Against Structured Propositions

Here’s a view about propositions:

A proposition P is a set of ordered pairs <A,G> where the first object is an individual and the second a property. These propositions are generally expressed by declarative statements such as my utterance of the sentence `Alvin is Green’. Call this `Structured Propositions’.

Here’s an argument against this view of propositions.

P1) If propositions are structured then one must either: a) become Meinongian, or b) accept gappy propositions.
P2) According to Russell, Meinongianism entails contradictions, so it’s unacceptable.
P3) Gappy propositions cannot explain informative speech acts where true negative existential are asserted. Such speech acts cannot express gappy propositions. So, gappy propositions are unacceptable.
P4) From P2 and P3, it follows that we should neither be Meinongian nor gappy proposition theorists.
C) Propositions are not structured.

Here’s an argument for P1.

Suppose that the name `Alvin’ is referenceless. I go on and utter the sentence `Alvin is Green’. Which proposition did I assert? If propositions are structured, then what I asserted looks either like this ‹A, G› or like this ‹ , G›. The first one requires you to believe that Alvin, who does not exist, IS in fact some kind of object: a non-existent, or subsistent one. That’s Meinongianism. The second option requires you to believe that there are gappy propositions: propositions with unfilled spots in their structure.

For Russell’s argument for P2, read his “On Denoting’’.

As for P3, here’s an argument. Suppose you reply to my assertion by uttering the sentence `Alvin does not exist’, which happens to be true. Which proposition did you assert? Assuming that you cannot be Meinongian, the proposition looks something like this: NEG‹A, E›, where `NEG’ stands for negation and `E’ stands for the property of existence. The proposition is, clearly, a gappy one. The problem is that it should not be.

Here’s why. We need evidence to show that a particular speech act ends up having a propositional gap. More often than not, this is signaled by a presupposition failure. The traditional example is Russell’s King of France: `The present King of France is bald’ presupposes that there is some such thing as a present King of France. Since this presupposition is not fulfilled then the utterance does not express a full-blown, truth-evaluable proposition. Suppose, for argument’s sake, that that’s a good paradigm of propositional gaps. Does that happen with true negative existentials?

The answer is no. In general no existential claim carries existence presupposition. If they had then they would all be useless: i.e., they would all be asserting what they are presupposing. Your reply, for example, would like these:

Let us suppose that Alvin exists. By the way, Alvin does not exist.

Or, think of the more problematic, affirmative existential claim:

Let us suppose that Alvin exists. Now let me tell you something about Alvin. Alvin exists.

The absurdity of these speech acts goes to show that existential claims in general do not carry existence presuppositions. If you want, you can run the usual “hey, wait a minute’’ tests and see this for yourself. The main point here is that existential claims in general do not carry existence presuppositions.

How does this help? Well, if they don’t carry existence presuppositions, then your utterance of the true sentence `Alvin does not exist’ doesn’t carry an existence presupposition. If so, then the fact that `Alvin’ is referenceless does not affect the presuppositions of your speech act. In particular it does not make it so that there is some presupposition failure involved in it. And this is important because if there is no presupposition failure then there is no reason to think that what you uttered has any propositional or truth-value gap whatsoever.

It follows then that whenever someone utters a true negative existential claim they manage to convey a full-blown proposition. But non-meinongian fans of structured propositions are forced to say that these speech acts express a gappy proposition. Which is simply false. True negative existentials cannot express gappy propositions.

Structured proposition theorists are forced to be Meinongians. I referred to Russell’s argument against this view. But here’s a brief, sketch of it. What would happen if you become Meinongian in order to save the day for Structured Propositions? Well, you would end up having something like this: NEG ‹A, E› which literally says that there IS an object that does not exist. Of course, Meinongians would like to distinguish between ways of existing. What the proposition really says is that there IS a Subsistent object that does not exist. But Meinongians should also accept that there is an important set of things that REALLY do not exist: like the square circle. Otherwise you run into contradictions. So they have to distinguish between objects like Alvin and the square circle, they all have different ways of not existing. And all that just seems ridiculous.


22 Responses to Against Structured Propositions

  1. Jon S. says:

    I suppose I should weigh in with what must be my expected response.

    (1) P2 is an explicit appeal to Russell’s authority on a historical question. As passing familiarity with his History of Western Philosophy would suggest, this is a bad thing to do. (More seriously, Meinongianism does entail contradictions, but they are only contradictions about non-existent objects. It’s unclear why something that doesn’t exist should be expected to obey a law about what can be true of existing things.)

    (2) In your argument for P1, you seem to equate the terms “non-existent” and “subsistent.” This is certainly a mistake if you’re trying to use Meinong’s terminology; plausibly a mistake even in standard terminology.

    (3) In your final paragraph, you’re missing what Meinongians would (or, at least, should) say. First, they don’t distinguish between ways of existing, at least not in a way that matters here. (Meinong himself distinguishes between existing, which is the normal every-day existence that you and I enjoy, and subsisting, which is existing minus “temporal existence,” which is the sort of non-temporal existence that numbers and Augustine’s God might enjoy.) The important distinction here is between having being (which includes both subsisting and existing, in the senses given above) and not having being, which is sometimes called absisting. This is not at all a distinction in ways of being–it’s a distinction between being and not being, full stop.

    (4) The proposition in your final paragraph doesn’t say that there is a subsistent object that doesn’t exist. (Although, the proposition expressed by the sentence ‘The number 5 does not have temporal existence.’ would say such a thing, in Meinong’s terminology.) The proposition says that some absistent object fails to exist, which isn’t particularly weird since “absisting” just means “being an object” and not “having being” or “subsisting” or anything like that. So, in fact, the Meinongian doesn’t distinguish between Alvin’s absisting and the round square’s absisting.

    (5) Re: the sort of worries Russell raises in “On Denoting,” Meinong addresses them in a subsequent edition of his main book, _Uber Annahmen_ or _On Assumptions_. For an explanation of why (he thinks) Russell does not deal his theory a death blow, you’d have to read my undergraduate thesis, conveniently located in Tanner library. I presume you aren’t actually that interested in it, since you can more or less take for granted that people think “ew” when you say “Meinong,” but you should at least take care to get the Meinongian view right.

    Let me know if something I’ve said about Meinong’s view or his terminology requires more explanation.


  2. Jon S. says:

    One further thought: wouldn’t non-Meinongian structured propositions people just deny that any proposition containing such an object as Alvin is expressed? My understanding is that they don’t think that there are any propositions about Alvin at all, even though they think that such a sentence as “Alvin does not exist” is true (presumably in virtue of being analyzed as having some non-apparent logical form). (Having Meinongian-structured-propositions-leanings but not much serious interest here myself, I don’t know what the other lot say.)


  3. Edú says:

    Awesome Jon, thanks a lot for the clarifications. I think I agree with you about mostly everything. I just want to add some things. Here’s what I want to add:

    1) Suppossing we agree on everything here, it follows that Meinongianism entails contraditions. They are contradictions about non-existent things, but nevertheless a contradiction. Unlike Jon, I think that it is really bad to have theories that entail contradictions, no matter about what. Word has it that contradictions entail everything. So contradictions in the Meinongian theory entail, among other things, that Meinong’s theory is wrong. That’s bad, isn’t it? When your theory entails that your theory is false?

    2) Nothing to add on Jon’s second point.

    3) I disagree. I think that Meinongian structure-proposition fans will not take Alvin to be simply an object that has no being, full stop. They want it to be an object, so that it can take part of the structured proposition. So the important distinction is between existing and subsisting objects. So, by your own comments, they do focus on ways of being, not between being and not being full stop.

    4) Same as above.

    5) You’re right. I’m not actually that interested in whether or not Meinong’s theory is dead after Russell’s blow. I’m interested in showing how unattractive it would be to be a Meinongian structured-proposition theorist. I hope to have succeeded on that count.

    6) On the last note, I’m afraid to confirm that non-Meinongian structured-proposition fans will not follow your suggestion. For one good reason: they want to explain how competent speakers manage to convey non-trivial information by uttering empty-name sentences. Following your advise would offer no explanation of such a phenomenon. The other, non-meinongian lot of people simply say that the propositions are gappy and that this makes them false, because the predicate does not map any object into truth.

    And thanks a lot for the clarifications again!!

  4. Dustin says:

    Hi Edu,

    Your argument for (P1) assumes that you have asserted a proposition when you asserted ‘Alvin is green’. Given your stipulation that ‘Alvin’ is “referenceless”, that assumption is highly dubious.

    That’s not to say that I don’t agree that you’ve expressed *some* sort of *thought*. I do agree that you’ve done that–I would just be extremely hesitant to say that you’ve asserted a proposition.

  5. Jon S. says:

    Some further clarification on points 1 and 3.

    On point 1, I should have been clearer about the kinds of contradictions a Meinongian will admit to endorsing. For Meinongians, it can be true of an object, like the round square, that it is both round and not round, and in general objects can have the property P as well as the property not-P. The inference from “x has the property not-P” to “x does not have the property P” or “not(x has the property P)” is blocked when it comes to non-existent objects, however. And so you don’t get contradictions of the form “S and not-S” for any sentence S. That is, objects can have contradictory properties, but these don’t generate actual contradictions at the sentence or proposition level. So, speaking strictly, for the Meinongian the law of non-contradiction is not violated, because that law says that propositions P and not-P should never be true together, and for the Meinongian they aren’t.

    On point 3, I should clarify further what it means for a Meinongian to say that something is an object. What matters here is that Alvin can be an object without subsisting. “Being an object” is just the same as “absisting.” The idea, basically, is that something can have properties (i.e., it can be an object, it can be various ways) without having being. So, for things that exist in the usual sense, the Meinongian distinguishes between two ways of being, viz., between temporal and non-temporal being. (Quine, I believe in “On What There Is”, grants that such a distinction would make sense, and only complains that it does not follow standard usage of the verb “exists”.) But for things that do not exist in the usual sense, for Meinong they neither exist nor subsist, they merely absist, i.e., they simply have no being. Nevertheless, they ARE objects, and can do the sort of object-y things you’d like them to do, like become parts of structures and be the referents of certain syntactic objects.

    One further point, which I think actually supports your view: Meinong actually acknowledges that the structures including objects without being (like Alvin or round squares) could never be propositions, because propositions are taken to subsist themselves, and all of their parts (i.e., all of the things from which they are constructed) must have being (i.e., must subsist or exist) as well. But he posits entities analogous to propositions, except that the entities in question may fail to subsist if they are structured from object which fail to subsist. These entities he terms “objectives”. Objectives subsist when all of the objects from which they are constructed subsist (or exist), but merely absist when some object involved fails to subsist. So, for Meinong, Alvin can be an object without subsisting, but this is no help to the structured propositions folk, because the received view that propositions subsist is maintained by the Meinongian. (A Meinongian, though, might plausibly hold a “structured objectives” view.)


  6. Jon S. says:

    Dustin chimed in at just the right moment–his response was what I was gesturing at in (what we’re calling) my (6) above.

  7. Edú says:


    I agree. I have presupposed that meaningful utterances of declarative statements do express propositions. But I don’t think that is highly dubious in any sense. And neither do any of the Neo-Russellians defend structured propositions do. Salmon, Soames, Reimer, Taylor, and Braun all agree that there has been asserted a proposition.

    I find your suggestion that some sort of thought, but not a proposition, has been expressed as highly ad-hoc. If you are going to split thought-contents from propositions, then you’ll need to tell me another story about what in the world are the contents of thought, if not propositions. And then you’ll also have to tell me how this helps understand human behavior, given the traditional belief-desire psychology model. I don’t know what you’ll have here, but I bet you’ll want to call it a proposition. But hey, you are supposed to think that propositions are structured to begin with!


    Thanks for the clarification on Meinong’s view of propositions. It does seem to be unhelpful for structured proposition theorists. I am puzzled, though, by your distinction between not having the property P and having the property not-P. Couldn’t I just say.

    The round square exists. It follows that it has the property of being round.

    We know that it is impossible for square things to be round, and we know that the round square is square. So we know that it is impossible for the round square to be round. Modal logic tells us that if it is not possible to P, then it is not the case that P. So it follows that the round square does not have the property of being round.

    From which it follows that both the rounds square has the property of being round and that the round square does not have the property of being round. That’s a contradiction, and that’s all I need isn’t it?

  8. Dustin says:

    Hi Edu,

    I would probably agree with you that ‘meaningful utterances of declarative statements do express propositions’. But I would take issue with whether in this case we have a ‘meaningful utterance of a declarative statement’. Well, I would agree that in *some* sense it is meaningful–but I would disagree that it is meaningful in the sense of ‘meaningful’ according to which it is true that every meaningful utterance of a declarative statement expresses a proposition.

    Anyway, here is a very general (and to my mind plausible) reason to deny that your utterance expresses a proposition: *whatever* propositions are, they should at least determine truth-conditions, but your utterance doesn’t have truth-conditions, and, hence, doesn’t express a proposition.

    Also, my separation of thoughts from propositions (the contents of thoughts) isn’t ad hoc at all. Many theorists have already accepted this distinction to account for many other puzzles (like Frege’s puzzle). In the literature, the view according to which thoughts are distinct from propositions often goes by the name ‘guise theory’.

  9. Edú says:

    Hi Dustin,

    I see your point. It’s well taken. But, perhaps the problem can be cleared out if we pick the adequate example. What we are dealing here with is not ‘Alvin’ in particular but utterances of empty-name sentences in general. So take another example. We are talking about Conan Doyle’s work, and I go on and say:

    Sherlock Holmes is a addicted to cocaine.

    What are your intuitions about this case? It seems to me that I have made not only a meaningful
    statement, it also seems that I have said something true. So this looks like it meets your own standards for ‘meaningfulness according to which there is a proposition being expressed.’ And yet, Sherlock Holmes does not exist, as we all know.

    A second point, with respect to guise theory. If what you refer to is something similar to what Salmon, for example, defends in his “Frege’s Puzzle” I am afraid it’s not going to work the way you want it to. Guises are proposition-guises and so their existence depends on the existence of a proposition. So if my utterance of ‘Sherlock Holmes is addicted to cocaine’ expresses a guise, it must be because it expresses also expresses a proposiiton. It couldn’t be otherwise.

    The way Salmon uses this for Frege’s Puzzle is simple: the difference in informativeness is owed to differences in guises but not in propositions. I don’t see how this can work for empty names. Salmon himself does not seem to think it does, since he prefers a Meinong-like view.

    Perhaps by guise theory you mean something else (Neri-Castañeda’s theory?). I don’t know: neither Castañeda’s theory nor how they would work. Could you add some more here?

  10. dtlocke says:

    Hi Edu,

    I have no idea what to say about cases where we speak about fictional characters. It isn’t clear (1) whether what we say in such cases is true/false rather than assertable/not-assertable, (2) whether we express propositions in such cases, or (3) if we do express propositions, what the structured-propositions theorists is committed to saying such propositions contain (see below).

    As far as ‘guises’ go, please note that *I* avoided using the term ‘guise’ when I stated the view. I did so, because the term ‘guise’ has the tendency to mislead people (as it did you) into thinking that there can’t be a guise unless there is something that is being guised. I used the term ‘thought’, and the idea was that thoughts (mental states) guise propositions (n-tuples/truth-conditions/whatever you take propositions to be). Think of masks–masks mask people, but not every mask masks someone (some are sitting on the shelf). That’s the way I think of thoughts–thoughts guise propositions, but not every thought is guising a proposition.

    The other alternative for rejecting your argument, is to accept that propositions need not determin truth-conditions (all by themselves) and then to insist that propositions don’t always contain the objects that utterances expresssing them are about. Consider the statement ‘I have pants aflame’. Russell would have said that the proposition expressed by this statement contains me (my very self). However, an alternative structured-proposition view is to say that this proposition contains a function from indexes to individuals. Givan an index, the function would yield an individual and so the proposition together with the index would determine truth-conditions. Similarly, a structured-proposition theorist could say that the proposition expressed by your utterance of ‘Alvin is green’ contains such a function. The explanation of why your utterance does not have truth-conditions would thus be that the aforementioned function is undefined at the relevant index.

    Either way, no Meinongian objects are involved.

  11. Jon S. says:

    Hi again,

    Sorry not to have gotten back to this earlier–here is the promised written version of the Meinongian response to your last query directed at me.

    First, the round square doesn’t exist. (While we’re at it, the existent round square doesn’t exist either–but I refuse to explain this here because I’d just be reproducing my thesis.) But it is round, so we can start there. Of course it’s impossible for existent objects (or, for that matter, for all possibilia) to be both P and not-P. But the round square isn’t to be found among possibilia. So what’s impossible for possibilia has no bearing on it. For the same reason, we can’t infer from the fact that the round square is square that it is also not the case that it is round. (There’s no need to block inferences from an atomic sentence of the form “x is P” to an atomic sentence of the form “x is Q” when all Ps are Qs. So we CAN infer that the round square is not round. But we can’t move this negation outside and infer not(the round square is round), just because merely absistent entities are not subject to the usual considerations that would allow such an inference–they can bear contradictory properties.)

    Introducing certain properties (like possibility or existence) into the descriptions of nonexistent objects (e.g., the possible round square, the existent red black thing, etc.) causes what look like more worrying problems for the Meinongian, but Meinong has a response (which, basically, denies that such properties can be attributed willy-nilly to objects, a la Kant’s denial that existence is a predicate, holding that objects either have them or fail to have them). It’s of course more complicated than just that, but I suspect that your interest in Meinong will (justifiably) not extend to looking for justifications of the things he says about this.

    As for your project, I still think Dustin is pointing in the right direction. Structured-propositions folks have plenty of resources (to give an example he hasn’t mentioned yet, they could go fictionalist) to avoid going Meinongian.


  12. Edú says:

    Hi Dustin,

    Here’s what I think

    (1) I think it is pretty clear that what we say in fictional discourse can be either false or true. Once speakers read Conan Doyle they come to know that it is true that Sherlock Holmes is addicted to cocaine. I don’t think there is anyone disputing this point, except for you of course. Furthermore, you do not need to limit to fictional discourse either. Consider a professor in an English Department who dedicates his life to study Shakespeare and, in particular, Hamlet. Are you inclined to say that all her work will simply be assertable non-assertable? I think this is ad hoc.

    (2) In so far as what we express by using empty names can in fact have an effect on your behavior, then it must be that what I say is something that can be the content of your propositional attitudes. And there is no way you get this without propositions.

    (3) I agree, it is not obvious what structured-proposition theorists want to say. But it is clear that they do not want to say what you think they can say. Putting in a function instead of an individual will simply give you a different kind of proposition. If you go back to my original post you will find out that structured propositions BY DEFINITION include the object to which the proper name refers IN the object slot of the proposition. Structured-propositions are supposed to imitate the syntactic Structure of the English sentence, not its logical structure (if you can find that one). You have to keep in mind that structured-proposition theorists tend to be Millians who believe in direct reference.

    (4) Finally, guises. It seems then that your mask theory has little to do with guise theory. So your words were misleading when you presented your view as a version of guise theory. In any case, I think there is an important disanalogy between masks and your non-propositional thoughts. Masks are relevant by themselves, as you want them to be, not because of what they cover. Their content, the masked individual, does not have any causal efficacy for the masks job. If that’s what you think of thoughts and propositions, then you might as well get rid of propositions. If so, then you might also get rid off propositional attitudes and get simple attitudes. You might also, perhaps, get rid off the causal relevance of propositions in belief-desire psychology. And once you are there, you might want your thought-stars to play all those roles. I guess you will want your thoughts to have truth-values. If you get to this point in your theory: wouldn’t you like to call your propositions thoughts?

    Jon, I will get back to your last comments soon.

  13. Edú says:

    Hi Jon,

    You are the authority on Meinong here. So I will not debate you last comment. I will agree, then, that Meinongianism, properly construed, is not inherently flawed. I still think, however, that it more expensive, ontologically speaking, than non-Meinongianism. So let’s keep that as a very last option.

    I think, however, that both you and Dustin are confused about structured propositions. So I will present the view once more. Fans of Structured propositions make, at least, two claims:

    (1) Propositions reflect the syntactic structure of the sentences, the utterances of which express them.

    (2) In the case of declarative statements with subject-predicate forme, the structure is filled in by : (a) an object in the subject position, and (b) a property in the predicate position.

    It goes without saying that whatever view that does not endorse (1) and (2) is NOT a view of propositions as ‘structured’. They might be structured, but that will be a different view that should go by a different label.

    In particular, Dustin think that we can change (1) for something like (1A)

    (1A) Structured propositions reflect the logical form of the utterances of sentences that express them.

    Only if we accept (1A) we can accept his claim that, instead of an object in the subject position, we get a formula that delivers an object. Notice, also, that accepting this latter claim would entail a second change in the nature of structured propositions. This is not, as I said, the same thing. It is not, at least, the view of propositions I am intending to reject.

    Here’s one more reason to realize that this is so. As far as I know, contemporary fans of structured propositions are Millians who called themselves ”Neo-Russelllians”. They believe in direct reference (claim 3 below) and structured propositions help them make their point more clearly.

    (3) Proper Names contribute their referent, and nothing else, to the content of a sentence in which they are involved.

    (3) together with (2) gives you propositions in which the referent itself is part of the proposition. There is no better way to claim that names contribute their referents to the content. That’s the attractiveness of structured propositions.

    Now, you (Jon) point to one other option for structured -proposition theorists: fictionalism. But the option is mistaken as well. This time it is clearly so. I take it that your confussion was with respect to structured propositions and not fictionalism.

    Fictionalism, as applied to a certain domain of discourse, is the view that discourse D is not really true or false, or that the objects to which D REFERS DO NOT EXIST. How can this be compatible with the object-need of structured-propositions theorists? Szabó, for example, characterizes fictionalism in the following way:

    “To be a fictionalist about Fs is to think that our naïve attitude toward F-discourse is only halfway correct: we are right in thinking that we use genuine singular terms that purport to refer to Fs, but wrong in thinking that they actually succeed in referring. In engaging in F-discourse we inadvertently slip into fictional talk”

    Briefly put: Structured-propositions require objects. Fictionalism takes out objects. How can they live together? I think they cannot.

  14. Edú says:

    One last comment, to Dustin:

    You said you have no idea what to say about uses of empty names in fiction. You think that things are pretty much unclear. I am wondering what you think of non-fictional uses of empty names. I tend to think that my claim that Sherlock Holmes does not exist expressess something that:

    – is a piece of non-fiction;
    – is truth-evaluable;
    – is meaningful;
    – may be embedded in a propositional attitude;
    – may be causally efficacious; and, most importantly
    – it is true;

    What do you think?

  15. Dustin says:

    Hi Edu,

    How do you think that the defender of *un*structured propositions is to handle the case of empty names?

  16. Edú says:


    I have thought of only one ‘un’structured proposition option. There should be more. But the one I like takes propositions to be sets of possible worlds. Stalnaker does this in a Millian-Friendly way! So, for example, when I say that Sherlock Holmes does not exist I express the proposition that is true in all and only those worlds that do not include Sherlock Holmes.

  17. Dustin says:

    Thanks, Edu. So one further question: which possible individual is this Sherlock Holmes and what is your (non ad hoc) reason for saying that “Sherlock Holmes does not exist” expresses the proposition that is true at all and only those worlds that do not include him?

    Also, no matter what you say to that question, your theory seems to postulate the existence of merely possible individuals (this Sherlock Holmes you spoke of). What exactly was wrong with Meinongianism again?

  18. Edú says:


    Thanks for your objections. I think there are many solutions to them.

    1) If you look back you’ll reallize that the possible-worlds representation of the proposition THAT SHERLOCK HOLMES DOES NOT EXIST, does not involve Sherlock Holmes in any way. It makes use of all and only those worlds where there is no Sherlock Holmes.

    2) The non ad-hoc reason to say that this is the proposition expressed is given by speaker intuitions: they seem to agree that there is a truth-evaluable, meaningful, proposition being expressed and that this proposition can be expressed by saying “There is no such thing as Sherlock Holmes”. I don’t see the ad-hocery here. Using possible-worlds semantics is just a way of representing this, one that does not run into the problems of structured propositions.

    3) Possible worlds semantics does not postulate the existence of possible individuals, just like using numbers to say how many apples you have does not by itself postulate the existence of numbers. It is a further question whether or not you want to argue that those possible individuals exist. I know you know this, but I will nevertheless say it: one need not be a modal realist to have possible worlds semantics. Boris knows of a way of doing this. If I’m not wrong it is called ‘Lagadonian’ and it claims that every object stands for itself. I won’t ge into these matters. There are other options as well, many different versions of actualism, and even fictionalism. I think John Divers has a fairly recent proposal on how to have possible-worlds semantics without possible individuals. Published in Mind a year ago or so.

    So possible-worlds semantics does not come with the ontological cost of Meinongianism. That’s what was wrong about structured propositions and, ultimately, Meinongianism.

  19. Dustin says:


    My points correspond to yours.

    (1) I take it that the proposition expressed by “Sherlock Holmes does not exist”, on your view, is *not* the necessary proposition. So there are worlds where it is false. On your view, this amounts to there being some worlds where Sherlock Holmes does exist. So I repeat my question: who is this Sherlock Holmes, and what is your non ad hoc reason for saying that “Sherlock Holmes does not exist” expresses the proposition where *he* does not exist.

    (2) The ad hocery I am alluding to has nothing to do with your taking “Sherlock Holmes does not exist” to express a proposition, nor does it have anything to do with your taking it to express a possible worlds proposition. The ad hocery comes in trying to say *which* possible worlds proposition it expresses [see point (1)].

    (3) OK, so just tell me: what is your account of the merely possible Sherlock Holmes? Your view talks about him. So you must either analyze him away, accept him, continue to play the ostrich, or admit that the problem of empty (fictional) names is a problem for everyone, not just for the proponent of structured propositions.

  20. Dustin says:


    At the end of (1), that should say “the proposition true at all and only those worlds where *he* does not exist”.

  21. Edú says:


    Enough ostrich playing. Here are two responses.

    1) I take the true negative existential claim that Sherlock Holmes does not exist to express a proposition that does not involve Sherlock Holmesin any sense. I take it to be synonymous with the proposition that there is no Sherlock Holmes. This is not hte same as the proposition that is true in all and only those possible worlds where *he* does not exist. The proposition I have in mind can be expressed without the use of the pronoun. I am not talking about any individual in particular. I am talking about possible worlds that where there is no thing as Sherlock Holmes.

    Now, you are correct in pointing that the whole analysis presupposes that there are worlds there is some such thing as Sherlock Holmes. What seems to be missing is this: all those worlds are such that, when uttering the name ‘Sherlock Holmes’ in those worlds you refer to different things for different worlds. Hence, Sherlock Holmes of w2 need not be the same as Sherlock Holmes of w3. Our world just happens to be one where utterances of ‘Sherlock Holmes’ do not manage to succed in refering to something.

    So again, there is no need to use the pronoun in the way you do. Doing so would actually be misleading. It would require us to arbitrarily pick one of the many worlds that in fact have Sherlock Holmes among their parts.

    2) I still think that possible-world semantic need not care about the metaphysics. But that is contentious and you don’t believe it. So I won’t keep pressing the point. But I will restate my answer.

    There are many different stories you can tell about possible worlds. Kripke says they are possible stories, Sherlock Holmes is in one (or more) of them. Stalnaker says they are ways the world might be. Others say they are fictions.

    This is interesting. Unlike structured propositions, possible-worlds propositions allow for the possibility of fictionalism. Since possible-worlds semantics does not require objects to fill in any gaps, possible-worlds propositions are happy enough to have fictional characters. There are other options, I am sure. The important thing is not to defend fictionalism about possible-worlds, but to show that structured propositions are directly tied with an ontological commitment in a way that possible-worlds semantics is not.

    Now, the way I see it, possible worlds ARE (with an ontological R) functions. I think we can accept functions as abstract and I think this is less commiting than Meinongianism.

    I can see you complaining about this move. So let me present you with the option I was considering at the very beginning: Lagadonian languages. The main claiim here is that every object stands for itself. Now, consider this world as a structured big object in parallel to the way in which sentences are structured syntactic objects. So, suppose the world has some or other syntax-like form. Now, you can, using your imagination, abstract the knots and bolts and conceive of other ways of tying them together. Different ways of doing this will give you different possible ways in which this world could be. So it might be that Sherlock Holmes is that: the result of rearranging the knots and bolts of this world in a different way.

    Is that still too ostrich like?

    Last, very last: I do agree that Empty Names are a problem for everyone. I just happen to think that some can solve the problem and others simply cannot.

  22. Jon S. says:

    Obviously I haven’t checked in in a while, but there’s one point to make about the ontological costs of modal realism vs. Meinongianism, which is to note that modal realism has an expanded ontology which includes possibilia, while Meinongianism doesn’t attribute _being_ in any sense to non-subsistent objects. (Abstract objects, concrete objects, etc., all have being, but objects like Sherlock Holmes and so on have no being whatsoever.) Rather, Meinong simply offers a definition of objecthood disconnected from being; roughly, to be an object is to be the object of a thought. That claim carries no ontological commitments whatsoever if you can convince yourself to give up the (so far as I can see, unmotivated) notion that whatever is an object of thought must be something that has being. Thoughts connect up with all sorts of funny things–why should something have to belong in our ontology for us to think about it?

    (Let me know if you respond to this or want some clarification, as I don’t check this blog very often lately.)


%d bloggers like this: