 Good morning. Good morning. So today we're carrying on with Evans' version of the causal theory of reference. Immediately after, we'll have one more session on Evans on Wednesday. And then I'm going to change the schedule. And we'll move on to Wittgenstein on following a rule, the central passages of investigations right after that. We'll do basically the same. I haven't yet made change the syllabus on the website, but I will do that by tonight. Okay, any questions about that? So straight on to Wittgenstein after this. Okay, so I want to begin just by going over quickly the stuff we did last time and making a remark about how that applies to natural kind words. Then a comment on what seemed to me the most, it's a little bit technical and a little bit hard to see what he's after in Evans' official formulation of his theory. And then what I hope we can spend most time on is putting some perspective on the causal theory of reference, because one thing you might have thought about causation in reference is what do they have to do with one another? Why is this happening? I mean maybe it's kind of plausible that reference requires causation, but is there a bigger picture in which you can understand why it is that causation seems to be a key notion? Okay, so last time we were talking about this example of Evans' here we have the youth, the famous emperor, and 1793, just before Waterloo, and most of the information that we have about Napoleon comes from this mid-period, the emperor of Europe, the great military strategist and so on. So if we have an individual, an alpha there, who covers this part of Napoleon's life, and then is replaced by an imposter, at this late stage, if beta comes in, at this late stage pushes Napoleon into a ditch and takes over the three-cornered hat right here, and we say, did Napoleon fight the battle of Waterloo? Then the answer is, no, Napoleon did not fight at the battle of Waterloo, right? This place was taken by an imposter. If on the other hand, the beta had taken over much earlier and had shoved the youth into a ditch and then gone on to become emperor of Europe and so on, and then you say, well, in that scenario where beta takes over at a much earlier stage, in that scenario did Napoleon fight at the battle of Waterloo? Then the answer is, yes, right? No? If what happened was that this youth was shoved into a ditch at the age of 21 or whatever, and then the guy who shoved them into the ditch won all the famous victories, unified Europe under the French Empire, then when we talk about Napoleon, who are we talking about? The youth or this guy? This guy, right? And did this guy fight at the battle of Waterloo on this scenario? Yes, so did Napoleon fight at the battle of Waterloo? Yes, right, okay. So that's the idea that what matters for reference is who's the dominant source of the descriptions you associate with the term? Who's the dominant source? The guy in the middle is the dominant source, so whichever one he is, that's who you're talking about. Any questions about that? Okay. There are also these cases of reference shift. Spot the dog was our early example. Evans also gives this case of Madagascar, a hearsay reporter of Mali or Arab sailors. And then you can understand what's going on in this kind of case by saying, what's going on here is that there's a shift in the dominant source of the information associated with the name. So what happens is that when the Mali or Arab sailors are talking, the dominant source of the information that they associate with the name Madagascar is a bit of the mainland as Marco Polo takes it over and uses it as a name of the island. The dominant source of the information we associate with the term comes to be the island. Was that too fast? Yeah, so you got a shift in reference there. So this is a kind of key example because it brings out what's wrong with Kripke's theory and write about Evans is why Evans's theory beats Kripke's because what's going on here is that you had an initial dubbing of that part of the mainland and everyone in the causal chain since then intended to refer to the same thing as the people before them. Everyone was doing their best to keep continuity with reference. This is not like the case in which I call my dog Wittgenstein where I'm not even intending to preserve reference. These people were all intending to preserve reference but the reference shifted nonetheless. So as against Kripke, reference is shifting despite there being a causal chain all the way back to the mainland and an intention to preserve reference all the way through. But what's happened is there's a shift in the information associated with the NEM. If I'm explaining this correctly that should be absolutely plain. Whether you agree or not the argument should be absolutely clear. Yeah? Okay. So what preserving reference demands is not that you intend to refer to the same thing as the people before you. What matters is that you keep the same dominant source for the body of information you associate with the NEM. Now you can apply all this stuff to natural kinds, words like water and gold and tiger and so on. Now what happens in twin earth is you've got a dossier of information associated with the word water. Participates in the evaporation condensation, precipitation cycle clear quencher stuff and so on. So you've got a similar dossier on earth as you have in twin earth but on earth the dominant source for that information is H2O. On twin earth the dominant source for that information is XYZ. So if you moved from one planet to the other then at first the reference would be the same because the dominant source of the information you associate with the term would be the stuff in your home planet but as time went on the dominant source of the information you associate with the term would be the stuff in your new home. So your reference would shift slowly from earth to twin earth. Yes? Okay you run a marathon in earth. What I've just said implies that you would be wrong but if you say well that's obviously not the right answer then I can fudge a little bit to try and accommodate that. So you run the marathon you drink the water it's just fantastic that memory is burning your brain. Whenever you think of water that comes back to you you're lost called that experience of drinking water you always wish you could quite recover. Something like that? Right. Aha. Okay. Well the thing is one of the disagreeable things about this theory is that there is room for a lot of flexibility in the notion of the dominant source and as Evan says the reasons for your interest in the thing might affect which information counts at which causal source counts as dominant. I suppose we ham up your example a little bit and suppose that what you're always trying to do is recapture the context of that perfect drink then I think it would be fairly intuitive that the dominant source is that early encounter and it stays that way even though time changes. Someone who just forgets drinking water over time on the other hand there would be no point in saying the dominant source has to do with what was happening 20 years ago. Your case is a little bit difficult because it falls somewhere in between. Yeah. I mean if you say well you really want to be in a position to say look you're wrong water doesn't come out of the taps here water doesn't fall from the sky you're never getting what you ask for when you ask for a glass of water. It seems to me intuitions are not clear about that it seems to me. My intuitions are not clear. Does that make sense? Okay. I think what your example presses at is the need to say more about what dominant means. The reason I give the Napoleon example is it's so clear that one you can see why one person rather than another is counting as a dominant source of the information. So there is something very intuitive here but making this into an explicit general account is not easy. Okay. So you can have just the same structure for names for substances as you have for names of people here. You could have unintended shifts of reference if you have dominant source shifts without you realizing it. So we don't need to think of natural kind terms as Kripke thinks of them as involving an initial dubbing and then a causal chain plus an intention to preserve reference. That's the Kripke theory, that two-part theory the initial dubbing and the causal chain with the intention to preserve reference. And what these kind of examples seem to show is that that two-part structure is not right. Because you can't have the shift of reference even though you have both these conditions being met. And Evan says the point of his paper is to restore the condition connection which must exist between strict truth conditions and the beliefs and interests of the use of the sentences and the technical notion of strict truth conditions is to be of interest to us. So this is supposed to be more of a real-world conception of reference. If it's the dominant source of your current body of information then it's something that matters to you. Whereas something that's merely at the end of a causal chain going back to an initial dubbing perhaps several centuries ago that's not really so important what went on in that chain of that initial dubbing several centuries ago. What matters is who's the dominant source of your present collection of information. Okay? That's just to state what the theory is. Okay, very good. So I want to work over just I think this is the most technical bit in Evans's paper and it seems to me needlessly technical I don't think it's actually... I don't even think it's correct, actually. But let me just work over this because I assume that you're still trying to do this thing of working back and forth between the readings and the section in the lecture. Yes? We are also doing the readings. Yes? Right. And there's certainly... Just smile at me. There's much more in Evans's article than I'm covering in the lecture as you'll know, right? It's a very rich article. Lots of examples. What I'm going to do now is just try and talk through what I think is the hardest to crack bit of the article, the hardest to see why he puts things the way he does. So he says a speaker intends to refer to the item that is a dominant source or is associated body of information. That's a little bit strange that I think he means that as a kind of definition of intending to refer. It's a little bit intuitive. The kind of way you'd use it is if I'm telling you a great long story about Louis XIV I'm saying to you how Louis XIV was very badly treated by the British and was eventually put in prison on Elba by them but obtained a glorious release and went on to fight at Waterloo. After a bit you might say to me you don't mean Louis XIV you mean Napoleon. You were intending to refer to Napoleon. So that because Napoleon is clearly the dominant source of the information that I'm using. So that's the kind of thing he means you mean to refer to you're trying to refer to that's just a definition the item that's the dominant source of your associated body of information. That is a little success in reference in any particular case is relying on common knowledge between speaker and hero that the name has been used to refer to X by the members of your community and this is the kind of anti-serial clause that's not uncommon knowledge that the referent meets matches predicates in the cluster. I know it's early in the morning but is it kind of plain that that's the anti-serial point? Serl said the referent is whatever matches the cluster of descriptions associated with the term this is not a match theory it's not satisfaction of the predicates in the cluster that's the important thing it's not whether he actually did escape from Elba it's whether he's a source of the information about escaping from Elba so what goes on is we have this practice of using a name Napoleon we all know that we're using it to refer to this person everybody knows that we're all using it to refer to this person and when I say Napoleon I'm relying on your knowledge that the name has been used to refer to that person before something like that is the theory it's a little bit complex because really most of us wouldn't really have thought about this stuff before starting a class on it yeah I mean did you really know that people intend to refer to the dominant source of that information before you did this class? I mean think about these poor, benighted people who've never taken a philosophy class but talk about Napoleon the whole time do they really know you know what I mean historians do they really know all this stuff? does anybody who's using a name really have this reflective knowledge of what's going on? Evans sums up his theory by saying Napoleon is a name of X a community C in which it's common knowledge that members of C have on their repertoire the procedure of using Napoleon to refer to X that is with the intention of referring to X that is with having X as the dominant source of the body of information associated with the term now is that common knowledge among English speakers? I do not believe it I mean even if Evans's theory is perfectly right this is really a little bit fancy yeah I know that we prefer our refutations of philosophical theories to be this is a contradiction but the objection to this and it's just a bit fancy it describes to people knowledge that they don't really seem to have so let me try and put this in perspective I mean do we really have to be explicit about this? I mean our ordinary thinking suppose you think about using a term like this piece of paper to refer to a particular piece of paper or that table or that person then what's going on there is you could say well this isn't using a name if I just say that man over there but it's the same kind of structure because in perception I've got a whole bunch of information about a particular individual and the person I'm referring to is the person who's the cause of me having all that perceptual information remember that thing with the tomatoes I got a tomato behind a mirror I got a tomato I can see in the mirror and I say that tomato which one am I talking about? here this is me I can see in the mirror this one is hidden by the mirror I use the term that tomato I say that tomato is red or that tomato is mine which tomato am I talking about? the one in the mirror I mean this one yes the one reflected in the mirror not the one hidden by the mirror so I can refer to the thing when there's that causal connect between me and it it's the dominant source of my perceptual information so Evan's theory applies here too but that doesn't require that I be thinking about that being the dominant source of my perceptual information it just is the dominant source of my perceptual information therefore that's the one I'm talking about I have here a simple piece of chalk now you can think about this chalk you can say what is he going to do with it does that require that you be thinking this chalk is the dominant source of my current body of perceptual information no you didn't need to think that but so long as the chalk actually is the dominant source of your current body of perceptual information you can think about it so you perceive an object you can think well I'm getting a cluster of information I'm getting a dossier on that thing and the dominant source that just is the object you're seeing so you've got a perceptual system here I mean just as just having vision you've got a perceptual system that organises itself around particular objects in your vicinity that's what your perceptual system is going at any animal is going to have a similar perceptual system not any animal but most animals that organises itself around the objects around them a tiger seeing its prey has got a whole bunch of information about that thing it is not reflecting on what's the dominant source of that information but it still sees particular objects so when you get that that flamingo or whatever then you are referring to whatever is in fact the dominant source but there's no obvious role here for explicit common knowledge or intention if you talk about photographs what a photograph is a photograph of you could say a photograph is a photograph of whatever is the dominant source of the information in the photograph but this stuff about common knowledge or intention doesn't really seem to the point even if the photograph was taken by a machine it's still a photograph of a particular person whatever anybody intended and it seems similar for memory so if you just think what kind of cognitive creatures we are we don't just use language we have perception, we have memory and if you ask who is it you are remembering if you remember a particular glass of water then that flash bulb in your memory which thing you are remembering is the thing that is the dominant source of your current body of information so you could think what's going on is in perception in perception or if you think of this over time in perception I get information of that input to me I'm referring to whatever is the dominant source of that information over time that perception, that causal chain goes on, I get wired up with other perceptual inputs I have and so on and you've got here a system for finding out about its environment and when you say well which object is it finding about in its environment it's just finding out about the objects that are the dominant sources of its current information and so on the way we talk to each other is just the same extension of those more basic perceptual systems that's to say they work in just the same way if you and I are talking about someone then we're getting information about whatever thing is the dominant source of our information that's just the way it works and you could think well here's the way it goes you get the person that we're talking about here inputting into my perceptual information then I talk about them to you you put that together with your perceptual information the whole information gathering system that you have whether it's perception or memory or language is all working in just the same way and it's not a matter of what we have common knowledge of or what we think about we don't need this kind of clause in Evans where we talk about the common common knowledge that we have in our repertoire we can just say language works in the same way as perception or memory I'll try and amplify that kind of point in a moment is that reasonably clear I'm really saying it's not so much that I'm trying to explain that relatively technical formulation in Evans as to say you can set it aside I don't really think this is was valuable about the theory yes I'm saying that's how it is I actually do not think Evans would have disagreed I think he didn't really address the question but I think if he had addressed it he would very likely have agreed so I don't think this is it's a wildly left field idea I'm expressing here but what I'm saying is if you think of an animal that can perceive and remember but it doesn't have language the way you describe what it's doing is finding out about particular objects and having memories of particular objects and that would work in that dominant source kind of way and language is just working in the same way human language is just working in the same way as the cognitive system of an animal without language so far as reference goes ok well on that cliffhanging point let's look a little bit about this in a much broader context so I'm actually not going away from this point if you're puzzled by that then you will have an opportunity to raise questions in just the next couple of minutes suppose you ask what the function of language is and what good does it do you having a language we talked about this a bit in the very first meeting but I want to come back to this with one particular answer in mind there are lots of answers you could give to the question what is the function of language but surely one basic thing is that it transmits knowledge from speaker to hearer that's to say if I can see round the corner and you can't and you say to me is he coming then I can tell you he's coming just for you just as if you'd seen him do you see what I mean should I go over that again and one way you can find out whether he's coming is to have a look yourself another way is to ask me so my talk does the work that perception could have done if someone reliable tells you something then it is for you just as if you'd seen it yourself that's kind of obvious with animals that an animal call sign there's a predator around can make the whole flock react just as if any one of them would have if they'd seen the predator themselves yeah that's why animals have these alarm cries so there's something really basic about that so the basic picture is the speaker makes an assertion the speaker says something the hearer trusts and understands them and then you get knowledge yes that's a really basic thing in language and obviously if you're very sophisticated you're not going to believe a lot of the things that people say to you but there's a really ground floor level in which you have to believe the things that people say to you and we do treat language as a way of conveying knowledge in really basic cases you know the lottery paradox these come across the lottery paradox put up your hand if you know about the lottery paradox one okay the lottery paradox is this suppose I buy a ticket in a lottery then suppose there are another hundred there have been a hundred tickets sold what are the odds of my winning the lottery one in a hundred right not a very good chance it's 1991 that I will not win do I know that I'm not going to win no I don't know I'm not going to win suppose there are a million tickets in the lottery yep so I have only a one in a million chance of winning do I know I'm not going to win no so no matter how high the probability is that I'm not going to win I still don't know that I'm not going to win so knowledge can't be identified with merely very probable belief I can make it a million million tickets I can make it a Google of tickets and I still don't know I'm not going to win yep okay so let me put it let me twist this a little bit suppose that you've bought a ticket in the lottery can I do I as a candid friend have the right to say to you you're not going to win no even though it's very very probable yes I don't know that right so if I'm going to assert something it's not just that what I say should be very probable it's that I have to know that it's right if I'm going to tell you it's one thing if I tell you you're not going to win because I know the thing has been fixed and you're definitely not going to win because someone else has already won or it's all been rigged but if I only have a high probability I don't know the right to say to you you're not going to win okay so when we're using language we're using it for the purpose of transmitting knowledge in these kind of simple basic cases and then you can think of language as a kind of long range perception it's an extension of your perceptual system if someone's telling you that P it is for you as though you'd seen for yourself that P people do sometimes say well you have to be very suspicious of what people tell you when I was nine years old someone gave me a a notebook in which they'd written son never believe anything you hear and this was given to me as a piece of golden information that would guide me well through life just don't believe anything that anyone tells you and it's really not good advice you can see what they mean right you should be suspicious you should be wary but the idea of not believing anything that anyone tells you I mean if you really went that far you couldn't so much as learn the language in the first place suppose that you're two years old and you've got this advice never believe anything that anyone tells you and they're saying and the Trump has tried to explain to you what does red mean and they're saying this is red and this is red and this is red that's not red that's not red and they're saying I don't believe any of that then you have no hope you have no way of ever learning the language and as for just ordinary knowledge of common affairs if you take some very exotic place like Scotland or Australia and you ask well how do most of us know there is such a place even you don't know because you actually did some visual inspection yourself yeah you rely on what people told you I mean you couldn't have a fraction of the knowledge you have unless you relied on what people told you so being able to trust and get knowledge from other people seems to be fundamental to the functioning of language we treat it that way in everyday life anyway in these Los Aliparados kind of cases but when you think about how language learning has to work and further to be a stable language at all you just got to be taking it that language can give you knowledge but then suppose you think of how you'd explain what an assertion is made using language well you can think of a perception as here I have knowledge in my head there is the world out there perception is just something that gives me knowledge of the world that's not a bad definition of perception if you're looking at an animal if you're looking at some creature that lives deep in the ocean deep in an ocean trench and you ask well can it perceive its surroundings what you're asking there is does it have something, some organ or something that will be an intermediary between what's going on in its environment and it getting knowledge of its environment perception is just something that comes in that if in between you and the world gives you knowledge of it so you could think of assertion like that you could think of assertion as being like perception there's what's going on in the world out there there's your knowledge and assertion is something that in language can mediate between the stuff that's going on in the world and you're getting knowledge of it so if you've got if you know about something and you tell me and then now I know about it then your assertion is what allowed me to get the knowledge that you had because you can think of assertion what you do is make a statement in a language that's really defined in terms of knowledge the possibility of it generating knowledge in the hero so what does cause have to do with this if we talk about reference then reference is really basic to language functioning the way it does but what does cause have to do with our objectives in using a language that's what's puzzling I think about the Kripke stuff that Kripke makes it very plausible, makes it compelling causation has got something to do with reference reference is very fundamental to the functioning of language but then you think it seems kind of arbitrary why is it cause there is something to do with reference why is it that they're connected but what is reference anyway reference has got to have something to do with the way language plays its functions if you're going to explain what reference is as x then whatever x is it's got some role to play in allowing language to function so what's the function of language class what's the function of language the transmission of knowledge from speaker to hero right very good okay for present purposes that's the function of language so if you want that to happen assertion making a statement stating a fact has to play some role in the transmission of knowledge and you can understand reference by what it's doing in those statements of facts about the world and what's knowledge what's knowledge class okay here's what's knowledge everybody agrees that in order well actually many people agree that in order to know something it has to be true you can't know if it's not true you've got to believe that it's true and it mustn't be an accident that you're right in believing it so you mustn't be just lucky yes that's epistemology right how many of you guys have done a class on epistemology some okay well for the rest of you you don't need it I mean this is this is really the main point okay so it's not an accident that you're getting it right that's what knowledge is yeah you can't be just lucky if I just guess what your name is and I get it right I don't know what your name is yeah just a lucky hit isn't so what's that isn't an accident I mean that's what epistemology is it's a study of what it means it's not an accident yes I'm not sure the GSI completely agree with that remark but anyway okay so I make the following claim it's not an accident it's a causal notion if I just guess your name and I get it right then that's not knowledge because the fact that that was your name wasn't what caused me to have the belief so what it means it's not an accident is that you know that P and the fact that P is true is causing you to believe that P yeah knowledge is a causal notion so if you think of perception as what gives you knowledge of the world then in order to be perceiving this tomato rather than that tomato I have to be causally connected to the tomato because it's a causal connection that means I can have knowledge that's how I can be having beliefs about the tomato which aren't accidentally right they're caused by the way the tomato is that's why I can have be perceiving this tomato rather than that one so perception of the object requires a causal connection to the object because perception is the intermediary between the way things are out there and you're getting knowledge knowledge requires causation so perception requires causation so if you're going to make a statement about an object then that's got to be something that's capable of providing you with knowledge of the object that's what we're saying the function is to generate knowledge an assertion's got to be something that can generate knowledge so an assertion about an object demands a causal connection to the object otherwise it couldn't be capable of generating knowledge of the object so to be able to refer to all if I'm going to be able to refer to one tomato rather than the other I've got to be in a position to make assertions about the object that transmit knowledge of it yes, that's what we've got to so being able to refer to an object is being in a position to transmit knowledge to the knowledge of it but that requires a causal connection to the object the reference to the object requires a causal connection with it well there you are now you know why cause is the right notion why it's not just arbitrary what Kripke found was the connection between cause and reference because reference is connected to a language to transmit knowledge throughout the community and that requires causal connections so reference to the object requires a causal connection with the object in virtue of which you're in a position to transmit knowledge of it so not just any causal connection will do and if you think of language like that you see that it really is very similar to all our other cognitive capacities is very similar to the perception works or memory works we don't ever have any real knowledge well because we are brains in a vat or something like that well on the face of it if we really were brains in a vat and didn't have knowledge of anything then what would we be able to refer to if you say we can refer to our own sense perceptions and so on do we have knowledge of them see the reason you're restricting it like that is just that you're thinking well I do have knowledge of that lot and you're just taking it for granted you can't refer to any of the stuff out there if you don't have knowledge of it and that's true even if suppose that suppose that in fact you are a brilliant Berkeley undergraduate sitting in a class just like this one listening to a spell binding lecture just like this one but you doze off and you have a dream in which you're a brilliant Berkeley student listening to a spell binding lecture just like this one you see what I mean so even though the world is just the way you're dreaming it to be you're not actually getting knowledge of any of it because it's just an accident that your dream matches the way things are you could have been having a quite different dream that makes sense so in that case suppose someone comes in the door and in your dream someone comes in the door but it's only an accident are you in your dream able to refer to the person who really came in the door no of course not it's like the tomato case there's something it seems to you there's a tomato right here but you're not able to refer to it because it's not causally connected to you your dream matches what's on going on out there but that doesn't allow you to refer to any of the stuff that's going on out there so just having the right kind of background of sensory impressions wouldn't be enough to let you refer to that stuff out there so you don't have the knowledge you don't have the ability to refer yes, language would just stop working completely the only possible exception that's the claim the only possible exception to that is your knowledge of your own mind that's a really interesting question and actually the next Putnam article we'll look at is called the brains in a vat and it's actually addressing the question what a causal theorist should say about these possibilities of where it seems like you don't know anything about the world Putnam's answer is quite radical and surprising as so often with Putnam so you might want to take a look at that if you're already if you're interested in the question plain as day so now we know what the function of language is and why reference is a causal notion more in evidence next time . . . . . . . . . . . . .