The Inga-Otto Thought Experiment

This entry was posted on 3/14/2007 5:02 PM and is filed under Metaphysics,Cognition.

Trackbacks

• No trackbacks exist for this entry.

Comments

• 3/14/2007 5:18 PM Neil wrote:
  It' difficult to know what force exactly these conditions either should have or are taken by C&C to have. I argue in forthcoming work that they are not necessary conditions. C&C themselves say that having these features makes it overwhelmingly plausible that Otto's notebook plays the right functional role for its contents to count as part of his memory: they do not say that they are necessary conditions. It's worth adding - and dialectically appropriate, since Ken argues not that extended minds are impossible, but that a a matter of contingent fact human minds are not extended - that an agent who does not automatically endorse thoughts is not a normal agent; Gilbert has some work on how endorsement is the default and automatic response, which is only effortfully overriden.
  Reply to this
  1. 3/14/2007 8:30 PM kenneth aizawa wrote:
     

     It is true that C&C do not explicitly offer their four conditions as both necessary and sufficient conditions for being a part of ones cognitive economy. One has to tyr to infer the intent of the conditions based on surrounding text. The last sentence of the paragraph preceding what is quoted above is “Is my cognitive state somehow spread across the Internet?” (C&C, 1998, p. 17), but the paragraph following that quoted contains the following, “The Internet is likely to fail on multiple counts unless I am unusually computer-reliant, facile with technology, and trusting” (ibid.) Reading the whole of those three paragraphs, it looks like the idea is that one's cognitive state is not spread across the Internet because the information does not satisfy the necessary conditions in the quoted paragraph.  If this is the incorrect reading, then what is the correct reading?
     
     I think I can agree that, in some sense, it is not normal to distrust the deliverances of one's cognitive apparatus, but C&C seem to be committed to more, namely, that this cannot happen.  It is in some sense impossible to distrust the deliverances of one's cognitive apparatus.
     

     
     Reply to this
• 3/14/2007 8:04 PM djc wrote:
  We don't put them forward as necessary and sufficient conditions. We say they are "features of our central case that make the notion [of belief] so clearly applicable there". This suggests something in the vicinity of sufficient conditions (though maybe not quite that), but certainly not necessary conditions.
  Reply to this
  1. 3/14/2007 8:33 PM kenneth aizawa wrote:
     Fair enough regarding this bit of text you cite.  How should I understand the text I cited in reply to Neil?
     
     Reply to this
     1. 3/15/2007 5:14 PM djc wrote:
        I'd say we're saying that it fails to satisfy a number of the conditions that made the Otto case a clear case, so it isn't a clear case. Maybe there's an implication that failing too many of these conditions is an indicator of failing to be a belief, and that therefore some disjunction of the conditions is somewhere in the vicinity of being a necessary condition (roughly, such that failing to satisfy such a disjunction is a strong indicator of failing to be a belief). But we're certainly not making a clear-cut claim of necessity.
        Reply to this
        1. 3/16/2007 5:58 AM kenneth aizawa wrote:
           What I am driving at, of course, is that in what Andy has called conditions of "trust and glue", the trust is not a contributor to what is part of one's memory.  Trust conditions should not be among the disjuncts. 
           
           Ok, so let me tweak the example.  Take Dotto1, who is a completely normal human being who trusts his brain-bound memory for faces, and who satisfies disjuncts 1-4.  Now let Dotto1 read a science article regarding memory which leads him to conclude, erroneously let's say, that memory for faces is not to be trusted.  Thereupon, Dotto1 becomes Dotto2 (Dotto2 = a later time slice of Dotto1).  Don't Dotto1 and Dotto2 have the same memories for faces?
           
           Reply to this
           1. 3/16/2007 6:16 AM kenneth aizawa wrote:
              

              Consider Dan and Doubting Dan.  Both Dan and Doubting Dan run through Ebbinghouse’s experimental protocols, being given lists of non-sense syllables and being asked to repeat them at various delays.  Dan and Doubting Dan produce exactly the same results on these tests.  They both utter the same outputs one the same lists at the same delays.  The only difference between Dan and Doubting Dan is that, for whatever reason, Doubting Dan thinks he is completely screwing up and never getting the right answers, where Dan has normal estimates of his performance.  Dan has normal memory for Ebbinghouse tasks, but it would seem that Doubting Dan does too.  Trust seems to be entirely irrelevant.  They both seem to have exactly the same information processing apparatus at work, the same apparatus that cognitive psychologists have tried to investigate for over 100 years.

              
              Reply to this
• 3/14/2007 9:41 PM Neil wrote:
  Surely there is a sense in which it is rationally impossible to distrust the deliverance of one's (all out) cognitive apparatus? One needn't be doubt free, but it seems that one must think that they are more likely than not. Moore's paradox looms if one rejects them.
  
  Here are further reasons why C&C conditions cannot be necessary (are not the mark of the mental):
  
  1. Constant availability: short term memory contents fail this condition by definition.
  
  2. Ease of access: when an Alzheimer's sufferer recalls something, the content is part of the mind. But they may frequently fail to recall that content. Indeed, normal agents may fail to recall content.
  
  3. These conditions are, in any case, only designed to capture one of the less interesting ways in which we extend our minds: storing representations. Extended means of manipulating information are more interesting.
  Reply to this
  1. 3/15/2007 10:24 AM kenneth aizawa wrote:
     In the scenario I have in mind Dotto does not think or say "That's Betty, but I don't believe it is."  Instead, he thinks or says something like "I'm inclined to think that's Betty, but I don't believe it is."  No Moore's paradox there.
     
     I think that in cases of visual illusions, to cite another type of case, we can distrust the deliverance of our cognitive apparatus.  My senses tell me that Twin A is larger than Twin B in the Ames room, but I don't trust those deliverances.
     
     I agree that the C&C conditions are not necessary (although there is reason to think that C&C think they are).  I'm not sure that your three points establish this conclusion, but I agree with the conclusion.  I hope we can come back to these three points.
     
     Reply to this
• 3/15/2007 7:31 AM Daniel Weiskopf wrote:
  Hi Ken. I admit that I'm not too clear on what 'endorsing' a piece of information is. Still, I wonder if the following isn't the thing to say about your example. When Dotto retrieves the name to match the face, he does remember the name. He even, in a sense, remembers that the person's name is 'N'. But he doesn't believe that the person's name is 'N'. Being a memory, then, might not depend on satisfying condition 3. Being a memory is just a matter of being an accessible informational trace of a certain sort. But being a belief requires more. I can have all sorts of things sloshing around in my memory that I don't necessarily believe. It's my taking those retrieved pieces of information in a certain way--the 'endorsing' way--that makes them my beliefs. What's wrong with taking C&C's conditions in that way?
  Reply to this
  1. 3/15/2007 11:08 AM kenneth aizawa wrote:
     Daniel,
     It looks like you are drawing a distinction between two ways of understanding the C&C conditions.  One is as N&S conditions on memory and another as N&S conditions on (dispositional?) belief.  You also seem to be suggesting that the conditions are not N&S for memory (which is what I maintain, so that you are with me so far), but perhaps are N&S conditions for (dispositional) belief (against which I must now provide further argumentation).   So far, so good?
     
     I would be reasonably happy to defeat these conditions as N&S conditions on memory.  John Sutton, for example, is very interested in extended memory, as opposed to extended belief.
     
     And what you say makes some sense of the text of C&C.  In earlier paragraphs, C&C move back and forth between talk of memory and belief, but that the immediate paragraph in which the conditions are introduced, the discussion is framed in terms of belief.  So,  endorsement may make more sense as a necessary condition on belief than for memory.  Agreed?  (But, there Chalmer's claim in response to my original post that the conditions are certainly not necessary conditions.)
     
     Ok.  So consider a new case.  Blindsight patients.  Do they believe, say, that the key is oriented in a particular way, even though they deny seeing that the key is oriented in a particular way?  Don't they believe it is oriented in a particular way, even though they do not endorse the claim that it is oriented in a particular way?
     
     
     Reply to this
• 3/15/2007 9:51 PM Daniel Weiskopf wrote:
  Hi Ken. I think Dave has said above that he doesn't take these conditions to be N&S; in any case, I've usually understood them to be roughly sufficient. (That's why I argued in my anti-C&C paper that they weren't really sufficient.)
  
  Anyway, yes, I agree that endorsement should be understood to apply to belief, not just memory generally. The blindsight case is an interesting one. I think that the vagueness attached to 'belief' is more or less precisely shared by 'endorsement'. So do blindsight patients believe that the line has a certain orientation? The possibilities are: Bel, End; ~Bel, End; Bel, ~End; ~Bel, ~End.
  
  We can rule out the first and last here, since you want to show that belief and endorsement can come apart. You propose interpreting them as being Bel, ~End. On this view, endorsement is stronger than belief. It requires some extra degree of commitment beyond being (say) in a state with a certain direction of fit. I think that what you're trying to get at is aptly brought out by the fact that they always regard their judgments as the product of a mere guessing process. If endorsement requires further epistemic certification, they don't seem to be endorsing, in the robust sense.
  
  Note that this isn't a threat if C&C's conditions are merely sufficient. But in any case, I think there is also a sense of endorsement that goes the other way. You can see this by interpreting them as being ~Bel, End. Here endorsement just means: verbally assenting to a sentence. This is, or can be, a relatively shallow cognitive act. You can endorse where you don't necessarily believe (or don't believe with the right degree of specificity, etc.). Here belief is the more robust attitude to hold.
  
  I think that this is just as easy a reading to get as the one you propose. But I'm not ultimately sure that either one is quite correct. That's because I'm not sure blindsight patients really do believe anything about the lines in their blind fields. After all, they are _guessing_. If I am asked to guess whether a coin will be heads or tails, and I say 'heads', I don't _believe_ it will be heads. In a sense I endorse the idea of its being heads, since I verbally committed myself that way. I am not endorsing in the more robust sense you may have in mind here, though. What I'm doing is, simply, guessing. And guessing is an attitude that isn't either belief or endorsement, in their robust senses.
  
  Anyway, that's my first thought on the matter. I realize that everyone thinks these patients can be induced to form beliefs about the contents of their blind fields. And in the shallow sense of belief, I agree. But I agree that they also endorse those judgments, in the shallow sense. Move to any more robust sense, and I don't think I agree. So belief and endorsement stand or fall together in these cases. In which case, they may not help against C&C's analysis.
  
  (Sorry if my replies to this thread are a little spotty. I'm traveling at the moment.)
  Reply to this
  1. 3/16/2007 7:14 AM kenneth aizawa wrote:
     There are, I think, other reasons for thinking that the conditions are not sufficient, but that's for another day maybe.
     
     The coin tossing case is guessing, but the blindsight cases apparently not.  The patients perform above chance when it comes to, say, finding the orientation of a key.  Take a real case.
     

     CLT is a 54-year-old male who had a right posterior cerebral artery stroke in 1987 (Fendrich, Wessinger, & Gazzaniga, 1992).  Magnetic resonance imaging confirmed a brain lesion to area V1 that deprived CLT of visual perception in most, but not all, of his left visual field.  In the experiment run by Fendrich, Wessinger, & Gazzaniga, CLT was asked to listen to two 0.6 second tones.  During one of these tones, a 1º black circle was flashed three times for 96ms on a white background.  Eye tracking equipment enabled the investigators to control the position of the display in the visual field.  CLT’s task was to make a forced choice regarding which tone co-occurred with the black circle.  When the black circle was flashed in his right visual field, he was correct more than 95% of the time.  When the black circle was flashed in his left visual field, CLT “insisted he ‘never saw anything’” (Fendrich, Wessinger, & Gazzaniga, 1992, p. 1490), but there remained a small island of preserved function in this left visual field, where CLT’s detection was significantly above chance.  

     Looks like CLT has beliefs in a loose sense, if you will, but in what (even loose) sense does CLT endorse the deliverances of his visual system?
     
     But, really, I think it would be a decent day's work to have moved the extended cognition folks to come out and say deny that there is extended memory and only extended (dispositional) belief.  That would be something of a concession, at least in my mind.  Andy sometimes hints at this kind of thing.
     
     
     Reply to this
     1. 3/16/2007 7:50 AM Eric wrote:
        I guess I am just agreeing with Dan here, but I dont see how any of these examples do any work to separate endorsement from belief. What they show, I guess, is that both belief and endorsement are idealizations. What we really have are credences and assignments of credences; and these are often less than full certainty.
        
        What the CLT example shows is that credences dont always match objective probabilities (frequency of being correct). But that's no surprise, is it? There are lots and lots of cases where people have credences that are lower than the freqencies with which they will be right
        
        
        What we would really like to see, if endorsement and belief are to be teased apart, would be a case where degree of endorsement didnt not match degree of belief, or credence. But that would seem to me to be just what "degree of endorsement" should mean. "Degree of endorsement" simply means the degree of belief, or credence, that I give to a piece of information as it becomes accessible to me.
        
        I dont see any reason to think there should be a difference in robustness between belief and endorsement.
        Reply to this
        1. 3/16/2007 11:08 AM kenneth aizawa wrote:
           

           It seems to me that there being no distinction between believing and endorsing (between a belief and what one endorses) would be bad for C&C. That would seem to make these sufficiency conditions (let us call them) for a belief pretty unilluminating. Condition three would then just be “upon retrieving information from the notebook he automatically believes it”  I don't see how you can defend C&C by effacing this distinction that they seem to need.
           

           
           Reply to this
           1. 3/16/2007 11:17 AM kenneth aizawa wrote:
              That is, the combination of, first, taking the conditions as sufficiency conditions on belief, rather than as sufficiency conditions on memory, and, second, equating endorsing and believing, would seem to trivialize the sufficiency conditions.
              
              Reply to this
              1. 3/16/2007 12:39 PM Eric wrote:
                 I'm equating endorsing and believing, tout court. I'm equating them with respect to epistemic status I assign them. There are claims I have endorsed that I dont believe, precisely because I dont remember them.
                 
                 What I am saying is that if something as all the properties (whatever those are) of being a belief other than the credence I assign to it, then my "endorsing" it ipso facto makes it a belief. Endorsing, is just, by definition, the process of taking something that has all the informational -integration properties of a belief and making it a belief.
                 Reply to this
                 1. 3/16/2007 1:43 PM kenneth aizawa wrote:
                    So, on your analysis, would you say that in the blindsight cases, we have instances in which all the informational-integration properties of a belief are present, but in which there is no endorsement, hence no belief?
                    
                    Reply to this
              2. 3/16/2007 3:58 PM djc wrote:
                 For us, endorse = occurrently believe, more or less. Obviously there's no problem with using the notion of occurrent belief in giving a sufficient condition for dispositional belief, which is what's at issue in the relevant case.
                 Reply to this
     2. 3/16/2007 8:38 PM Daniel Weiskopf wrote:
        The fact is that, although blindsight patients are better than chance, they're still guessing _in the epistemic sense_. They have no confidence that their (correct) statements about their visual fields are any more grounded than my judgments about coin flips. I don't get the sense that CLT's 'beliefs' are any thicker than his/her 'endorsements' in this case.
        
        (And I think I also 'endorse' what Eric has said elsewhere in this thread.)
        Reply to this
        1. 3/17/2007 8:18 AM kenneth aizawa wrote:
           The Dan-Eric line seems to be coming apart from the David line.
           One reason to think this: For Dan and Eric, believing that p is a matter of having what epistemologists sometimes call "a pro attitude toward p".  For Dan and Eric, endorsing p is this pro attitude.  Right?  By contrast, for David, where endorsing p  = occurrently believing p, endorsing is mere tokening.
           
           Another reason to think this: For Dan and Eric, it looks like endorsing p is an essential ingredient in believing that p.  Endorsing p is necessary for believing p.  By contrast, for David in an earlier entry endorsing p is not a necessary condition on believing p.
           
           Reply to this
• 3/16/2007 3:56 PM Eric wrote:
  If we assume that all the informational-integration properties of belief are present in blindsight cases (and I dont claim that they are--but it seems plausible), and we assume that the blindsight patients assign credences to the bits of information they report that are below the threshold, (if there is one) for what we would normally call belief--that is: they fail to endorse them--then it seems to me that answer would have to be yes.
  Reply to this
  1. 3/17/2007 8:48 AM kenneth aizawa wrote:
     David, can we more clearly formulate your understanding of the C&C conditions as one of the following?  (I think there is room for improvement in the statement of condition 3), but maybe you can comment on that too.)
     
     

     Version A:

     S dispositionally believes that p if S satisfies at least one of the following

     1) in cases in which the information that p would be relevant, S will rarely take action without consulting p,

     2) the information p is directly available without difficulty,

     3) S automatically endorses p, i.e. occurrently believes that p,

     4) the information that p has been endorsed at some point in the past, and indeed is there as a consequence of this endorsement.

     
     

     Version B:

     S dispositionally believes that p if S satisfies at least two of the following

     1) in cases in which the information that p would be relevant, S will rarely take action without consulting p,

     2) the information p is directly available without difficulty,

     3) S automatically endorses p, i.e. occurrently believes that p,

     4) the information that p has been endorsed at some point in the past, and indeed is there as a consequence of this endorsement.

     
     

     Version C:

     S dispositionally believes that p if S satisfies at least three of the following

     1) in cases in which the information that p would be relevant, S will rarely take action without consulting p,

     2) the information p is directly available without difficulty,

     3) S automatically endorses p, i.e. occurrently believes that p,

     4) the information that p has been endorsed at some point in the past, and indeed is there as a consequence of this endorsement.

     
     

     Version D:

     S dispositionally believes that p if S satisfies all four of the following

     1) in cases in which the information that p would be relevant, S will rarely take action without consulting p,

     2) the information p is directly available without difficulty,

     3) S automatically endorses p, i.e. occurrently believes that p,

     4) the information that p has been endorsed at some point in the past, and indeed is there as a consequence of this endorsement.

     
     

     
     
     Reply to this
     1. 3/18/2007 7:29 PM djc wrote:
        None of the above. We're really not trying necessary and sufficient conditions, or even sufficient conditions -- I think it's usually pretty hopeless to try to do this sort of thing. Of your versions above, the claim in the paper is probably closest to version D read as a sufficient condition (i.e. not reading 'if' as 'iff'), but again it's more of an indicator than a sufficient condition. As for necessary conditions and the like, I think our attitude is summed up pretty well by the immediately following paragraph:
        
        "Insofar as increasingly exotic puzzle cases lack these features, the applicability of the notion of "belief" gradually falls off. If I rarely take relevant action without consulting my Filofax, for example, its status within my cognitive system will resemble that of the notebook in Otto's. But if I often act without consultation - for example, if I sometimes answer relevant questions with "I don't know" - then information in it counts less clearly as part of my belief system. The Internet is likely to fail on multiple counts, unless I am unusually computer-reliant, facile with the technology, and trusting, but information in certain files on my computer may qualify. In intermediate cases, the question of whether a belief is present may be indeterminate, or the answer may depend on the varying standards that are at play in various contexts in which the question might be asked. But any indeterminacy here does not mean that in the central cases, the answer is not clear."
        Reply to this
        1. 3/19/2007 5:09 AM kenneth aizawa wrote:
           

           Ok. Then, would you say that, on this theory, S' s having a dispositional belief that kangaroos cannot fly is contraindicated by the fact that S has never had an occurrent belief that kangaroos cannot fly (condition 3) and that he does not have this dispositional belief that kangaroos cannot fly because of his having endorsed this in the past (condition 4)?

           
           Reply to this
           1. 3/21/2007 1:09 AM djc wrote:
              No. We're really not trying to give a theory of belief here. The kangaroo case is a different sort of case, involving beliefs whose contents are consequences of other things one believes, and it raises different issues.
              Reply to this
              1. 3/21/2007 4:47 PM kenneth aizawa wrote:
                 I guess I am losing a grip on what you mean by a dispositional belief now. I thought that one's dispositional beliefs could at least include contents that one could infer "relatively easily" from other things one believes. But, that's ok. Your project can be what you want it to be. Is there some X for which you are giving, not necessary or sufficient conditions, but some "indication" conditions? Is this what you are up to? If so, what is that X?
                 Reply to this
                 1. 3/22/2007 5:01 AM kenneth aizawa wrote:
                    

                    To elaborate, consider the following . . .
                    

                    Let's say that the belief that kangaroos cannot fly is a “faux dispositional belief.” It is not the kind of thing for which 1)-4) are meant to be “indicator conditions.” Instead, 1)-4) are indicator conditions for what we might call “genuine dispositional beliefs.”

                    Ok. C&C should be able to develop indicator conditions for whatever they wish. So, what is a genuine dispositional belief? There seems to be a tension. On the one hand, given what was said about the kangaroo belief, it looks like conditions 3) & 4) are somehow “built into” the concept of a genuine dispositional belief. They are “built in” in order to avoid the kangaroo counterexample. On the other hand, if conditions 3) & 4) are just part of the concept of a genuine dispositional belief, then conditions 3) & 4) look to be necessary conditions on something being a genuine dispositional belief, which is contrary to their purported status.

                    So, what won't work would be to say that a dispositional belief just is one that has been endorsed at some point in the past, since that would make condition 4) a necessary condition on a dispositional belief.

                    So, what C&C need here is to make an indicator connection between 3) and 4) that is tight enough to exclude the kangaroo example as not in the target, but not so tight as to make a necessary connection between 3) and 4).

                    
                    Reply to this
                    1. 3/22/2007 5:13 PM djc wrote:
                       (I wrote this a day or two ago, but it doesn't seem to have gotten posted.)
                       
                       Sure, the kangaroo belief is a dispositional belief. Again, we're not trying to offer conditions for being a dispositional belief. Rather, we're pointing to features of the Otto case that are highly relevant to its counting as a dispositional belief, and also observing that insofar as other relevantly similar cases lack these features, they won't count so clearly as a dispositional belief. My point in the previous message is that the kangaroo case is beside the point here as it isn't a relevantly similar case: ceteris isn't paribus, because (inter alia) the content is a consequence of other things one believes. (I take it that this is a familiar phenomenon: e.g. features X, Y, and Z are highly relevant to Monopoly's counting as a game, relevantly similar cases that lack these features count less clearly as a game, but quite different cases such as ring-a-roses may lack these features and nevertheless count as games in virtue of different features.) Obviously, if we were offering a theory of dispositional belief, one would need us to say more here, but we're not.
                       Reply to this
                       1. 3/23/2007 11:48 AM kenneth aizawa wrote:
                          So, do you mean to say, following up on the game story, that the kangaroo belief is a dispositional belief, only not by having the features 3) and 4), but by having the features 1) and 2)?
                          
                          Reply to this
                          1. 3/24/2007 3:27 AM djc wrote:
                             No. The point was that it counts as a dispositional belief partly in virtue of some features distinct from 1-4 (inter alia, the feature that the content is a consequence of other things one believes).
                             Reply to this
                             1. 3/25/2007 10:19 AM kenneth aizawa wrote:
                                You said that S dispositionally believes that kangaroos cannot fly is not contraindicated by the fact that S has never had the occurrent belief that kangaroos cannot fly or by the fact that S's having had an occurrent belief that kangaroos cannot fly is not part of the reason S dispositionally believes that kangaroos cannot fly.  Are you also saying that the fact that S has never had the occurrent belief that kangaroos cannot fly and the fact that S's having had an occurrent belief that kangaroos cannot fly is not part of the reason S dispositionally believes that kangaroos cannot fly does not at all count against the claim that S dispositionally believes that kangaroos cannot fly?
                                
                                And what about the tip of the tongue case?  The fact that S cannot recall that the name for the Australian teddy bear like creature that eats eucalyptus leaves does not contraindicate and does not at all count against S dispositionally believing that "Koala" is the name for the the Australian teddy bear like creature that eats eucalyptus leaves?
                                Reply to this
  2. 3/17/2007 8:49 AM kenneth aizawa wrote:
     Dan, Eric, can we formulate your vision of the C&C conditions more clearly as follows?
     

     D-E Version:
     S dispositionally believes that p if S satisfies condition 3) and one or more of 1), 2), and 4)

     1) in cases in which the information that p would be relevant, S will rarely take action without consulting p,

     2) the information p is directly available without difficulty,

     3) S automatically endorses p, i.e., takes the endorsing “pro” attitude toward p,

     4) the information that p has been endorsed at some point in the past, and indeed is there as a consequence of this endorsement.

     
     Reply to this
• 3/17/2007 5:33 PM Neil wrote:
  I'm surprised by Eric's responses here. First, given what he rightly says elsewhere about introspection, he shouldn't build such strong first-person authority into his belief conditions. Availability without difficulty seems to be an introspectionist condition (unless something like "available to guide behavior" is meant, in which case (2) threatens to collapse back into (1)(; automatic and past endorsement also suggestions something introspectionist. Eric should talk about actually guiding some significant proportion of responses, I think. What proportion? Well, that's the second thing that has surprised me about his responses. He seems here to want to put much more precise conditions on what counts as belief than he is entitled to, given his dispositional account. Satisfying a sufficient proportion of the dispositional stereotype associated with a belief is too vague to lend itself to such specification; indeed, it is likely to be highly contextual what cunts as sufficient.
  
  All that is ad hominem; let me add that what Eric should say goes for the rest of us too, since Eric is right about both introspection and belief.
  
  Does Otto satisfy a sufficient proportion of the dispositional stereotype to count as believing what's in his notebook? Obviously his is not a paradigm belief state. We shouldn't compare his state to mine in believing that today is Sunday. Instead, we should compare his dispositions to those of other cases which are more marginal, but in which it is plausible to count agents as believing. Blindsight cases are a good start. Some others: self-deception, Capgras cases, other delusions.
  Reply to this
• 3/18/2007 7:23 PM Neil wrote:
  Oops, I may have the wrong Eric. I took 'Eric' to refer to Schwitzgebel, not Thomson. Sorry.
  Reply to this
• 3/22/2007 5:21 AM kenneth aizawa wrote:
  Tip of the tongue phenomena:
  S is playing a trivia game. The question is “What is the name for the Australian teddy bear-like creature that eats eucalyptus leaves?” S groans, “Oh, I know this, oh, oh, oh ..., it's on the tip of my tongue.” One might think that this is a case of S having a “genuine dispositional belief” that “'Koala' is the name for the Australian teddy bear-like creature that eats eucalyptus leaves.” This “genuine dispositional belief” has occurred to S at some point in the past (condition 3) and this “genuine dispositional belief” is present because it was entertained at some point in the past (condition 4). But, it appears not to be readily available without difficulty (fails condition 2). It also looks as though S does not consult this information when it would really be relevant (hence failing condition 1). So, it looks as though S's “dispositionally believing” that “'Koala' is the name for the Australian teddy bear-like creature that eats eucalyptus leaves” is contraindicating by conditions 1) and 2).
  
  Knowledge of language:
  Maybe there is no innate knowledge of language, but perhaps only “genuine dispositional beliefs” about language. Non-linguist S “genuinely dispositionally believes”, say, that every sentence must have a determiner phrase and a verb phrase. Maybe this “genuine dispositional belief”is consulted whenever relevant and is readily available, but this is presumably not a “genuine dispositional belief” that ever occurred to non-linguist S and hence not one that is present because of its past occurrence (it is, after all, innate). So, S's having a “genuine dispositional belief” that every sentence must have a determiner phrase and a verb phrase is contraindicated by conditions 3) and 4).
  Reply to this

Leave a comment