine that they have being be “participation”.

Stove, however, suggests that it is only this doomed yearning for a defeated religion that explains the prevalence of idealism in the nineteenth and early twentieth philosophy, which he shows to extend far beyond crabbed old Victorian professors, embracing figures such as the mathematician Henri Poincaré. It must have been the yearning, Stove reasons, because the primary argument advanced in favour of idealism is “so bad that it is hard to imagine anyone ever being swayed by it”

In fact, Stove singled this argument out as the worst argument in the world, baptising it with the sarcastic epithet “the Gem” (scholars today know it under Andre Gallois’s perhaps more benign “Master Argument”). It is found originally in Paragraph 23 of Berkeley’s The Principles of Human Knowledge. Berkeley advanced three arguments for idealism, but “it was the Gem, virtually unaided, which made idealism the philosophy of the Western world for more than a hundred years. It was the ugliest toad of them all who got the princess and the kingdom.”

In the original form from Berkeley, the Gem reasons that it is impossible for any thing to exist outside or independent of the mind, since the attempt to give any example of such a thing—say, a tree deep in the forest, unthought by anyone—leads immediately to contradiction. If you are offering it as an example then you must be thinking of it. So it isn’t independent of or outside the mind after all. If the very concept of an object unthought of by anyone entails a contradiction, then there cannot be any such object. All objects must be thought of. Thought pervades the world.

Stove sees this as a mere non-sequitur. Its one premise is:

You cannot have trees-without-the-mind in mind, without having them in mind.

And its conclusion is:

You cannot have trees-without-the-mind in mind.

I admit to not feeling the ridiculousness of this argument as immediately as Stove seems to expect. If I have “trees-without-the-mind” in mind, then it seems that they are not trees without the mind—they are in my mind. So the attempt to have them in mind leads to a contradictory concept: trees without the mind that are in my mind. If the attempt to have trees without the mind in mind leads to contradiction, then the thing must be impossible. I cannot have trees without the mind, and nor can you or anyone else. What cannot be conceived of cannot be asserted to exist, not even possibly.

Stove needs to work harder to establish the silliness of the argument, but he doesn’t really; he goes right on to his pathologising in full confidence that you already agree that nobody could really have been swayed by it..

Later, he has another go, treating the argument as analogous to the argument that you can’t eat an uneaten oyster, since if you eat it is isn’t uneaten. This isn’t much help, since there is a clear difference in the cases. You can, of course, eat an uneaten oyster by making it into an eaten one, and you can conceive of this possibility because there is no contradiction in the concept of an uneaten oyster. But there is, according to “the Gem”, a contradiction in the concept of an unthought tree. If the concept is not empty then its object—the unthought tree—must fall under it, and an object falling under my concept cannot be unthought. Or, alternatively, the object must fail to fall under its concept in order to fall under it. No object can perform such a contradiction, and so there must be no such object.

A.N. Prior does a better job of trying to explain where “the Gem” goes wrong. But here the ascribed error is too subtle to support Stove’s belief that nobody could really have taken the argument seriously. He accuses Berkeley of equivocating between:

1. x is imagining truly that there is something which is not-thought-about

and:

2. There is something which x is imagining truly to be not-thought-about.

Prior holds that 2 is a logical falsehood; it is simply contradictory to say that something that x is imagining is not thought about. But 1 is perfectly acceptable. The subtle point here is that x can imagine that there is something which is not thought about without thinking about something that is not thought about.

Prior holds that Berkeley’s equivocation here is akin to what Aristotle called the “fallacy of composition”, e.g. reasoning from: “he is not able to write-when-he-is-not-writing” (which is true because writing-while-not-writing is a contradictory and impossible activity) to: “he is not able-to-write when he is not writing” (which is often false, since many who are not currently writing are nevertheless able to write). The first is like 2 in the above pair of propositions; the second is like 1.

I don’t find this analogy so useful, since, in terms of logic, the “fallacy of composition” involves a modal operator. It involves confusing “<possibly> he is writing and he is not writing” with “he is writing and <possibly> he is not writing”. The case of 1 and 2 doesn’t seem to obviously involve modal operators.

In fact, Prior’s argument involves devising an operator “__ is thinking about __”, which is in part a propositional operator like a modal operator, since the second slot is to be filled with a proposition, just like the operator “<possibly> __”.

But there is an awful lot of logical baggage with this. Prior holds (you can find the clearest arguments in Objects of Thought) that not all contentful thoughts require objects in the traditional sense. This is why Prior holds that one can imagine that there is something which is not-thought-about without having to think (contradictorily) about some thing which is not thought about. What one is thinking about, in this case, is no object at all but rather the proposition: “something is not thought about”.

But what is the subject of that proposition—isn’t it the contradictory something not thought about? To avoid that, this would have to be an unusual case, in which the subject of one’s thought is the predicate: “__ is not thought about”, while the predicate is a higher-order predicate: “__ is satisfied by some object” (but we are not thinking about which object). In other words, when someone thinks that there is something that is not thought about, the logical form of this thought is: “the predicate ‘__is not thought about’ is satisfied”. Only in this way can one have no object in mind when thinking that there is something that is not thought about.

This line of thinking derives from Frege’s innovation in analysis, which is to regard quantifiers as higher-order predicates, so that “some x is F” is not to be analysed as predicating F of the subject “some x” but rather as predicating a higher-order predicate “__ is satisfied” of F.

This analysis is not uncontroversial, and I think that the old Victorian idealists might have found much to justifiably object to in the notion of taking a predicate as the subject of a thought. Also, you might say that when you think that the predicate “__ is not thought about” is satisfied, you must implicitly be thinking about the whole domain of objects that might satisfy it (to use a quantifier without specifying, and thus thinking about, its domain leads to all the logical difficulties of unrestricted generality). But then, as predicted by the Gem, the very attempt to think the thought makes it false.

In any case you have to take a long and complicated path into the logical analysis of thought to arrive at this possibility, as a way of avoiding the conclusion of “the Gem”. It seems like bad history of philosophy to suppose, as Stove does, that only an irrational religious yearning could explain why anyone was swayed by an argument whose conclusion can only be blocked by a highly technical and controversial analysis of higher-order predication.