 Okay, good morning. So, I guess you've all had a chance to express yourselves on the essay on what you make of all this. So, today is our last day in the causal theory of reference, and I'm just going to, in the first two sections today, wrap up on some relatively technical things about causal chains and informative identities. And then I want to look a bit at the broader implications of this. There's something about the causal theory of representation that really changes, it really affects your picture of the relation of the mind to your knowledge of your surroundings. I want to try and catch some of that in the last section today. I've been trying to restructure the class quite a bit, and after moving all the pieces around quite a lot, I realized that there are actually quite good reasons for keeping it the way it is. So, not for the first time. Okay, so on Friday, we'll go on to Putnam's article, brains in a vat. Okay, oh boy, okay, brains in a vat on Friday. Okay, so where we got to last time, we're talking about the function of language and why you'd want a language at all, and maybe the basic function of a language is to let you transmit knowledge from speaker to hearer. Then, if you're thinking of reference in that context, reference in the context of having a system that you can use to transmit knowledge about the things in your surroundings, then what reference to the object is going to require is that if you're going to refer to the object, that's the same thing as being in a position to transmit knowledge about that object. And if you think that knowledge is a causal notion, that in general knowledge of the object requires that you be causally connected to it, then that's just going to imply that reference to the object demands that you be causally hooked up to it. There'll be a causal connection between the thing and you. So putting that around the other way, reference to the object is going to require a causal connection to the object in virtue of which you're in a position to transmit knowledge of the object. I know obviously not any kind of causal connection is going to do for knowledge. I mean, if you just hit me with a baseball bat, then you will make a big impact on me causally, but I won't thereby be in a position to refer to you, if you see what I mean. Yeah, I mean, I'll be unconscious. So not any kind of causal connection is going to be good for reference. So this suggests that the kind of causal connection that's going to matter for reference is one in virtue of which you're in a position to transmit knowledge of the object. Evans's formulation was the item you're trying to refer to is the item that's a dominant source of your associated body of information. So the idea there was there might be two objects, one of which is your dominant source, the other of which is getting some information into the pool of information you associate with a term, but it's not the dominant source. That's the model of that Napoleon example is whoever's a dominant source of the information you associate with the term. That's the one you're referring to. Or I gave, I tried to do a photograph of two different people where you're mostly getting information about one of them, but some information from the other is getting in there. You'd say that's really a photo of the person who's a dominant source of the information in the photo. So in the case of perception, the dominant source of your current perceptual information, that will be the object that you're seeing. So you can refer in perception using terms like this table or that table, but if you think a little bit further about the way it goes in perception, if you take a term like this table or that person, then when you or I are looking at the table, there's a whole bunch of things that are actually causally involved in perception of it. For example, there's the sun, right? Our friend Mr. Sun is involved in here because it is light rays from the sun that are bouncing off the table and striking your retina. The window is involved because if it wasn't for the window, then there couldn't be that causal process of the light getting in. The lamp overhead may be causally involved. There might be lots of stuff, lots of things that are causally involved in the generation of your perception, but you wouldn't say just as you look at the table that you are seeing the sun or that puts you in a position to refer to the sun. The lighting might be kind of hidden, so you don't even know where the lamps are. You don't know what the illuminates are, so these things are causally involved, but you're not seeing them or in a position to refer to them. And when Evans talks about dominant source, it's not obvious that that really helps sort out why that is. And why is the causal connection that matters, the one between you and the table rather than the one between you and the sun? Every bit of information that you get from the table, well the sun is causally implicated in you getting that information. You see what I mean? So just saying it's the one that's involved in generating most of the information, doesn't disentangle the sun and the table. Have I stated that plainly? I mean you see there's a problem here that we want to specify more exactly, which causal connections matter for reference and which don't? Solutions? It's not that hard. Frankly, as you'll see in a moment, yes. Dominance source. Dominance source, yeah. Dominance source, the thing about that is that it depends exactly what you mean by dominant source, but in the face of it it sounds like something like quantity matters, yeah. But the sun, the table is involved in you getting lots of information, but every bit of information you're getting from the table, if it's only the sun that's allowing you to see the table, then the sun is causally involved in generating every bit of information that the table is involved in generating. So it's not clear why the table rather than the sun counts as dominant. It's not just quantity of information being generated. You see what I mean? There must be something right about that, yeah. And that's very intuitive. You're focusing on the table in a way that you are not focusing on the sun, yeah. But the trouble is to spell that out a bit, but I agree that that's the intuitive answer to this puzzle, yeah. The sun could be replaced for that. That's really an interesting idea. It depends what you mean by replaced, though. I mean, if the table was replaced by a duplicate, yeah, then I don't know, in some sense things would go on as before. If the sun was replaced by a duplicate, all right, but if the sun was just taken away, you wouldn't be able to see anything at all, yeah. Right, and if you take away the sun, you wouldn't see the table at all. Yeah, they're both involved. You see what I mean? Yeah. The causal immediacy of the table. Yeah, yeah, the sun was way downstream in the table, yeah. That too is an interesting idea, I mean, but what about our old friend the mirror, the unexpected mirror that you don't realize is there? Yeah. Reflection, yeah, right. Yeah, but you know, if I've got the mirror here, right, the unexpected mirror, then it's causally in between you and me, yes. But I don't even realize there's such a thing there, but I can see you in the mirror. If you wave to me and I wave back, that's not crazy, because I know it's you, right? I see reflection, yeah. I was trying to wave to you, if you see what I mean. I see you reflected in the mirror. Okay, one last, yeah. I'm sorry, I don't mean to make too, I think this is interesting, and all of these ideas are chipping away. Figure ground, yeah. Yeah, that is a bit like that focusing idea. You know, it's when the thing is lifted out for you as figure from ground, and you can choose what to focus on, yeah. Right, right. Let me try this on you. So this is just stating the puzzle I just gave. There are different causal roles for the object you're seeing, the thing illuminating it, and if you see it through a telescope or through a mirror. I see the force of saying you might just see the reflection there, but it's also equally colloquial to say I saw you through the telescope or I saw you in the mirror. Yeah, I mean, if you're looking at yourself in the mirror, you're looking at because you want to see how you are. You don't want to see how your reflection is. If you see what I mean, it's not just but you're interested in how your reflection is this morning. You want to find out how you are. Okay, so what about this? I mean, I don't know that this is a comprehensive answer to the puzzle, but something like when you're making a judgment about the table and you're saying the tables are brown or the tables rectangular, then you're responsive to the brownness or rectangularity of the table. You see what I mean? Whereas you can't make a judgment like that about the sun just by looking at the table. You see what I mean? And actually, there's a range of characteristics the table has. How big it is, just where it is. There's a stack of things you can find out. Is it shiny? Is it matte? There's a stack of things you can find out about the table just by looking, but you can't really do that with the sun just by looking at the table. So there's some sense in which your verbal reports about the table are differentially sensitive to the characteristics of the table, but they're not differentially sensitive to the characteristics of the sun. If the sun turns up or turns down, then you can see the table better or worse, but you're not varying your judgments about how the sun is. Flawed visual perception. Yeah, very good. Okay, so this way of putting it implies that perception is always accurate, which is probably a bit too strong. You're trained philosophers who live in a world of hallucination and illusion and madness and so on. Okay, but I make the following claim. In ordinary perception, if you're going to be able to talk about the cup at all, it's got to be pretty much the way it looks. What I mean is, if I seem to see a badger there, then it's not that I'm looking, I see the cup all right, but I'm just having some illusions as to what, it's not like straight or curved. Straight or curved are pretty, how should I say, they're pretty subtle distinctions compared to that kind of thing. If it really looks like a badger, then that's a hallucination. I'm not seeing the thing at all. I'm not able to refer to it. So unless I'm getting it mostly right about the thing, the thing I'm seeing in the distance, I'll write straight or curved as one thing, but roughly what size, what kind of thing, is it a fleet of aircraft or is it a car or an elephant? Vision can't be completely useful about these things and still leave you in a position to refer to the thing. Well, we have some ability to keep track of objects perceptually through changes in their characteristics. What I mean is, just as a person waves their arms and so on, there's change in the position, I can demonstrate this for you now, there's change in the position of the arms, but you have no trouble keeping track of the object through those changes. Whereas some changes in an object, you have trouble keeping track of the object through those changes. So with the case of the flowers over a period of days, it seems to me the spectrum of cases, if you're just looking at a bunch of flowers and they're, let's say, withering before your eyes or just waving in the breeze, then you're changing all right, but you have no trouble keeping track of them. So you don't need an identity judgment, that's the same flowers again, but sometimes you do, I mean, if you come back after a period of days and they're all changed colour, then you really might be a bit puzzled what have they been replaced or what? I'll say more about that kind of case in just a moment. Yep, very good. Let me give you another, can I put that in hold just a second, because I want to give an example that I think might be an example just like that, okay, and then come back. So you see how this works for proper names, but if you get all your information about Sally down a telephone line, I keep telling you about Sally, then I'm implicated, the telephone line is causally implicated, but the role that's a causal connection that Sally plays is different because it's Sally's various characteristics that are responsible for all the judgments you make using that term, whereas variations in the telephone line don't make a difference as to whether you're going to say that Sally is tall or not, that's okay. So there's something systematic about the role that Sally plays in that kind of case, the telephone line or your informants don't play, but this is what I think might be an example of what Dylan was talking about. Suppose we go back to our old friend, the sinister Goodle, so when you're making your judgments about Goodle, you might say well look my judgments about Goodle are all about the theorems that he proved and the proofs he gave of them, so I make judgments about how innovative Goodle's use of a recursion theory was and that is actually systematically reflecting what Schmidt did, you remember the story? The Schmidt did the proof and okay, but then my judgments about Goodle, using the name Goodle, are systematically varying with what Schmidt did. You see what I mean? Is that your kind of case? So why are we going to say isn't this guy really just the conduit through which all these different characteristics of Schmidt, the mathematician, came to be causally affecting our verbal judgments and in that case when we say Goodle, we'd be referring to Schmidt, that's the wrong answer because Goodle stole the credit. Do you see what I mean? That is Dylan's kind of case. Do you want to supplement that with another case? You're talking about reflections. Yeah, so this is a case where Schmidt is it where stands behind Goodle, so here you have Goodle, here you have Schmidt and here you have the verbal reports and Schmidt's causing Goodle to talk and Goodle's causing our verbal reports, but we're differentially sensitive to the characteristics of Schmidt, who we're talking about in the maths. And your kind of case is one where you've got Goodle, you've got a mirror reflection of Goodle and that's causing the verbal reports. Yeah, and you might say well we're systematically sensitive to what's going on in the mirror. Yeah, these are puzzling cases. You might say well this is Kripke's revenge, this is where Kripke comes back in because really to look at this, you can't address this kind of distinction in terms of dominant source or differential sensitivity because they're equally dominant, they're equally differentially systematically sensitive. So you've got to actually look at the details of the chain and in this case say but there was some funny business here, if you see what I mean, when Goodle stole the credit that blocks the reference from going all the way back to Schmidt or the mirror reflection is somehow transparent in a way that means that the reference does go all the way back to Goodle here. So Kripke might say well this just means you've got to look back at this historical stuff that Evans was saying forget about that. You see what I mean? Yeah, this is a homework exercise. I don't have a crisp sorting out of this yet. Yes, that's right. Yeah. Okay, suppose we take the Napoleon case where the swapover happened very late. Is that the kind of case you're thinking of? So we've got, as you put this mature, the mature imposter takes over very late. The thing is in the Goodle-Schmidt case, Goodle is really transparently transmitting, I mean conducting all these characteristics of Schmidt because whatever Schmidt does, Goodle just does that, replicates that and it's proof. So when we make the verbal judgments as to what we say Goodle was doing in his proof, we are actually causally sensitive to what this guy was doing. In the Napoleon case, there's the imposter and there's the emperor, the mid-period emperor and there are our verbal reports and they are being caused by the late imposter and the mid-period emperor. But one's not acting as a conduit for the other. You see what I mean? It seems to be a different, maybe you're reading it some different way but my picture is that here you're getting a lot of information from this guy coming down here and here you're just getting a little bit of information from this guy coming down here and that's what makes the distinction. You don't have that same kind of sneaky kind of structure as you do here. That's very good. I mean if you're going to say it's a way of transmitting knowledge of the thing, let's take that very seriously and now you've got Goodle, you've got this figure who is an imposter, a fantasist, one of the flotsam and jitsam of Vienna in the early last century. What he's doing is he's hijacking this sober series of mathematician stuff and projecting this fantasy onto it. Then you say, well is he a reliable source? No, he is not a reliable source. He's a kind of guy who goes around pinching mathematicians ideas. Although he might actually have just written it all down, so in fact it's all just the same way as Schmidt had it, he's not a reliable source so you can't get knowledge of what Schmidt was doing from this but you could get a knowledge of what Goodle was doing from this chain. Although this is a causal chain all right back here from Schmidt to Goodle. It's not, how should I say, epistemic. It's not, it doesn't have to do, it's not a chain through which we can get knowledge. You see what I mean? So that's lost here but this bit of the chain we do get knowledge and that's why we'd say we're referring to Goodle rather than Schmidt. That's really powerful I think. There's something right about that. I think from an Evan's point of view there's something about this though that is, it's against the thrust of Evan's account. Evan's idea, so this is a very good way of, I think something peculiar about the causal chain going through, back through Goodle to Schmidt in virtue of which you're differentially sensitive to Schmidt's characteristics. Yeah, it's not the right kind of causal connection to give you knowledge. Evan's I think when he's just talking about the dominant source of your current information, the idea was forget all that stuff about going back into history. So it's alien to the spirit of Evan's account and I think within Evan's own account you'd have to be saying something like, look what we know about Goodle is after all not just the mathematics. What we know about Goodle is something about where he was born, was he married, where did he live, did he have children. You have this picture of someone with a life and the mathematics is just a little bit of the life, I mean with all due respect and what matters is that Goodle was contributing, this person here was contributing to covering most of that biography rather than just the mathematics. So I don't know, these are two different lines and I think they both have something to them. So I don't know, I guess that's my idea of wrapping up, is to raise a lot of loose ends. I'll leave them as homework. Okay, anyone else want to take a shot at that? Okay, I think that is actually pretty much the state of play today. I think we don't have a tight necessary sufficient causal account of reference but something like that gives you a picture I think of where we are today. If you can think of something which you might well be able to or that this sorts out some of this then that would be great. Okay, so I want to just review where we are with informative identities, where we began. So where do we got to an informative identities? Remember, Frege, this is where we began in about, I don't know, a long time ago. A equals A holds a priori and according to Kantus to be labelled analytic while statements of the form A is B often contain very valuable extensions of a knowledge and cannot be established a priori. So then I showed you these distressing photographs. Yeah, when you say well that's informative alright and then you think about it a bit and you think well those are informative too. Yes, because it's informative. If you have a causal theory of reference then it's informative to be told that it's the very same thing that was causally responsible for this photo as was causally responsible for that photo and if you think that's a model of reference in general then it's very hard to see how if you have, if signs have their references fixed by causal chains, how could you ever get an uninformative identity? The same as the sign isn't going to be enough. So how are you ever going to get an uninformative identity? But it seems like we need uninformative identities and if you take this inference and you say well that needs an extra premise, does that need an extra premise class? Yes, what is that extra premise? Okay, so we got the morning star is the evening star. Okay, so now we've got an inference of three premises right? Is that valid or does that need another premise? Sorry? Is it good the way it is? Put your hand up if you think it's good the way it is. Yeah, you really have to say that. Put it in your hand if you think it needs another premise. I don't think that's a daft response if you think it needs another premise. The thing is if you said it needs another premise you'd really be in trouble. If you took seriously the point that it's very hard in a causal theory to see how the identities can ever be uninformative, yeah? Then what you say is well I don't after all, this is like the case of the photographs when I have the morning star here and the morning star up there. It's like the case of these photographs because I don't know if what's causally behind this use of the phrase the morning star and that use of the phrase the morning star is that the same thing? You see what I mean? So you might say this, I mean it's not a daft thing to say well you do need another premise because after all the causal chain back here might be going back to something different to what the causal chain up there goes back to. So in that case you would need another premise, namely something like the morning star is the morning star and similarly for the evening star, yes? But can you see the problem at this point? Because are you done adding premises? No you are not done adding premises, in fact you have only just begun, right? I mean you're really in trouble at this point because it's just going to keep exploding and the board is not going to be big enough, yeah? So at some point you just have got to be able to say that inference is just valid as it stands. We really need these uninformative identities. You couldn't have any kind of inference involving names unless you had, you could at some point stop and say I'm just going to use the fact of identity here in the inference. But then how do you ever get to that? I mean how could that ever be valid? If it's not ever valid then we really have lost the use of singular terms. We really can't use names for gossip and so on anymore. We should be tragic. But on the other hand this thing about how is it the same causal chain? That seems very powerful. I mean how do you know a priori that the causal chain is going back to the same thing? Well again this is like another loose end. I think there isn't a canonical answer to this. It is something that's still discussed as a kind of open problem in the literature. That's what I mean when I started that Frege's puzzle really looks obscure and must be fairly straightforward when you start. It is by no means straightforward. But let me just say something about how far I think we've got. It's something like this. What certainly seems a priori and maybe analytic is if you say this general form of argument a is f, a is g, a is f and g. Arguments like that. You can say this schematically the general form and it could be a priori that arguments of that general form are analytic. Just the same way you can say something like p, if p then q, so q. That general form of argument is valid. Now here I'm not talking about any particular argument. I'm just saying if you've got an argument of that form then it's valid. So here what Frege is right about is if you've got an argument of that form then it's a priori that is valid. But when you're arguing this table is f, this table is g, this table is brown, this table is rectangular, so this table is brown and rectangular, yeah? Then that is not a priori there that you really have an argument of that general form. And I'm not saying it's very unlikely but in principle it's possible that just as I talk some force from outside is zapping the table and replacing it with a qualitatively identical counterpart. I don't say it's likely but it's not a priori that it's not happening. And of course there are all these illusions and tricks and so on where people do substitute one thing for another and you don't realise, yeah? So this goes back to this thing about the flowers, actually changing. What seems to happen is that in perception when you're talking about the same person it might be that you and your evil twin are both in the class, yeah? But that you just move in and out so the two of you are never in the class at the same time, right? That's a possible scenario, yes? And then if people talk about you and say that person's f, that person's g, that person's both f and g, then that's not actually a valid argument, yeah? So when somebody does put information together from different sources, sorry, does put information together from different premises about you as true of the same thing, then you're assuming that that's not happening. There aren't being substitutions you don't notice. So you could say, well, what's happening is the general form of argument, what is this general form of argument require? A is f, A is g, A is f and g. Well, what that requires is that you have the same sense for A every time, right? It's not just the same sign, it's the same sense. But what does same sense require? Same sense requires that you've got the sign A and you've got a causal chain back to a particular object. So think of sameness of sense not as a matter of sameness of description or something like that. Think of same sense as a matter of your way of causally locking on to the object and keeping that the same so that if I say to you, watch this piece of chalk, watch this very carefully, right? You can do that, you can lock on to the piece of chalk and keep tabs on it over a period of time. That's what sense is. It's a way of causally locking on to an object and keeping tabs on it over a period of time so that if you've causally locked on to the same object in the same way all the way through here and you're just expressing that, then you've got a valid inference. It's not a priori that you write because you might have made a mistake. You might think you'd causally locked on to the same thing but not of causally locked on to it. But if you have, then when you say you've thrown that piece of chalk once, you've thrown that piece of chalk twice, you've thrown that piece of chalk three times, right? You're putting together information over time and you're using this perceptual ability to keep tabs on the sameness of the object over time. And something like that seems to go on in speech too. If you're talking about your friend Sally, you might say, Sally the singer, Sally the singer and just use a tag like that to so I can be sure it's the same Sally you're talking about over the period of time. It's not a priori that you're really managing, that I'm really managing to keep tabs on one in the same object but if I am doing it then I have the right, if I am managing to keep tabs in the same object, then I have the right to make this kind of inference without any further premises. I think something like that is where we are with informative identities. If you say that does not sound like a neat and tidy theory to me, I think you're right. It does not sound like a neat and tidy theory to me either. But something like that is a state of play. I don't know, maybe the GSIs, there's some cutting edge stuff that I don't, but I think that is where we have got to. Does that make sense anyhow? You see what I'm suggesting? That was the right answer. Okay, let me talk a little bit more broadly now about how this way of, these ways of thinking about reference affect your knowledge of your own mind. We talked about this earlier that you can have, it's very tempting to think if you just shut your eyes, close out the world, you can have authoritative knowledge of your own thoughts. You know whether you're thinking and you know what you're thinking just by inner reflection. It doesn't matter what's going on around you. But if you buy a causal theory then that kind of claim really doesn't seem so good anymore. You know what you're thinking. You know when you're thinking the same thought, when you're thinking a different thought, right? So it's not just that I know that I'm thinking. I know which things I'm thinking. But if I know which things I'm thinking, then I must be able to tell when I'm thinking the same thought or a different thought. And the trouble is that a causal theory seems to threaten this kind of idea. We came upon that earlier. But I want to just put this round another way. Suppose you say, look, let's suppose the causal theory is correct. I don't see why we can't hang on to the idea that even in splendid isolation I can know whether I'm thinking and know what I'm thinking. Yeah, let's try and do that. So in that case, suppose you shut out the world. You're sitting in your study and you think, well, what's going on? And you have a thought, as one does, about Goedl. You think, is that really right that Schmidt stuff? Goedl is innocent, you think. Goedl was framed. So what happens is you're struck by a particular thought about Goedl. And then you say to yourself, I'm thinking about Goedl. Okay, you can do that. Right there you are. That's you. I'm thinking about Goedl. And then you think, well, if I'm thinking about Goedl, then by the causal theory of reference, there must be such a person as Goedl. And Goedl must be causing me to think this very thought. So since I'm thinking about Goedl, Goedl exists. Right. And if you've read Kripke, then you know that. Okay. So there you are, locked away, thinking, and suddenly you've established the existence of Goedl. How do you do that? Right? I mean, of course, if it doesn't just work for Goedl, I mean, it will work for water, too. You are locked away with your own thoughts. You have a thought about water. You think, I'm thinking about water. Therefore, there is such a stuff as water. And you know that just by reflecting in the contents of your own mind. How did that happen? You realize now the benefits of knowing about the causal theory of reference because it's knowing about the causal theory of reference or lets you do this, right? Well, again, this is, it's not that I'm suggesting a tidy resolution here, but let me suggest a way of thinking about this. Suppose you've got it that you're thinking about Goedl and you get from that to Goedl exists. Well, maybe that transition is all right. I mean, if the causal theory of reference is right, then that is a kind of a priori transition, right? You couldn't be thinking about Goedl unless Goedl existed. You couldn't be thinking about water unless water existed. I mean, that's it's quite intuitive anyhow. But how do you know that you're thinking about Goedl? Well, maybe you could only conclude that you're thinking about Goedl if you already know about Goedl's existence. So maybe that thought justifies you in saying you're thinking about Goedl. But does it only if you already know about Goedl's existence? I suppose you didn't know whether Goedl existed. Suppose you thought that Goedl was just someone who figured in a thought experiment of cookies, right? Then it could happen that you were actually having a thought about Goedl, but didn't realize that this was a thought about Goedl because you thought it was a thought about an imaginary person in a thought experiment of cookies. Yeah. So only if you have knowledge of Goedl's existence, the first thing there justifies you in thinking the second thing. So you've got here an argument, a set of transitions that are correct, but they don't actually generate knowledge of the conclusion because you had to know the conclusion to make the transitions. Yeah. So that might be one thing that would block that kind of inference. And if it is, it's an example of a much more general phenomenon. This is in the trade. This is called transmission failure, where you start out with something you know, and you make valid correct transitions from that. And you get to a conclusion, but you don't know the conclusion on the basis of these three transitions, because you had to do the conclusion already in order for these transitions to be correct. So the reason I say it's quite general is you might think that something like that goes on with Descartes' Kajito. How about Descartes' Kajito? You guys all seem pretty clued up about Descartes' Descartes. Doodle Descartes. Okay. So Descartes' Kajito is kind of like that if you think about it. It's only these inferences point you from the thought out to the world, right? And that's what's weird about them. Descartes' inferences point him from the thought back to his own existence, right? That's what he was doing, I think, therefore I exist. But they seem vulnerable to the same kind of complaint, because what's going on here is Descartes has a thought. He thinks, that's me. I'm doing that. Look at that. That was me. I'm thinking. And then he says, so I exist. And then people said, well, can you just do that again? How did that prove your own existence? Didn't that in some way presuppose your own existence? And you could think that's got the same kind of structure actually. If you think how you get from just having a conscious thought to the judgment I'm thinking, well, if you already knew you existed, if you already knew about yourself, and then you have the thought. And then you can say, that's me. I'm doing that. I'm having that thought. Therefore I exist. But you had to know of your own existence to get the conclusion out. Do you see what I mean? Yep. If you didn't know of your own existence, and you're just struck by a thought, you wouldn't be able to tell it was you that was doing it, because you wouldn't know there was any such thing. Just as here, if you didn't know of Goedl's existence, then when you had a thought, you wouldn't be able to say to yourself, well, I'm thinking about Goedl. One other, I mean, I think this is, I guess what I really want to get at is, these inferences might strike you as pretty weird. And in a way, they are pretty weird, right, from your own thoughts to where the external world is. But they are part of a family of really interesting and that have the same structure. And they are really interesting and it's hard to diagnose exactly what's going wrong with them. Here's one other last case. I don't know if you'll know about Descartes, I guess. What about GE Moore? This is one hand and this is another. GE Moore gave the following proof of the existence of the external world. The skeptic questions the existence of the external world says there are no material objects. GE Moore gave a very long talk in which he discussed the nature of material objects and so on. And then said, look, here I can prove the existence of material world. This is one hand, this is another. Hands are material objects. Therefore, there are material objects. Therefore, there is an external world. That's a common reaction to all of these arguments. Hands are material objects. Therefore, there are material objects. Therefore, there's an external world. And his point was you can be a certain of the existence of a hand as you can of the premise of any skeptical argument. That's just where we come in. But you might think this argument too has that same structure. You get a perceptual experience. You make the judgment this is one hand and you say external objects exist. But your perceptual experience only lets you draw the conclusion this is one hand. If you already know that external objects exist, yeah? So Moore is trying to get from the experience to the external world and he faces that same structure of problems. So whatever you think of the particular diagnosis I'm giving in terms of transmission failure, I really want to just end by highlighting these kinds of inferences. These are really interesting inferences. Okay, so Putnam on Friday. Okay, thanks.