London Ed wants to pin the 'existential fallacy' on me. He writes and I respond in blue:
I am trying to locate the place where we have the real disagreement. It always seems to be about whether a certain type of thing exists or not, rather than basic assumptions. We are agreed about the 'phenomenology', i.e. the truth of sentences like 'Bill is thinking of a unicorn' or 'Peter has an imaginary girlfriend'. And we are agreed (I think) that you are not privileged in the sense that you have access to intuitive or revealed truth or empirical observations that I don't have, such as knowledge of intentional objects, or of their existence. So it remains that our dispute is semantic or logical. There is a disagreement about your move from premisses that we agree upon, to conclusions which we don't. So the disagreement is either semantic (i.e. one or the other of us or both is misinterpreting the meaning of premisses or conclusion) or logical (we both agree on meaning, but disagree on the validity of the move).
I think we both agree that a sentence like (A) below is or could be true
(A) Tom is thinking about a unicorn
We also agree on some or all of the existential premisses below
(B1) Unicorns do not exist
(B2) Nothing is a unicorn
(B3) There are no such things as unicorns
(B4) The world does not contain unicorns
I classify the premisses as follows. The first is properly 'existential' i.e. it contains the verb 'exist'. The second two are categorical sentences with explicit use of the copula 'is'. Christopher Williams claims they are really identical, because Anglo-Saxon doesn't like the verb 'is' beginning a sentence. The fourth doesn't use the copula 'is', although some logicians claim that it embeds it, and is really equivalent to 'the world is such that it does not contain unicorns' or 'the world is a non-container of unicorns' or 'no unicorns are contained-in-the-world'. The fourth also introduces the problematic concept of 'world'. I read it as 'domain of quantification', but that is problematic too.
Anyway, I imagine we agree on at least some of the B-premisses - it would be helpful if you could indicate which, in your view. (Obviously I agree with all of them, as a thin-theorist who believes in the univocity of the existential concept).
BV: I think that there are some further assumptions that Ed is making that need to be clearly stated:
(C1) For any x, x exists iff x is both extramental and extralinguistic.
This is not intended as a definition of 'x exists' but as a clarification of how Ed is using 'exists' and cognates. We could call (C1) the Independence Criterion: the existent is independent of (finite) mind and language.
(C2) It is not the case that some items are nonexistent.
We can call this assumption Anti-Meinong. Alexius von Meinong famously and very controversially maintained that items such as the round square and the golden mountain have no being whatsoever without prejudice to their being perfectly legitimate objects. To assume (C2), however, is to assume that everything exists. Without this assumption, Tom's thinking about a unicorn could be construed as his standing in the intentional relation to a nonexistent item. I am quite sure that any such Meinongian construal of (A) is anathema to Ed. He has that "robust sense of reality" of which his countryman Russell spoke.
I think it is important to appreciate -- and here I expect Ed to object strenuously -- that (C2) is not a truth of formal logic, but a substantive truth, if it is a truth, of metaphysics. For its negation -- Some items are nonexistent -- is not a formal-logical contradiction. 'Some' is a purely logical word: it implies nothing about being or existence. It is a logical quantifier, the particular quantifier, and nothing more. Meinong and the Meinongians (Routley, Parsons, Castaneda, Butchvarov, et al.) may be wrong in their metaphysics but surely they are not contradicting themselves when they say or imply that some objects do not exist, as they would be were they to say or imply that some objects are not objects. Note also that the logical form of 'Some items are nonexistent,' namely, Some Fs are Gs, has both true and false substitution instances, not all false substitution instance as would have to be the case were the form contradictory.
A third assumption Ed makes:
(C3) All logically kosher uses of 'exist(s)' and cognates are univocal.
Call this the Univocity Assumption. It is a semantic thesis that rules out different senses of 'exist(s)' whether equivocal or analogically related. There is exactly one sense of 'exist(s)' and that is the sense specified in (C1). The ontological counterpart of (C3) is the following No Modes of Being assumption:
(C4) There are no modes of being or of existence: everything that exists exists in the same way.
Thus Ed will have no truck with any such distinction as that between esse intentionale and esse reale, merely intentional being and real being. I think that Ed's acceptance of (C4) is at the back of his belief that I commit the 'Existential Fallacy.' More about this in a moment.
(C5) The one legitimate univocal sense of 'exists' is fully captured by the 'existential' or particular quantifier of standard first-order predicate logic with identity.
Anyone who accepts the (C)-assumptions will also accept the (B)-premises, assuming that 'world' in (B4) refers to the totality of what exists, die Gesamtheit der Dinge, nicht der Tatsachen, to turn the Tractarian Wittgenstein on his head.
So why are we disagreeing? In my opinion you are committing the existential fallacy by moving from premisses like (A) which are not existential, to a conclusion which is existential. And in a very odd way indeed, because you justify the move via B-type premisses. I.e. since the term 'unicorn' cannot correspond to any existing object, you infer that it must correspond to some non-actual or intentional object. Thus at some point (supported by other premisses that I don't include here) you move to
(C) Some objects are intentional objects
or something like that. Something which looks a bit like an existence claim, but which cannot consistently be interpreted as an existence claim in the sense of any of the B-sentences above. Otherwise you have committed the 'existence fallacy', i.e. you have moved from premisses which do not assert or imply existence, to a conclusion which does assert or imply existence. Indeed, the 'double existential fallacy'. Not only do the premisses fail to assert or imply the existence of the things asserted or implied in the conclusion, they explicitly assert or imply their non-existence!
BV: What Ed suspects me of is the illicit inferential move from
(A) Tom is thinking about a unicorn
(A*) Tom is thinking about something
(A**) Something is such that Tom is thinking about it.
We can call the move from (A*) to (A**) the Exportation Move: 'something' is exported from the intentional sphere or context to the non-intentional sphere or context. The first move, from (A) to (A*), strikes me as logically innocuous. Call it Particular Generalization within an Intentional Context, or 'within the brackets' with a nod toward Husserl. If Tom is thinking about a unicorn, he is thinking about something, not nothing! 'Something' is here of course (C1)-existentially neutral. One can particularly generalize within a dream with no commitment to (C1)-existence. A charming feminist appears in a dream and I conclude, hot damn, some feminists are charming! This is consistent with there not being any such feminists in reality.
The Exportation Move is fallacious for Ed because for him 'something' is existentially loaded. 'For some x, x is F' is synonomous in Ed's head to 'There exists an x such that x is F,' where the second sentence asserts existence in the sense clarified by (C1). Now I think there is a clear fallacy in the vicinity, call it the Existential Fallacy, an example of which is the illicit two-step move from
Sally wants a baby
Sally wants something
There exists an x such that x is a baby and Sally wants x.
In the normal case, when a woman wants a baby, what she wants is not some particular baby already in the world; what she wants is motherhood: she wants to be the receptacle through which a new baby comes into existence. Clearly, the Sally inference commits the Existential Fallacy in the second step. It is plain that if I want something that may or may not exist without prejudice to my wanting it, it does not follow that the object exists in reality apart from my wanting it. The same goes for 'hunting' and 'looking for' and 'pining after' and other intentional verbs.
We need to distinguish among the following three concepts:
1. Existentially neutral particular generalization within an intentional context. This strikes me as a logically kosher move. Suppose I am thinking about a distant relative, Angelo, whom I believe to be alive, but unbeknownst to me died last month. If I am thinking about Angelo, then I am thinking about something despite the fact that Angelo no longer exists. Suppose I am thinking about him while he expires. His dying in no way terminates or alters the object-directedness of my thought. Particular generalization, therefore, is valid within the intentional sphere.
2. Existentially neutral exportation. This too strikes me as logically valid. If I am thinking about Pegasus, then I am thinking about something, not nothing. But that is equivalent to saying that something is such that I am thinking about it, so long as both occurrences of 'something' are existentially neutral. But if the second occurrence were existentially loaded, then the inference would clearly be invalid.
To see this more clearly, let us suppose that Meinong is right and that some items are beingless, Pegasus being one of them. Well, if Pegasus is nonexistent, then surely some item is nonexistent, both within and without the Intentional Context. This should not be surprising since Meinongian objects could be thought of as hypostatizations of merely intentional objects whose mode of being is esse intentionale.
Well, what about such merely intentional objects, objects that have no being apart from the mental acts whose objects they are? If Sam is thinking of something that is merely intentional, does it follow that something outside of the intentional context is such that Sam is thinking of it? This is a difficult question. I don't think I am required to answer it on the present occasion.
3. Existentially loaded exportation. This is an invalid move. Sam cannot search without searching for something, say, the Lost Dutchman Gold Mine. Searching for that mine, he is not searching for something immanent to his consciousness: he is searching for something which is such that, were it to (C1)-exist, it would so exist in splendid transcendence of his and everyone's mental states. Nevertheless, Sam's searching for something does not show that he is searching for something that exists in reality.
Ed's complaint may be that I am commiting the Existential Fallacy by engaging in existentially loaded exportation. Well, let's see. One thing is clear: if Tom is thinking about a unicorn, then he is thinking about something, not nothing. He has a more or less definite object before his mind. That object is not nothing; so it is something. This is a phenomenological datum that every theory must cater to or accommodate. To think about a unicorn is not to think about a flying horse or about the Lost Dutchman Gold Mine. This object is the object of the mental intention, or, the intentional object. It may or may not exist in reality apart from the mental acts trained upon it. So one cannot infer the extramental existence of intentional object O from O's being an intentional object.
So I have no trouble accommodating Ed's logical insight. His point is that if someone is thinking about something, it does not follow that there exists in reality -- in the (C1) sense -- something about which he is thinking.
So I plead innocent of the Existential Fallacy. Crucial to my defense of talk of merely intentional objects is my rejection of Ed's assumptions (C3), (C4), and (C5). I say that there are different modes of being or existence. When Tom thinks of a unicorn, he thinks of something, not nothing. This item lacks existence in the (C1)-sense. It lacks esse reale. But it enjoys esse intentionale. So while merely intentional objects do not exist in the (C1)-sense -- this is true by definition! -- there is a sense in which they exist. They exist intentionally.
In this way I avoid both the Meinongian position and Ed's draconian reism, all the while doing justice to the phenomenological data.
Obviously there are ways round this, but I would be interested to know what you think they are. You have to start with a sentence like (A), of form 'S-V-O' containing subject phrase 'S' ('Tom'), verb-phrase 'V' ('is thinking about' and object phrase 'O' ('a unicorn'). Then a second sentence denying existence in some sense of things correspond to 'O'. This sentence can only be of the following 3 forms: containing the word 'exist', containing the verb 'is' (or cognate like 'there is/are') or just containing a verb ('contains'). The term 'O' will have to appear somewhere. Finally a conclusion which asserts existence in some sense. To avoid a fallacy, you need to establish that the sense of 'exist' used in the conclusion is different from the sense used in the second premiss.
BV: It seems to me that that is what I have done above. It seems to me that the following propositions are logically consistent:
a. Tom is thinking of a unicorn
b. Unicorns do not exist in the (C1)-sense.
c. Tom's mental state is object-directed; it is an intentional state.
d. The object of Tom's mental state does not (C1)-exist.
e. The merely intentional object is not nothing.
f. The merely intentional object enjoys intentional existence, a distinct mode of existence different from (C1)-existence.
Ed's critique rests on his (C)-assumptions. But they are all reasonably rejected.