Knowledge and its Limits reading group: Session one, “Introduction” and “Chapter One: A State of Mind”
This is the first ’session’ of TEM’s summer reading group on Tim Williamson’s book Knowledge and its Limits. I will begin with some brief remarks on the introduction, then there will be a précis of chapter one.
I. Some remarks on the introduction
I will not say anything substantive about the introduction. All of the main lines of inquiry that TW lays out will be argued for in great detail later. We will consider the cogency of those arguments at a later date. Since this is the second time I have read this book, I do have some things to say about the introduction, however. When I first read the introduction (the first time I read the book), I was (i) puzzled about TW’s remarks in the introduction, and (ii) fairly sceptical that he would be able to persuasively argue for all of the things he claims in the introduction. However, now that I have read the book, I now realize that the introduction is rather marvelous. TW manages to introduce all of the main lines of argument is a fairly comprehensive way. Moreover, it very clearly written. In short, I think that this introduction is the best I have ever read in a philosophy book.
II. Chapter One: A State of Mind
1.1 The main claim that TW builds his whole theory on is that knowledge is a mental state. Obvious mental states include pain, pleasure, love, belief and desire. The former two (belief and desire) are central to much analytic philosophy. TW claims that knowledge is a mental state in the same way those other states are. More rigorously:
If attitudes are relations of subjects to propositions, then the claim is that knowing itself is a mental relation such that, for every proposition p, having that relation to p is a mental state. Thus for some mental state S, being in S is necessary and sufficient for knowing p (21).
In some way, most previous conceptions of knowledge have a constituent piece of knowledge as being a mental state. Viz., belief. However, on those views, being in a mental state, S, is only necessary for knowledge. TW’s claim is that being in some mental state, S, is necessary and sufficient for knowledge.
On most other views of knowledge, knowledge is thought to not be mental because it is factive; i.e., knowledge of p entails p. Initially, TW’s evidence that factive states can be mental is that they share so many features with non-factive mental states. Moreover, he claims that our pre-theoretical understanding of the mental includes factive states such as knowledge. He admits that that fact does not preclude us from discovering theoretical reasons for excluding factive attitudes from the mental. His goal in the first part of the book is to show that such putative theoretical considerations are not plausible.
1.2 TW starts with the common thought that one is always in a position to know what mental state one is in. This thought leads one to what TW calls the transparency thesis.
TT: For every mental state S, whenever one is suitably alert and conceptually sophisticated, one is in a position to know whether one is in S (24).
If TT is true, then knowledge is not a mental state. This is because one is not always in a position to know that one knows. Consider a familiar example: N.N. is a well-informed citizen who, at time T, knows that Lincoln is president. At T1, Lincoln is assassinated. At T1, N.N. does not know that Lincoln is president because it is no longer true. N.N. is not in a position to know that Lincoln was assassinated. Thus, she is not in a position to know that she does not know that Lincoln is no longer president.
TW argues that TT is false. In fact, he thinks that nearly no mental states satisfy TT. He gives two examples. First, imagine that you are watching a basketball game at T. You feel nothing but indifference about whom you would like to win. However, when the game ends at T1, you are disappointed that team x didn’t win. TW claims that at T you hoped that team x would win, and that you weren’t in a position to know that you hoped x would win. This example is peculiar to me. It seems to me that at T one is in a position to know that he hopes x will win. That is, if he reflected on whether he wanted to x would win, he would realize that he hopes x will win. This is because he would consider the counterfactual scenario where x loses, and realize that he would be saddened by that fact.
The other example he gives is believing v. fancying. Since (i) a key difference between believing and fancying is that when you believe x you rely on it in practical reasoning and (ii) in part, one’s dispositions to use x in practical reasoning can only be brought out counterfactually, it follows that belief is not transparent. Or, at least that’s TW’s claim. Again, I am unsettled. Wouldn’t one come to know that one believed rather than fancied p upon reflection simply by reflecting on such counterfactual situations? [I am going to skip the part about scepticism because we will have occasion to talk about that in-depth later. Of course, we can also talk about it in comments.]
1.3 TW now moves on to introduce another major claim of the book: knowledge is unanalysable. He begins this discussion by contrasting paradigmatic mental states with non-mental states. The germane non-mental state he analyzes is ‘believing truly (BT).’ He claims that BT is not a mental state because it has non-mental content; viz., truly. In other words, BT is the combination of a mental state (B) with a non-mental condition (T). This leads to a possible paradox for his account, however. The following brings out the paradox:
(i) The state of knowing it is raining.
(ii) The state of truly believing it is raining.
(iii) The state of believing it is raining.
The fact that (ii) is sandwiched between (i) and (iii) is prima facie strange because (i) and (iii) are mental states and (ii) is not. TW argues that it is not that strange. He offers an analogy. Consider the following:
(πi) The property of being an equilateral triangle.
(πii) The property of being a triangle whose sides are indiscriminable in length to the naked human eye.
(πiii) The property of being a triangle.
(πii) is not a geometrical property because it is partly contingent on the limits of human eyesight. However, (πi) and (πiii) are geometrical properties. Thus, this is analogous to (i), (ii), and (iii) and this is not paradoxical like (i), (ii) and (iii) are. Another way to see the analogy is by considering the following: every agent that satisfies (i) satisfies (ii), but not vice versa; every agent that satisfies (ii) satisfies (iii), but not vice versa. Similarly, every object that satisfies (πi) satisfies (πii), but not vice versa; and every object that satisfies (πii) satisfies (πiii), but not vice versa. Thus, TW concludes that there is no structural reason why something that is sandwiched between two mental states must be a mental state.
This discussion leads to a key point: since BT is not a mental state, and knowledge, ex hypothesi, is a mental state, knowledge cannot be comprised of true belief as a component part. For the same reasons, knowledge cannot be comprised of justified true belief as component parts. TW generalizes this point:
Such an argument applies to any of the concepts with which the concept knows is equated by conjunctive analyses of the standard kind… analyses of the concept knows of the standard kind always involve irredundant non-mental constituents, in particular the concept true…Given that the concept knows is mental, every analysis of it of the standard kind is therefore incorrect (30).
Thus, TW concludes that standard analyses must be incorrect. This argument bothers me a bit simply because he hasn’t rigoursly defended the claim that knowledge is a mental state. I don’t fully feel the pull of this argument because I am not persuaded yet that knowledge is a mental state. That notwithstanding, TW finishes this section with some general considerations that do convince me that the traditional program is bankrupt.
First, he considers the large importance we put on knowledge, and then asks: why should we care so much about knowledge if it just turns out to be a long, complicated, ad hoc list of necessary and sufficient conditions? I don’t know. I would think we wouldn’t find it to be so important if all it was was a long list of ad hoc conditions.
Second, he points out that many concepts don’t seem to have analyses in more basic terms. E.g., ‘means’ and ’causes.’
In response to the second point, he considers the rejoinder that perhaps we should think that knowledge is an exception to this trend because we already know what one of the necessary conditions is; viz., truth. He argues that this is based on a fallacy:
If G is necessary for F, there need be no further condition H, specifiable independently of F such that the conjunction of G and H is necessary and sufficient for F (32).
In support of this, he offers the example of ‘being red.’ ‘Being colored’ is a necessary condition for ‘being red.’ However, you cannot add something conjunctively to ‘being colored’ and have an analysis of ‘being red.’ TW states is conclusion about this thusly,
Necessary conditions need not be conjuncts of necessary and sufficient conditions in any non-trivial sense (33).
Let me just say that the second half of this section is brilliant. It literally gives me goose bumps while I read it.
1.4 TW next moves to a positive description of knowledge on his view. His main claim is that knowledge is the most general factive mental state operator (FMSO). A state is factive just in case, necessarily, ‘one has it only to truths.’ A state is stative just in case it is not a process.
FMSOs have five main characteristics:
(i) “Syntactically, an FMSO Φ has the combinatorial properties of a verb” (34)
(ii) “Semantically, Φ is an unanalysable expression” (ibid).
(iii) “Φ typically takes as subject a term for something animate and as object a term consisting of ‘that’ followed by a sentence (ibid).
(iv) “Φ is factive” (ibid).
(v) “‘S Φs that A’ attributes a propositional attitude to S” (35).
With these attributes in hand, we can now state TW’s proposal. His claim is that if Φ is an FMSO, then ‘S Φs that A’ entails that S knows A. So, if S sees that it is raining, then S knows that it is raining; if S remembers it was raining, then S knows that it was raining etc.
TW summarizes his discussion with three principles:
(1
If Φ is an FMSO, from ‘S Φs that A’ one may infer ‘A.’
(19) ‘Know’ is an FMSO.
(20) If Φ is an FMSO, from ‘S Φs that A’ one may infer ‘S knows that A’ (39).
(1 is true by stipulation. (19) is highly contentious. (20) is highly contentious, even from within the ‘knowledge as a mental state’ paradigm.
1.5 He finishes the chapter with a discussion of beliefs role in his account of knowledge. Since his account makes no explicit reference to belief, it might be thought that one does not need to believe p to know p on his account. Although the discussion is interesting, it is not central to the main points of the chapter. Briefly, he does hold that belief is necessary for knowledge. See 41-47 for his arguments.
I do want to mention some of his remarks at the very end of the chapter. He writes,
If believing p is, roughly, treating p as if one knew p, then knowing is in that sense central to believing. Knowledge sets the standard of appropriateness for belief… Knowing is in that sense the best kind of believing. Mere believing is a kind of botched knowing. In short, believe aims at knowledge (not just truth).
Does it follow that if belief aims at knowledge, then knowledge is the norm of belief? Since one is not always in a position to know that one knows, I don’t see how (although TW will have more to say about this later when he introduces his craz…hmm…controversial epistemic view about evidence).
[...] Knowledge and its Limits reading group has begun Jump to Comments For those of you who are interested, The Excluded Middle’s reading group on Knowledge and its Limits is now underway. Check out the first installment here. [...]
Errol,
Nice précis! A few comments before I have to run:
(1) “TW’s evidence that factive states can be mental is that they share so many features with non-factive mental states.”
I thought this was very weak. Sure, factive states share lots of features with non-factive mental states–that’s why they’re both *states*. But that doesn’t show that we should call the latter *mental* states; there is, after all, a big difference between the uncontroversially paradigmatic mental states–belief, desire, pain, etc.–and the factive states that TW wants to call mental: the former aren’t factive! Am I being too simple-minded?
(2) “Thus, TW concludes that there is no structural reason why something that is sandwiched between two mental states must be a mental state.”
I think Jonathan Lowe makes a very good point about this in his review of _Knowledge and its Limits_. Let me see if I can summarize it. The reason why the (πi)-(πiii) case isn’t paradoxical is that we have a nice explanation of how it could be that (πii) is a non-geometrical property, while (πi) and (πiii) are: the set of things with property (πii) can be partitioned into the set of things that have property (πi)–the property of being an equilateral triangle–and the set of things that *aren’t* equilateral triangles, but that look like they are to normal folks. The existence of this partition removes our surprise at the fact that anything with property (πi) has property (πii) and anything with property (πii) has property (πiii). But no such explanation is available for (i)-(iii). Williamson’s position explicitly blocks the partitioning; we can’t say that the set of true beliefs can be partitioned into the set of true beliefs that constitute knowledge and the set of true beliefs that don’t constitute knowledge. So, (i)-(iii) still seems paradoxical. I think that’s Lowe’s point (and if not–that’s my take!). It seems pretty good to me.
More later!
-Adam
Adam,
W/r/t (1), I should have mentioned the examples TW gives to vindicate the point about the two being so similar. He says that it would be very odd if there was a mental state for fearing, but not regretting; similarly, it would be very odd if there were a mental state of imagining but not of remembering. I feel the pull of those examples, especially the latter one. Imagining does seem to be incredibly similar to remembering in its ‘feel.’ Moreover, I see no pretheoretical reason to discount the pull of that similarity.
W/r/t (2), I don’t take that to be Lowe’s point (as long as we are talking about his review that was in The International Journal of Philosophical Studies, which can be found here). I take it that Lowe’s point is that the relationship between (πi) and (πii) is such that (πii) is a property that logically precludes that one could discriminate between the lengths of the sides. This property “has nothing to do with the powers of the human eye as such.” Thus, having the property of being an equilateral triangle does not essentially involve being indiscriminable to the human eye. However, it seems as if knowing essentially involves true belief. So, the point can be put like this: knowing seems to essentially involve a non-mental condition, whereas being an equilateral triangle does not seem to essentially involve a non-geometrical property. This does seem like a good point, although it is somewhat difficult to figure out what Lowe is saying.
Errol,
Thanks for the response. Here are a few thoughts.
(1) “I feel the pull of those examples…”
I very well might be alone on this one, but I just don’t find Williamson’s examples persuasive. I don’t think he’s given sufficient reason to discount the major difference between the paradigm mental states and the factive states–factivity. I’ll leave it at that for now. But if the position bears the fruit Williamson suggests it will, then that’ll be a powerful argument in its favor.
(2) “[It] is somewhat difficult to figure out what Lowe is saying.”
I agree. I found his talk of essential properties to be confusing, which is why I tried to avoid it when formulating what I took to be the objection. I think he’s on to something, though.
Adam,
I think I have thought of a better way of putting Lowe’s point. Let P be the state expressed by (i) and let P’ be the property expressed by (πi). I think Lowe’s point is this: the property expressed by (πii) is merely entailed by (πi). Let’s call this Q. In other words, (πii) is just some trivial entailment of (πi) that doesn’t have much to do with the limits of human eyesight (because it is trivially true that one can not discriminate between two things that are the same length). Disanalogously, the state expressed by (ii) is essentially a part of (i). In other words, without (ii) you cannot have (i). So, while the relationship between (πi) and (πii) can be expressed by saying ‘P’ (then) Q’. The relationship between (i) and (ii) can only be expressed by P.
Errol (and Adam),
Sorry for the tardy response, and I’ll probably elaborate more a little later.
First, I appreciated your comments on the introduction. I have not read the book before and I was also skeptical that anyone could pull off what the introduction promises. However, if anyone can it might very well be Timothy Williamson and I must say the intro definitely sparked my interest!
1) I think I echo Adam’s worry about factive states. I guess I have a hard time just picturing how a factive state could be JUST a mental state (at least, when the facts are not themselves solely “mental facts”). However, I think some of my worries will actually be addressed in the next chapter so I’ll suspend disbelief for now.
2) I was also troubled by the sports game example. I think TW included the “… whenever one is suitably alert …” condition to preclude your approach, so that if the subject were to reflect at T about whether they had a preference for either team to win, the subject would find no preference either way (because the subject was already “suitably alert”). I think this is what “suitably alert” must mean (something like “sufficiently reflective) in order to block the type of argument you present to show that we are in a position to know that we hope team x wins at T.
If this is right, then you can only “discover” that you had a preference after the game is actually over at T1. However, I wonder whether that really means you did hope at T that one of the teams would have won. If, upon reflection at T, you don’t discover a hope that either team wins but still feel disappointed at T1 when one of the teams loses, why does this mean that you did have a hope for a particular outcome at T? Perhaps the end of the game itself was what made you feel disappointed. If you can’t realize that you hope for team x to win at T even after reflection, then why should we think you did have that hope at T, rather than a wish(?) at T1 that the game would have ended differently?
More generally, I guess I would be suspicious of (some) attributions of mental attitudes to a person that are not “discoverable” by that person upon reflection, which seems to be what TW needs. However, I’ve generally thought that knowledge does not entail knowing that one knows (I sympathize with dretske and nozick here) and that one can know p without being in a position to know that one knows p. So, even if it looks like we will have to wait until chapter 4 for TW’s argument that no non-trivial mental states are transparent I guess I’m already on-board with denying that knowing is a transparent mental state.
3) Another worry I had was I was having a hard time seeing how some of the other attitudes TW proposes as FMSOs should really count as such.
Two, non-knowledge paradigm cases he cites are “seeing” and “remembering.” However, neither mental attitude seems (to me) to count as “factive.” His example for how a non-factive state (ie, one where we don’t get deductive validity for’ S thetas that A’ to ‘A’
functions is “guessed”
(1) I guessed incorrectly that he was guilty; and
(2) I guessed that he was guilty and you guessed that he was innocent
TW points out that the substitution of ‘knew’ for ‘guessed’ in (1) or (2) yields a contradiction. However, it doesn’t seem that a substitution of ’saw’ or ‘remembered’ into similar sentences also give rise to contradictions. Consider
I saw incorrectly that there were trees and and a pond in front of me (really, it was a mirage).
I remembered that I had won at Halo the last time we played, while you remembered that you had won the last time we played (we checked the game log, and actually you had won).
(1′
(2′
Probably it is just that my intuitions need sharpening, but neither sentence appears contradictory to me. Rather, I guess I think of both “saw” and “remember” as attitudes we have towards both true and non-true propositions, and so as non-factive.
TW’s response to this type of worry, I think, is to separate “seeing that A” from “seeing a situation in which A” (and analogously for remembering, on pages 38-39). First, only the former requires that you grasp the proposition A. However, it is possible to see a situation in which A and be able to grasp the proposition that A. Here TW makes a second claim : “[Y]out cannot see that it is raining, precisely because you do not know what you see to be a situation in which it is raining).
If that is all, this leaves the explanation for why “seeing that” and “remembering that” are factive fundamentally by appealing to knowledge. This is a similar feature to TW’s characterization of alleged FMSO’s “could feel” and “could hear,” which are (roughly) characterized as “She knew by the sense of touch/hearing that.”
If all FMSOs are fundamentally just ways of knowing, then why posit other, separate factive mental states? Why not just “knew that x by way of y,” where y could be the sense of touch for “could feel” or the process of recollection for “remember” (although knowledge is not a process, surely the result of a process could be a way to acquire knowledge).
I guess I just don’t see the motivation for saying that there are other factive mental states (even granting that knowledge is such a thing). But I have a more-than-sneaking suspicion it is only because I don’t understand fully how mental states can be factive in the first place and perhaps some naivety on the part of my intuitions about the philosophy of mind, which I presume the next few chapters might help out.
Anyways, sorry for the late posting again, and I thought the write-up was very good, so thanks for getting this all going. Even if I haven’t drunk the kool aid yet, I’m enjoying reading the book so far :).
Oops, “phi” for “theta” in my post above.
Another post on the run…
Errol,
I like your formulation of the objection, and I think it’s pretty powerful.
Jordan,
Re: (2) This is a nice objection that’s similar to what I was thinking about that section.
(3) Very interesting. I’ll think more about this.
Oh, and I should note that I *am* enjoying the book, even though everything I’ve said so far has been critical…
-Adam
Jordan,
Re: (2) I think that the whole entire book hinges on chapter 4. This is because if he cannot show that luminosity is false, then we should not think that knowledge is a mental state because we are obviously not always in a position to know that we know. I think that his argument fails in chapter 4 (not that I have a positive argument for luminosity. I just don’t think he has shown us that luminosity is false). I will present my argument for that in a few weeks (I have a paper on my personal blog that elucidates the objection).
Also re: (2) I think you are right to say that one does not really hope that x wins at T if one is not at all aware of such hope at T. That seems silly.
Re: (3) When I first read the book I definitely didn’t think “see” is factive. I am still ambivalent about this, although I am more willing to accept it now. I think that a main reason why I am inclined to think that it is not factive is that I am often pulled hard by the force of scepticism, and thus am very weary of the power of our senses. TW certainly does not think this way.
I am fairly convinced, however, that “remember” is factive. 2′ seems to me to be clearly a contradiction. When you remember that x took place, x must be true and you must know x took place. It is very intuitively plausible to me that this is true (although I am having hard time constructing a good argument right now!).
Adam,
I was wondering why you found Lowe’s objection so compelling. I guess I don’t grasp the prima facia reason why a mental/non-mental/mental “sandwhich” should be objectionable. I’m sure I’m being dense, but I’d be interested in hearing what you find problematic about it.
Errol,
2) I guess I don’t know why luminosity should be correct, either. Being in a position to know that we are in a particular mental state may be a mark of mental states but I don’t know why it is always the case that we are in such a position with respect to those mental states. It’s certainly going to be interesting to see how TW argues against it, and I’ll look forward to your argument when we get there.
3) Again, I might just have poor intuitions regarding remember. I think I feel the same skepticism towards memory that you initially felt towards sight. It certainly seems like we have “false memories” that are phenomenologically indistinguishable from “true memories.” I think TW can get around this (apparent) fact, but it trips me up.
Jordan,
Here’s another drive-by posting. See if this makes any sense. I have a funny feeling that I’ve missed something obvious…
I’ve been mulling over your question–what’s so paradoxical about mental/non-mental/mental sandwiching? I guess my problem has to do with the “top” of the sandwich–true belief being a necessary condition for knowledge.
On Williamson’s view, knowledge isn’t a kind of true belief, but there is a necessary condition between K and TB. My question is this: what explains that necessary connection? Is it just a brute fact that there is such a connection? Or is a better explanation that K is a kind of TB? Think back to Williamson’s example about being red and being coloured. Necessarily, anything that’s red is coloured. That connection would seem awfully strange if we didn’t have the simple explanation that we do–red is a kind of color. But on Williamson’s view, there’s a necessary connection between two kinds of (at least partially mental) states—K and TB–that doesn’t get explained in this way or any other; that strikes me as weird. That’s partially because I can’t think of another necessary connection b/w two mental states that’s not explained by appeal to the fact that one is just a particular kind of the other; can anyone think of such an example?
(Interesting thought: one might object by saying that seeing and remembering are such states (recall Williamson’s stuff about K being the most general FMSO—that provides the necessary connection between those and K). But I can respond by taking Jordan’s position—seeing and remembering in the senses Williamson talks about are just kinds of knowledge!)
-Adam
Adam,
Thanks for indulging my question :).
Anyways, as for a connection between two mental states that fulfill a similar “sandwich,” I would think any of the “could x” FMSOs would serve such a function, where “x” would be the other mental state in conjunction with “truly.” Consider “could feel”
1) She could feel that that her bone was broken
2) She felt that her bone was broken and her bone really was broken
3) She felt that her bone was broken.
2) necessarily follows given 1) if “could feel” is a FMSO (if it is factive, then it is true that the bone is broken) and the mental state “could feel” entails the (different, non-factive) mental state “felt.” 3) trivially follows from 2).
The above looks like it satisfies the sandwiching conditions, right? It relies on us classifying “could feel” and “felt” differently, as TW does, and considering “could feel” to be an FMSO. So, on the one hand it seems like we should get a similar sandwich for every FMSO (as long as the FMSO entails us having some other mental state as well).
On the other hand, this probably doesn’t help all that much, since the “could feel” example is likely just as problematic as the “knows” case. I think this might point to an answer : it isn’t that K is a kind of TB, but rather that B is a kind of K (a botched kind, according to TW). This explains the B part, given K. K is also factive according to TW, and necessarily a factive attitude is one we only have to truths. So this explains the T part. Put them together in a conjunction and we get the non-mental state TB. Does that sketch sound promising?
I think the main issue is how mental states can be factive. If a mental state X is factive (and thus entails truth), and that mental state X entails some other mental state Y, then a factive mental state X will necessarily give us “Y & true.” And, of course, if we have “Y & true” we get “Y” for free (and the bottom half of the mental sandwich). I guess I don’t think the problem is really “the sandwich” but rather whether a mental state can simultaneously entail some other mental state as well as some non-mental condition. If so, there is no paradox (the other mental state in the non-mental-state conjunction will trivially entail itself). If not, then TW is in trouble.
In other words, I think if we grant TW that there are factive mental states, then we should also grant him that there will be “mental/non-mental/mental” sandwiches. But I also feel like I might be missing something.
Jordan,
I am planning on replying to your post. It might be a day or two, though.
Jordan,
“So, on the one hand it seems like we should get a similar sandwich for every FMSO (as long as the FMSO entails us having some other mental state as well).”
Like you point out, this is not going to help with the puzzle. This is because every other FMSO’s sandwich is going to be the same know’s (I think). The other relevant mental state is going to be belief, and the non-mental condition is going to be truth. So, for ‘remember’:
(i*) The state of remembering phi.
(ii*) The state of truly believing that phi happened.
(iii*) the state of believing phi happened.
This is just the same as the knows case, and, at least to me, it is just as paradoxical.
Here is why these sandwiches are paradoxical to me. It is because of the conceptual connections between (i) and (ii). It is odd to me that (ii) must be present in all things that satisfy (i), yet (ii) is non-mental and (i) is mental. I think that that idea borders on incoherent. Moreover, I don’t think TW’s analogy works (for the reasons Lowe points out). Thus, I am left puzzled by the sandwich.
Errol,
Would you agree that the problem, then, is how a mental state can be factive in the first place? It seems that any FMSO will present a similar “sandwich” problem for us, and any FMSO will impose the non-mental “truth” condition. The problem of how (ii) is present in all things that satisfy (i) is therefore a problem of how a mental state can be factive, that is, of how a mental state can be such that it provides us with a non-mental condition (truth) in all cases.
If you do, then I think I agree : I am having a hard time wrapping my mind around how a mental state can be factive, and thus lead to non-mental conditions in all cases. Presumably TW’s answer lies in his commitment to externalism (which we’ll cover in the next two chapters). If internalism is true, then mental states are all contained solely in the head, and it seems incoherent to maintain that these mental states could always lead to non-mental conditions. If internalism is false, though, then mental states are not just physical brain states but also “something else” about the world external to ourselves. So even if FMSOs cannot be reduced to neat mental and non-mental components, perhaps this “something else” could still suffice to ground a non-mental condition for the various FMSOs.
Jordan,
Yes, I think that it is a general problem with FMSOs. It is more acutely a problem when K is the most general FMSO. My point above was just that you cannot cite the structure of other FMSOs to explain away the strangeness of the sandwich. In fact, it seems pretty tough for TW to cite other evidence to neutralize the strangeness. The strategy he does adopt–analogy–is probably the best way to go. However, I think the analogy fails.
You’re right that it is going to turn on the externalism/internalism issue. I look forward to that debate, and I look forward to your précis tomorrow.
[...] TW starts by setting the context. As we saw in chapter one, TW’s main claim is that knowledge is a mental state in the most general, fundamental sense [...]