Could Propositions Exist Contingently? A Response to Ben Wallis

Counter-apologist and valued commenter Ben Wallis has posted some criticisms of the argument for God from logic. (His post is basically a synthesis of the comments he posted here.) His approach is to attack the claim that if there are necessarily true propositions (i.e., necessary truths) then those propositions necessarily exist by appealing to the distinction between truth-in-w and truth-at-w (a distinction employed by Kit Fine and Robert Adams, albeit with different terminology). Drawing on this distinction, Ben proposes a view of propositions according to which necessary truths exist contingently. In this follow-up post, I explain why I believe Ben’s proposal isn’t viable.

Ben’s Proposal

Given what Ben writes in his post, it appears he is committed to all of the following claims:

  1. There are some necessary truths, i.e., some propositions that are necessarily true.
  2. Propositions exist contingently.
  3. Propositions are dependent on minds; in the absence of minds there can be no propositions.
  4. A proposition is not identical to any particular thought, but is instead a sort of “similarity class” of thoughts.
  5. A possible world is a maximally consistent set of propositions that describes a hypothetical state of affairs.
  6. We can meaningfully distinguish between truth-at-w and truth-in-w, where w is some possible world; only the latter entails the existence of a truth (a true proposition) in w. To say that some proposition p is true-at-w is only to say that p is a member of w; it is not to say that p exists in w and is true in w.

I assume that Ben is committed to (1) because he doesn’t contest it, because he’s a sensible level-headed fellow, and because he’s a mathematician (mathematical truths being paradigmatic necessary truths). But he’s also committed to (2) because he wants to argue against the claim that propositions exist necessarily so as to evade the theistic argument.

(3) follows from Ben’s assumption that “propositions exist in any world only insofar as beings with sufficiently-developed minds express them.” Indeed, he thinks that propositions exist contingently precisely because he thinks the only minds that exist are contingent minds (but not necessarily human minds; I’m guessing Ben is open to the possibility of non-human contingent minds, e.g., advanced alien life forms).

(4) is stated directly in his post:

For since we consider multiple thoughts to express a single proposition p, then since those thoughts are not identical to each other they cannot all be identical to p. Instead, we need to say something to the effect that p is a sort of “similarity class” of thoughts, i.e. that the different thoughts among human beings all exhibit some similar structure or character, as we might say that there is only one Ace of Spades, even though it has multiple incarnations.

My impression is that Ben isn’t firmly committed to this account of propositions, but thinks that at least something close must be true. For now I’ll assume Ben is committed to (4) unless he expresses a different view. (In any case, what I argue below can be easily adapted to apply to similar views.)

As for (5) and (6), these are affirmed in the fourth paragraph of his post:

Regarding the first point of disagreement, we appeal to a distinction made by Robert Adams in his paper “Actualism and Thisness” (1981). Adams prefers to treat possible worlds as maximally-consistent sets of propositions which tell “world-stories,” that is, which describe hypothetical states of affairs imagined by us. We can then distinguish between a proposition p being true at a world w, whereby it appears in the set associated with w, and being true in w. The latter sort of truth involves the proposition not only existing here in the actual world where we can use it to describe a hypothetical state of affairs and assign it a truth value in that capacity, but also existence within w, where denizens of that world can express it and assign it a truth value from their own point of view. Given that propositions exist in any world only insofar as beings with sufficiently-developed minds express them, this distinction seems intuitive and meaningful, and hence required in order to avoiding conflating existence inside a non-actual world of some truth there with its existence here in the actual world. Indeed it appears to have been championed quite independently of Adams, including by myself before I read his paper, and by Kit Fine under the labels “inner” versus “outer” truth (cf. “Plantinga on the reduction of possibilist discourse,” 1985).

Some Problems

In response, I’m going to point out some problems for Ben’s proposal before focusing in on one significant issue he needs to address.

In the first place, his definition of possible worlds as maximally consistent sets of propositions is vulnerable to a Cantorian objection (put forth here by Bill Vallicella). That is one reason why Greg Welty and I used a different definition of possible world in our paper.

Second, Ben’s understanding of propositions seems to reverse the relationship between thoughts and their propositional content. Propositions are primary truth-bearers and possess original intentionality (i.e., they are intrinsically ‘about’ things; they are not ‘about’ things in virtue of something more fundamental). A thought such as my belief that squares have four sides is true in virtue of its propositional content; it is true because the proposition that squares have four sides is true. So in an important sense propositions are logically prior to thoughts.

According to Ben’s proposal, however, thoughts are prior to propositions, since propositions are a sort of “similarity class” of (human) thoughts. On this view propositions are true (or false) in virtue of the thoughts that constitute them. But this gets things back to front.

Third, it follows from (4) that if propositions exist they are abstract entities (thought-types rather than thought-tokens). As such, they’re at least partly external to any particular human mind, and they’re not identical to any human thoughts or parts of human thoughts. But elsewhere Ben has written:

Certainly you [Anderson] never say that propositions are conceptually out of reach. But you do seem to suggest that propositions are external to our thoughts, i.e. they are not themselves our thoughts or parts of our thoughts. To my reckoning, that puts them conceptually out of reach.

By Ben’s own reckoning then, propositions (as he construes them) are “conceptually out of reach.” (I don’t agree with his reckoning, but the point is that if it’s a problem for my position it’s equally a problem for his.)

Fourth, Ben’s proposal has the odd consequence that nothing was true before human minds (or any other sufficiently developed minds) came into existence and nothing would be true if all such minds were annihilated. Assuming for the sake of argument that the only sufficiently developed minds in the cosmos are human minds, if a nuclear holocaust were to wipe out the entire human race then it would no longer be true that 2+2=4, that the sun is more massive than the earth, and that E=mc2. Bizarrely, it wouldn’t even be true that a nuclear holocaust had wiped out the entire human race.

Fifth, there simply are too many propositions — too many truths — for them to depend on human minds in the way Ben suggests. For example, for every natural number N there is the truth that N is not a purple armadillo. But most of these propositions, even though we know they’re all true, have never been entertained by any human mind and were true well before they were entertained by any human mind. (The counter that they might have been entertained by superior alien minds doesn’t have much mileage to it, for reasons obvious enough.)

Moreover, Ben takes possible worlds to be maximally consistent sets of propositions. It’s widely accepted, I think, that sets are ontologically dependent on their members; so if these possible-world-sets exist then their member propositions must exist too. Yet surely no human being has entertained every proposition in the set that is the actual world, never mind all the other possible worlds. So whose minds furnish the metaphysical basis for all these (arguably innumerable) propositions?

Finally, there are difficulties in explicating the distinction between truth-at-w and truth-in-w. Ben suggests that p is true-at-w iff p correctly describes w (or more precisely, describes the state of affairs that would obtain if w were actual). Thomas Crisp has pointed out that it’s hard to see how p could be true-at-w without being true-in-w. Doesn’t it seem obvious that if p correctly describes w then p would be true if w were actual? But if that proposition were true then (as we argue in the paper) it would exist. Nevertheless, Ben has a response to Crisp’s objection:

It seems to me, however, that Crisp’s key premise that p describes w iff, were w actual, p would be true is false. Our intuitive understanding of what descriptions are informs our statements about descriptions — not the other way round. As long as we have such an understanding, we are not required, I don’t think, to explicate it in English, or to construct a definition in terms of possible worlds semantics. If this bothers Crisp (or Anderson), then we can do as well with the following: p describes w iff, were w actual, p would be the case.

The trouble with this response is that being the case isn’t a property of propositions, and even if it were, one would still be stuck with a property-bearing entity (bearing the property being the case instead of the property being true). So the existence of propositions in w still hasn’t been eliminated. Perhaps this problem can be patched up in some other way, but even then Ben hasn’t addressed the other “awkward consequences” of the truth-at-w/truth-in-w distinction. So the viability of this distinction remains in doubt.

A Challenge: Making Sense of Necessary Truths

I now want to focus on one particular challenge that Ben needs to address. Since he’s committed to (1) he owes us an account of what it means for a proposition to be necessarily true, and one that is consistent with his other commitments. Here I’ll consider four possible accounts and point out why each one is unsatisfying with respect to Ben’s position.

Account #1: p is necessarily true iff p is true in every possible world.

This is perhaps the most natural explication of necessary truth, but clearly it’s one Ben wants to avoid because (as I think he concedes) it entails the necessary existence of propositions.

Account #2: p is necessarily true iff p is true in every possible world in which p exists.

This account avoids the entailment of the previous one by employing a weaker conception of necessity. A proposition p is necessarily true if it cannot fail to be true; that is, there is no possible world in which p exists but is not true. The problem with this account (as we point out in our paper: fn. 30) is that it has some absurd entailments. For example, the proposition that propositions exist turns out to be necessarily true, despite the fact that (according to this view) propositions exist contingently.

Account #3: p is necessarily true iff p correctly describes every possible world.

This account would seem to be attractive for Ben, since it employs his idea of truth-at-w (as opposed to truth-in-w). It apparently avoids the entailment that propositions exist in every possible world, but the question is whether it’s consistent with his other commitments. Remember that on Ben’s view, propositions arise from, or are abstractions from, actual human thoughts. But how could actual human thoughts have enough content to support all necessary truths? How could the human minds that exist today (never mind in earlier generations) sufficiently represent every possible world so as to allow necessary truths to “come out right” (i.e., to actually have the modality we take them to have)?

Account #4: p is necessarily true iff p is a member of every possible world.

This account is similar to the last insofar as it also appeals to the notion of truth-at-w. Recall that Ben defines possible worlds as maximally consistent sets of propositions. Why not say then that a necessary truth is just a proposition which is a member of every one of these sets?

The problem here, once again, is that human minds simply aren’t numerous enough and complex enough to account for all the propositions that constitute all these possible worlds. The domain of possibilities is far, far greater than the domain of actual human thoughts. In mathematics alone, there are countless truths that have been, and never will be, entertained by human minds. So it seems clear to me that this fourth account, while it may avoid the unwelcome (to some) entailment that propositions necessarily exist, can’t be reconciled with the sort of conceptualism or quasi-conceptualism that Ben wants to endorse.

I’ve suggested four accounts of necessary truth, none of which seem viable for Ben in light of his modal and metaphysical commitments. I believe the burden lies with him to come up with a coherent alternative account that can be reconciled with (1)-(6).

The ball is in your court, Ben!

8 Responses to Could Propositions Exist Contingently? A Response to Ben Wallis

  1. Dr. Anderson,

    Hey, thanks for putting so much time into this! I may not get to respond in any detail for the next couple weeks, but I did want to let you know I saw it and will respond before the end of the month. Also some preliminaries…

    First a quibble: I’m not sure how relevant it is in this context, but technically a similarity class will reduce to our experiences—probably our thought experiences in particular. (Remember I’m something of a Berkeleyan idealist.) So my affirmation of (4) is going to be what you might call in-model, that is, within the bounds of a certain mental picture of the world. However we can always broaden the model, in which case the elements of the smaller model will appear in the larger model as ideas (sometimes thoughts). This is the reason I don’t challenge your premise that propositions are thoughts—because once we jump outside the (smaller) model they are. But when we restrict ourselves to talking within the model, they aren’t thoughts but just abstract objects (though still related to thoughts, even in the model).

    As I said, though, I’m not sure if that’s entirely relevant—just a quibble.

    Second, I’m glad you don’t want to characterize possible worlds as maximally consistent sets of propositions. Frankly, neither do I. Instead, I use the notion of MCSPs only due to a recurring timidity which results from fearing to flout too many popular conventions. I didn’t realize you shared my rejection of that characterization, but since you do, so much the better!

    Also, I can give you right away a general direction of how I’d handle the “fifth” problem. In short, I’m not going to say that there are infinitely many extant propositions. This is something we have to deal with in general, for example when considering the paradoxical set of all truths. You may be familiar with Patrick Grim’s argument against omniscience where he appeals to tricks of the sort you describe. I respond here (particularly relevant is the section beginning with “A defender of Grim might insist…”).

    Anyway, hopefully I can have a real response in the next few days. But it might take longer, until later in the month. In the mean time, thanks again for the comments! They are greatly appreciated.

    –Ben

    • Thanks, Ben. Take your time.

      Briefly, regarding the fifth problem, my main point is not that there must be an infinite number of truths, only that there are far more truths that could plausibly be grounded in extant human minds.

  2. Pingback: Friday Links

  3. Dr. Anderson,

    Okay, so the thing that was keeping me from taking the time to respond is over and done. So just to recap from my initial comment, we need to tweak position 4 which you attribute to me. I do not deny that a proposition is identical to any particular thought. However the way it might reduce to a thought would be in terms of idealism, so that we still are going to want to treat propositions, for the most part, as abstract objects in our most common conceptual models of language/truth/logic/etc. In such a model we would want propositions to pick out not the individual thoughts we have, but rather a particular structure of thought. This seems most easily accomplished by thinking of propositions as a sort of similarity class, but as you gathered I’m not completely wedded to this view. The important thing is that the proposition picks out certain features of thought whereby we can distinguish between people conceiving versus not conceiving a given proposition, believing vs. not believing, etc. Hence my Ace of Spades example. We have certain necessary and sufficient criteria (not necessarily explicit) for what an Ace of Spades must be, so that we can distinguish a physical object from being an Ace of Spades or not. Similarly, we need to have criteria for distinguishing a mental object from being a particular proposition or not. Those criteria, in one way or another, determine what we mean by a proposition in the abstract. I like the similarity class approach, but if you have a better idea I’m game for it. Unfortunately to say that propositions are divine thoughts doesn’t get the job done because it doesn’t set us up to pick out the criteria we need. (Additionally we have no reason, as I argue, to think divine thoughts exist.)

    Also as I mentioned in my last post, I don’t really agree with 5 or 6 quite as stated, though again they are not too far off from what I would affirm. In this case it is my fault for not being clear, and I’m sorry about that. You may have noticed that, in my initial discussion with Paul Manata in the other blog post, I discussed modifying the “maximally-consistent set” (MCS) approach, and in my own blog post I described it as being what Adams “prefers.” Well, all this noncommittedness on my part is due to the fact that I agree with you that MCS is not the best way to think about possible worlds. However it seems like most everyone in professional circles does this, and since all I have to offer in its place is an intuitive and imprecise idea of a way reality could have been, or something like that, I initially thought it best not to challenge the view. But since you already agree we should toss it, that’s great! The important thing here is that we distinguish between the stories (or descriptions) we tell about a world and the world itself. So for instance we could say that a proposition is true at world w if we use the proposition to correctly describe w, and that a proposition p only exists in world w if “proposition p exists” is true at w (this would involve conscious beings existing in w and conceiving proposition p). My apologies for not being bold enough to clearly state my position for fear of sounding even more crazy than I already do by rejecting the widely-popular MCS!

    Oh, and I do want to be clear about something else up front: You say that I want to argue against the necessary existence of propositions “so as to evade the theistic argument.” However this is not my motivation at all! In fact before I ever encountered your argument, I used the in/at distinction (using different terminology since at the time I had not yet read Adams, but expressing the same idea) in this blog post to resist philosopher Patrick Grim’s anti-theistic argument! Rather, my motivation is that I find the view intuitive, and cannot make sense of its denial.

    But for the most part you have captured my view well enough. I do agree with 1,2,3, and although I disagree with 4,5,6 in their current form, the requisite modifications are not too dramatic, as explained above. So anyway, with these initial differences in mind, I can proceed to address your six criticisms.

    (I) The Cantorian objection is aimed at that part of my post which does not accurately reflect my own view. If you don’t like set theory as a model for propositions, then great! Neither do I. But this is not necessary to defend in/at distinctions, as explained above.

    (II) You object that propositions exhibit original intentionality which is not derivative of anything underlying in their makeup. Certainly this view runs counter to mine, but it’s not clear to me how you justify it. When I read your paper I first took “intrinsic” intentionality (apparently mistakenly) to refer to a sort of inseparability of intentionality from propositions. So, for instance, we can consider the similarity class of thoughts whose content is the proposition “Mark Twain is Samuel Clemens,” and this similarity class seems to necessarily (or inseparably) be about Twain/Clemens, even though it is not a thought (in our immediate model, anyway). It’s going to exhibit the marks of intentionality (directedness and aspectual shape) you describe in your paper, and which you argue makes it possible to assign truth value. But it’s not originally intentional insofar as it depends on the intentionality of its underlying makeup for its own intentionality. (Also, please note that you don’t have to accept my view that propositions are similarity classes in order to merely see that similarity classes exist and appear to exhibit this inseparable intentionality.) Now, you may take the position that original intentionality is required to assign truth values, but I did not see an argument for this in your paper, nor elsewhere. Rather, I saw only your argument that intentionality in general (i.e. that which exhibits directedness and aspectual shape) is prerequisite to assigning truth/falsity. So as long as similarity classes have it, we’re in the clear as far as your argument goes for assigning them truth values.

    However, I should warn you that I’m not entirely convinced propositions really are ever about something in the strictest sense. For instance you might object that the similarity class I described just now is not intentional, but only marks out the intentionality of other objects (our thoughts). But the same criticism can be directed towards your view of propositional intentionality. For although we might be tempted to think of propositions as (at least sometimes) being about one thing or another, perhaps that temptation only comes from the fact that propositions mark out the intentionality of each individual thought which conceives it. So if we are sufficiently strict in our characterization of intentionality, then we can avoid similarity classes having it—but then we remove our support for the claim that propositions have it either. And if we are more flexible, then we can find abstract objects like similarity classes of thoughts of propositions which exhibit a similar (or identical) sort of intentionality as propositions themselves. So it looks to me that whichever approach you prefer—strict or lax—is going to undermine your objection here.

    (III) You object that as abstract objects, similarity classes are external to our thoughts and hence, on my view, “conceptually out of reach.” Now, first off I need to acknowledge that my criticism there was badly expressed. Sorry about that, but allow me to clarify: If I recall correctly, the problem I saw (and see) is that if we posit entities external to our own minds then we need to justify their existence inductively. Since we can’t justify the existence of God inductively (as I argue elsewhere) then we won’t be able to justify the existence of propositions as you conceive them (being divine thoughts). But this is far afield from where I’d want to go in this conversation, so I regret bringing it up before. Suffice it to say here that the externality abstract objects have is an in-model externality, but they still ultimately exist as ideas within our own minds.

    (IV) I’m not really sure what is the substance of your fourth objection. You suggest that it is “bizarre,” for example, that “it wouldn’t even be true that a nuclear holocaust had wiped out the entire human race” even if indeed that does occur. But given the in/at distinction, what would make it so bizarre? Perhaps you find the distinction itself bizarre, but it seems to me that it is rather quite intuitive! Do you not also intuitively sense the distinction? Maybe you think in the end it doesn’t pan out, but don’t you see the intuitive pull which caused at least two philosophers and myself to independently arrive at such a view? But in any case, even if you really don’t see this intuitive attraction (and I would beg you not to be too hasty in your reflection), the fact remains that others do understand the distinction, and so however bizarre or counter-intuitive you personally find it to be does not constitute a good reason for us to reject it. (I also don’t think it would be a good reason for you to reject it either, but that’s more than I require in my defense here.)

    (V) You object that, on my view, there would be too many truths for human minds to actually conceive them. But considering my rejection of MCS, this is not so. I do not take the position, say, that for every natural number n there is a proposition “n is a natural number.” Rather, the set of natural numbers is characterized by a procedure for constructing numbers of a certain kind, and similarly the set of all propositions of the form “n is a natural number” is characterized by a procedure for constructing propositions of the given kind, not some sort of collection of actually existing propositions.

    (VI) You complain that in my response to Crisp I don’t avoid invoking the existence of an entity which must bear properties. And you are correct! But the property of being the case (that p) is a property had by the world itself, and not some proposition describing the world. So it is quite true that if a world w were actual, then w would exist. But I don’t see how you can deduce from this the existence of a proposition at w. So this objection, at least, doesn’t seem to track. You mention other objections to be found in this SEP article section, but Crisp’s was the only one I found there which was not resolved in the article itself while still being applicable to my view (e.g. most of the others seem to rely on MCS, which you and I both reject). Did you have some other objection in particular in mind?

    So those are my six responses to your six objections.

    Finally, you issue the challenge of deciding on an account of necessary truth. I like account #3, but with at least one modification: A proposition p is necessarily true iff, for every possible world w, p is consistent with a correct description of w. This seems right to me, at first blush, and it does not require us to posit the existence of more propositions than there are thoughts in human minds. But let me suggest (again) to you that if you can find a problem with it, our response should be to work to fix the problem rather than throwing out the project. Just because it is difficult to properly explicate an account of necessity doesn’t mean we can’t exploit it. As long as we have an intuitive understanding of what we mean—and it seems clear that we do—that is enough.

    Anyway, thanks again for the feedback. Although a worthwhile project, it takes a lot of thought (no pun intended) to hammer out a robust account of these matters, and so I appreciate the prodding. In particular it’s nice to face genuine challenges (not so very common on the internet), which you generously provide. ; )

    –Ben

  4. One quick clarification: instead of using the potentially confusing term “inseparably intentional” I should have said simply always intentional. So in other words, a person who thinks that propositions are always intentional thinks that every proposition has intentionality. That’s what I originally thought you meant by “intrinsic” intentionality—an inability to find non-intentional propositions. But I see now that you mean original intentionality as you describe it here on this blog.

    –Ben

  5. Dr. Anderson,

    Thanks again for carrying on the conversation as long as you did. I was pretty disorganized, and looking back I see that I was downright confused on certain points, so it’s cool that you were willing to put up with that.

    Anyway, I thought maybe I could persuade you to answer two last (and related) issues for me, if you have time:

    First of all, on your view, how do we represent God’s proposition-thoughts with our thoughts? That is, if propositions are thoughts in the mind of God (external to us), and we cannot have those thoughts ourselves, then what is the relationship between our thought “of” a proposition and the proposition itself which exists as a thought in God’s mind?

    I’m tempted to treat thoughts, at least for the most part, as sort of self-contained things which have a characteristic structure, and which can satisfy a “type” (this is probably better terminology than “similarity class”). So we could represent God’s thoughts by having thoughts of sufficiently similar structure to His. Or, if not exactly that, at least we could say that our thoughts represent God’s thought if they both belong to something like the same thought type. I’m not sure how else we could make sense of *representation*.

    This brings me to my second issue. For if that’s how representation works, or close enough, then it seems odd that we should need an existing external object to embody the ideal structure or type we have in mind when we evaluate the structure/type of our own human thoughts. So for instance, if I have a thought, it seems to me that I recognize it to be “of” the law of noncontradiction by evaluating it against whatever standard I have in mind for distinguishing between thoughts of the law of noncontradiction and other thoughts. Maybe I say, “look, there I go mentally manipulating these logical symbols in this particular way I associate to ‘noncontradiction.’ ” But in the end it’s still my own standard I have to check. I don’t see why I need anything external to myself to do that.

    So my second question is, do you agree that we don’t (always) need an external object of reference to privately distinguish between one type of thought and another? Or do you maybe think we really do need an external embodiment of a *type* of thought before we can talk about types of thoughts?

    I’m not sure if this will interest you, so if not then no worries. But it concerns me, so I thought I’d ask.

    Thanks,
    –Ben

  6. Ben,

    Sorry for taking a while to reply. Addressing your last couple of questions:

    First of all, on your view, how do we represent God’s proposition-thoughts with our thoughts? That is, if propositions are thoughts in the mind of God (external to us), and we cannot have those thoughts ourselves, then what is the relationship between our thought “of” a proposition and the proposition itself which exists as a thought in God’s mind?

    This is just a special case of the general problem of how propositions relate to our thoughts. Thoughts (beliefs, etc.) are generally taken to ‘express’ or ‘contain’ propositions. But this doesn’t shed much light on the relation. Given the theistic model, I’d have to characterize it as some kind of similarity or instantiation relation. I don’t have much useful to add at this time. But it’s important to note that our theistic argument doesn’t require us to be more specific about the relationship between our thoughts and propositions. It’s based on an analysis of propositions as such.

    This brings me to my second issue. For if that’s how representation works, or close enough, then it seems odd that we should need an existing external object to embody the ideal structure or type we have in mind when we evaluate the structure/type of our own human thoughts.

    I’d say this is a confused picture of how we think. We don’t so much think about the propositions as simply think them (i.e., form thoughts with propositional content). When I believe that the moon is round, I don’t evaluate the structure/type of my thought, or check it against some external standard, or anything like that. (Actually, this would lead to a vicious regress.) I simply form that belief — a belief that has propositional content, and thus is related in some relevant way to the proposition that the moon is round.

    So my second question is, do you agree that we don’t (always) need an external object of reference to privately distinguish between one type of thought and another?

    I’m afraid I’m really not sure what you’re asking here, Ben. It seems to presuppose the picture above, which I reject as confused.

  7. Thanks for the response. I wasn’t suggesting that we have to evaluate the type of thought we’re having before we can have the thought. But if we want to classify our thoughts and distinguish them as being of this or that type, then we have to perform some kind of evaluation. I’m not sure belief is a good example, because that involves much more than just a moment of thinking. But in order to know that, say, I’m conceiving of a circle, I have to be able to evaluate my own thoughts at least well enough to conclude that I am indeed conceiving of a circle.

    While we don’t typically do this—instead we just go about thinking our usual thoughts without much reflection—sometimes we do want to classify our thoughts as being of this or that type. Indeed it seems that on your view you need to be able to distinguish between thoughts which represent proposition P and thoughts which represent some other proposition Q. Otherwise we have no means of deciding which proposition, if any, we are conceiving.

    In short, representations only seem to help if we can tell one representation from another. That’s all I mean by evaluating thought types. I have to wonder how that private process would be different for us regardless of whether the object represented actually exists external to us.

    –Ben