 Okay hello good morning. So today and on Friday we'll look at Fodor's causally symmetry theory and then on Monday we'll sweep triumphantly on to Wittgenstein on following a rule a different part of the forest and today what I really want to do today is just to set out what Fodor is doing here there's a general problem he calls a disjunction problem that you will probably recognize from the Dretschke and then his causally symmetry theory which I'll state then next is his solution to that disjunction problem then we'll do a little bit of a workout a very simple workout on the causally symmetry theory and then I want to look at there's something puzzling about what the general problem is here and I'll try and indicate that at the end and then on Friday we'll move on to a blistering critique of Dretschke and Fodor but today all I really want to do is set out what the idea is and the general problem I've been saying this right since the first lecture is there are standards are right and wrong for sentences we want to be able to say there are standards are right and wrong for sentences in virtue of something or other and then whatever that something or other is is going to be something from the natural sciences and the whole world is a natural world and the whole world is describable by the natural sciences so since there is language since there is thinking in a natural world since language and thought has standards of right and wrong for them we have to be able to explain how that can have happened how common a natural world there are standards are right and wrong and what Dretschke is trying to do with his talk about biological functions and so on is to explain how that can have happened I mean even if you just look at this general form of the problem it's kind of obvious that it's going to be really difficult to get a convincing solution because in terms some of you at any rate will recognize this is a problem about deriving an ought from an is about deriving remarks about what right and wrong consistent how you ought to think or talk from statements about what in fact is the case and it's notoriously difficult to do that but we have to look at the details so the general structure of the answer that we were working towards all the time is something like this the sign X refers to X's in virtue of the fact that the use of the sign X is typically caused by X's yes good old refers to good old in virtue of the fact that the same good old was caused by good old yes water refers to water in virtue of the fact that uses of the same water were typically caused by water in these terms there is junction problem is this suppose that X's cause the use of the sign X to be caused the use of the sign X suppose that Y's cause the use of the sign X well take anything you like that produces an illusion of X-ness take any ringer for X anything that might cause you to falsely say X then that's a then anytime an X is produced wrongly there's some why that was causing it but if it's X is causing X to be produced that makes X refer to X then anything that's wrongly causing the production of the sign X should also be something that the sign X refers to let me go over this the puzzle is there are going to be some things that cause you to produce the sign X rightly namely the X's there are going to be some things that cause you to produce the production the site to produce the sign X wrongly let's call them Y's and why do we say that the sign X refers to X's rather than to the Y's how can we distinguish the causal connections that are fixing the reference of X from the causes of the production of X that don't fix the reference of X and a factor produced by things that X doesn't refer to that's the general structure of the problem and the basic constraint on a solution is we have to explain that in naturalistic terms we're not going to presuppose anything from outside the natural sciences so we're getting a distinction between two types of causation one that matters for the fixing of reference and we've got to understand why that kind of causation should matter for setting standards of rightness and wrongness so we really need an example here that's a bit abstract so here's Fodor giving an example I see a cow which stupidly I misidentify so this is what it's like when you're out for a walk in the country with a philosopher you see a cow the philosopher takes it to be a horse and then says horse it's really different when you're going out and about with Fodor and when you're going out and about with Oscar Wilde but anyway that's that's a kind of dialogue right the point look for he's looking right at the cow and he says horse so Fodor is a city boy so how come your utterance of horse means horse well because of a causal connection between horses and uses of the word horse right that was a lesson of all that stuff about water and twin earth and all that yeah the reason you word horse refers to horses is that horses are the things around here that cause cause the use of that word but then if cows can produce utterances of the word horse to then why don't we say that the word horse actually refers to cows to and the folder good at right there what we just learned from that example was that you can be making a mistake something that's not a horse can be you causing you to use the word horse and since there are some called these causal relations between uses of the word horse and cows why not say the word horse refers to cows to if the causal relations are the same and if causation makes representation well how can the semantic connections not be the same too that's a very simple example but it's obviously quite general because anytime you make a mistake there's going to be something or other that caused you to make that mistake why not say that causal connection means that you're referring to that thing so just put it abstractly you've got two kinds of connection from horses to the what use of the word horse and from cows to the use of the word horse and we say the causal connection on the left means that the word horse is referring to horses the causal connection in the right does not mean that the word horse is referring to cows why is that they're both causal connections what's as special about this one on the left how how come that one generates reference and the one in the right doesn't? That's the disjunction problem. And we go to explain in naturalistic terms, in terms of the natural sciences, why there's that asymmetry, why there's that difference between the two kinds of causal connection. And ultimately, we want to understand, get some explanation of why it is that this kind of causal connection is generating standards of rightness and wrongness. So that's a very simple example, that cow's horses thing, but obviously I think the problem is really general. And until you address it, then a causal theory can't explain the distinction between rightness and wrongness. Are you comfortable with that? That's where the disjunction problem is. And you see, it's really a very general form of the kind of problem when he was saying, how come the magnetosome is representing this magnitude rather than that magnitude? Okay, are we all on board? Okay, so this is the problem. Explain the difference between the causal link on the left and the causal link on the right. Here comes Fodor's answer. If horses didn't cause the word horse to be produced, then cows would not cause the word horse to be produced. I mean, presumably, if horses didn't cause the word horse to be produced, it's really a bit early for this kind of thing. I mean, I really do sympathize with that. But anyway, if horses didn't cause the word horse to be produced, well, why would that be? I mean, presumably that would be the word just meant something altogether different, something like that. Yeah, that could only be what was going on. So then, I mean, if horses meant were named for some bit of an engineering machinery or car components or something, then cows wouldn't cause the word horse to be produced. Yeah, so that's pretty plausible. If horses didn't cause the word horse to be produced, then cows would not cause the word horse to be produced. But even if cows didn't cause the word horse to be produced, suppose that cows didn't cause the word horse to be produced. I mean, presumably, if all the city guys were sent to some agricultural training college and given some kind of animal awareness, elevating your levels of animal awareness so that everybody got much better at distinguishing the two species, then hot cows wouldn't cause the word horse anymore. Everybody just got better at telling the difference between cows and horses. You can imagine that, yes? Now, if that happened, if everybody just got better at spotting that cows are non-horses, then would horses still cause the word horse to be produced? Yes, you just got better, right? So even if cows didn't cause the word horse to be produced, horses would still cause the word horse to be produced. So there's a difference between the two causal relations. Suppose that the one on the left vanished with the one on the right vanished. Yes, that's what we just agreed to. If the one on the left vanished, the one on the right would vanish. Yes? But if the one on the right vanished would the one on the left vanish? Aha, so they're different, yes? That was easy, right? So Fodor puts that by saying the connection between cows and the word horse is asymmetrically dependent on the causal connection between horses and the word horse. That's to say, I mean, it's not very technical. The one on the right depends on the one on the left because if the one on the left went, the one on the right would go. Yes? But the one on the left doesn't depend on the one on the right because if the one on the right went, the one on the left would be just fine. So the one on the cow's horse connection is parasitic on the horse's horse connection. The cow's horse connection is only coming in because cows look kind of like horses. The one on the right is kind of piggybacking on the one on the left. So if this one got thrown off, the one on the left would still be there. But because this one is piggybacking on this one, if you took this one away, then that one would go, yep? That's why we say this one is the one that matters for reference. That's the theory. The causal connection between cows and the word horse is asymmetrically dependent on the causal connection between horses and the word horse. Is it perfectly clear what that means anyway? What are you saying? Why hold an asymmetric causal dependence? Yep? Okay. So if there's an asymmetric dependence between horse's cause X and cow's cause X, well, if one of these is asymmetrically dependent on the other, the dependent causation is the one that's causing the mistakes. The non-dependent causation is the one that generates reference. This is a little bit abstract, but it's very, very simple. If everything's not perfectly clear, then stop me. Yeah? Sure, yeah. If zebra's cause the word horse, then okay, suppose he did, right? I wouldn't be surprised. I mean, you put a zebra in front of a photo and he says horse, right? He's just not very good at that kind of thing. I mean, no disrespect, but yeah. Okay, but what of it? Oh, okay, there are two things here. One is that zebras are causing the use of the word horse and the other is that horses are not causing the use of the word horse. Yeah? Okay, well, suppose we have that scenario. Okay, so suppose we have... So what we've got up here is zebras, right? Well, then, and suppose we have the asymmetric dependence like this, then the word horse would be referring to zebras. That's the theory. I mean, the idea is it wouldn't make sense to say that, and it wouldn't make sense right now to say that all of us have been mistaken the whole time about which things are horses. It's actually the zebras that are horses. Well, it's even hard to state this, but it couldn't be right to say that what's been causing us to use the word horse all the time is actually something to which the word doesn't apply. It was actually the stripey things that we were referring to, even though they never, ever caused anyone to use the word horse. That makes no sense. That's the idea. So when you consider a situation in which the independent causation, the causation that doesn't depend on anything else is between zebras and the word horse, then you're considering a situation in which the word horse refers to those zebras one, two. Zebras and cows look a little bit alike. Okay, sure. Yes, that's right. Oh, I see. I see. Was that your point? Okay, I missed that. Okay, right, right. Right, so if you install IC, IC, okay. Since cows look, I mean, let's suppose, since cows look like horses and cows look like zebras, yeah. If the reason this causal connection went was that you'd put zebras in there, yeah. Then you'd still have that. So you could have this one going, but that one staying. So this is not dependent on that one. I mean, that's absolutely correct. Okay, I didn't get that in the first line. Okay, that's really important. The question is, when you consider, it's all right if I call that a counterfactual, figure hand-off, if that's not a problem, if I call that a counterfactual. Okay, some, but not all. Well, I just mean something that says something about what would have happened. I should be being through all this possible world stuff. What's going on in some other possible world, right? So we're talking about a possible world in which cows don't cause horse to be produced. Yeah, and we're talking about what goes on in that world, yeah. And the question you guys are raising is which world are we talking about? Yeah, and in general, and in general there's always that issue with counterfactuals. I mean, I just think of these standard examples from the literature. If Bill comes, it'll be a great party. Yeah, if Bill comes and the carpet catches fire, it'll be a great party. Yeah, it doesn't, right? It's not always clear exactly which world you're talking about. That's what politicians really mean when they say, well, you know, suppose Iraq had had weapons of mass destruction. What then you say, well, I'm not going to answer that because I don't really know exactly which possible world you're talking about is what they mean, yeah. So when we're talking about cows not causing the word horse to be produced, I don't mean because you'd substituted zebras. It's like I don't mean and the carpet catches fire. Yeah, but then there really is a question how are we going to specify which world exactly we have in mind? Yeah, I mean, the simplest way to get the effect you want would be to suppose, no, the zebras are causing horse to be produced, but suppose it happened that cows started causing horse to be produced in a non-dependent way. Yeah, suppose a word, I mean intuitively, suppose a word just came to mean cows, then cows would still be causing the use of the word horse. Yeah. I know this, everything that's being said here is very simple but I realize it has this air of rigmarole to it so pause me if any of this is not making sense. Okay, there's another question. How did horses cause us to call horses horses? It wasn't the horse? Yeah, I mean, it's not as if there was some kind of formal introduction where we said to the horse, we are humans, what are you? That's what you mean. It's not as if it kind of manipulated us into it. Yeah, that's right. That's another question about exactly what the kind of causation here is. It's not that they formally introduced themselves, it's that we said, well, look at them, we better give them a name. Yeah, what are we going to call them? So we were responding to that presence and we said we better give them some name or other and then we whipped up the word horse. Why it was horse we produced is a further issue. Yeah, why it was that sound rather than any other. But we had given all these things milling around out there that caused us to think, let's get a name for that and then as time goes on, what you do with that term is causally responsive to what they are all doing. Yeah, if you say, they need to be fed three regular meals a day and no baked beans or whatever you say using the word horse. They are the ones that are making you do that. Is that addressing the question? Yeah. You could think of it in terms of the initial baptism or you could think of it in terms of the ongoing dominant source of your uses of the sign. Yeah, so these same issues we talked about with Kripke and Evans, both causation either way. Okay, so so much for the causal asymmetry theory. So the causal connection between cows and horse is asymmetrically dependent on the causal connection between horses and the word horse. And these questions that people were raising, I said you can put that in terms of possible worlds. So the unnatural way to put it is to say, you remember our little blue friends? Yeah, so if this is us in the actual world and in the actual world horses cause the use of the word horse, then what that counterfactual, what that statement if horse didn't cause horse anymore. What that's talking about is the nearest worlds in which horses don't cause horse anymore. Yeah, so you know that some possible worlds are very close and some are very far away. How about that? That's a good question. I mean, maybe, maybe not. I'm not really sure. It depends on exactly what the phonological constraints are on the generation of words. I mean, there are lots of words. That's right, that's right, that's right. So I think this is the right framework to address that question. Whether, I mean, there are lots of words missing in English, right? So you can say, let's see, you can say, pan, pen, pin, pond, pun, right? So they're all words except pond. There's kind of a missing word. Or tin, tan, tin, tin, tun, tun, t-u-n. There's no such word as t-u-n. Why not? So there's kind of missing words. So if horse went, it would probably be just a missing word like that, right? And if we're saying, I mean, the idea that some possible worlds are very close and some are very far away is very intuitive, yes? You can say to me something like, well, that's a very remote possibility, yeah? I say, suppose some madman comes running in. We should barricade the door. Suppose some madman comes running in here with a hatchet and you say, well, the contingency is very remote, right? What you mean as well is a possible world which happens all right, but it's very far away, yeah? So let's not bother barricading the doors. You see what I mean? Yeah, that's how we think. So when we say, what's the nearest world in which horses don't cause horse? Well, I guess that's because we're using a different sign for them or something like that, yeah? Sorry? Or we don't use, yeah, we don't use horse at all, yeah? Oh, yeah. So the nearest world in which that happens, the minimal adjustment you have to make to this world to get that effect is one where it's simply not being caused by anything, yeah? So not zebras, not kangaroos, not anything, yeah? So here we're talking about this world here, whereas the zebra's world is presumably a little bit further out. Yeah, something like that, yeah? That's the idea, anyway. So the nearest world in which horses don't cause horse, that's some subset of the nearby worlds we're talking about, yeah? And in those, cows don't cause horse either. But the nearest worlds in which cows don't cause horse, they're the worlds in which we've all had our agriculture awareness heightened. And so in those worlds, which are a different set of worlds in which cows don't cause horse anymore, in those worlds horses still cause the use of the one horse. Okay, there we go. There's the asymmetric dependence theory. I think it's fair to say that's really photo's point. There's something else? You see what I mean? I think it's a relatively simple theory. I think the exposition is baroque, I mean, but it's a great theory. I mean, it's very simple, it's very general, it's very austere in what it's appealing to. But yeah, the exposition is not so great. Okay, so that clear what the theory is? Okay, I just want to do a little bit of a work out in this. Suppose you've got a sign that really is disjunctive. Suppose you've got a sign that applies to both horses and cows. Right? Let's say D. So when photo points to a cow and says D, that's right. When he points to a horse and says D, that's right too. Yeah? So cows cause D and horses cause D. So if horses didn't produce D anymore, would cows still produce D? No, because D would be meaning something different. Yeah? So there's no reason to think either of them would still be causing D. So horses didn't produce D anymore. That could only be because D didn't mean the same thing. And in that case, presumably cows wouldn't produce D anymore either. Yeah? So in that case, the cow arrow is causally dependent on, is counterfactually dependent on the horses arrow. Yeah? Because if you didn't have the horses one, you wouldn't have the cows one. And it goes round the other way too. If cows didn't produce D anymore, that could only be because D didn't mean the same thing anymore. And so horses wouldn't produce D anymore. So if we've got these two, then this one depends on this one and this one depends on this one. Yeah? You take this one away. You take that one away. What do you think? The two different situations? That's right, the two different situations where you're describing if cows didn't produce D anymore and if horses didn't produce D anymore. But they have the same upshot. Namely, neither of them is going to be doing the causing. This is like a simple test of whether you're at home or let's talk about causal asymmetry, yeah? Asymmetry dependence, yeah? Is this idea worth it? No, it surely, I mean, just as a matter of fact. Surely, I mean, if you decide, if you don't build platypuses, aren't mammals, then don't build platypuses will then stop you from using the word mammal, yeah? But not for the rest. But no, for the rest, yeah. I think that's right, yeah, yeah. And that would just show that the causal connection between don't build platypuses and the use of the word mammal was not a reference fixing one. That was a mistake, yeah. If you don't get this, if you don't get what I'm saying here, you really have to complain because it's a very simple idea, and it just means I'm not explaining it very well if it's not kind of obvious. Some of you look extremely puzzled, and I'm, yeah, exactly, right, right. So the point of the exercise is just to try and bring out what's meant by asymmetric dependence, yeah? And so this is just a kind of simple finger exercise to see what it means asymmetric dependence. That's right. Do you refer to both or to either? Both, I mean, to both of those, it means it's either a horse or a calf, yeah. I mean, the photo might well result to using such a sign, right, if he gets enough ridicule, right? He would say, well, that's a D. It's one or the other, right, yeah? Call it a D, yeah. So then you've got two causal chains, two causal connections, but they're symmetrically underdependent in one another, but they're symmetrically dependent. So one of them does not get singled out as a reference-fixing relation. I don't know why you're looking so puzzled, folks. Okay, it does seem fine. Okay, maybe it's just a habit, okay. So if D stands for being a horse or a cow, then horses cause D, cows cause D, and they're symmetrically dependent on one another. Yeah, go happy with that, comfortable with that. Okay, suppose D is ambiguous. Suppose that there are some uses of D in which it means horse, and some uses of D in which it means cow. Then you've got two causal connections. You've got a causal connection between horses and D, and a causal connection between cows and D. But if the causal connection between horses and D goes, will the causal connection between cows and D go? No, I mean, if you think about it, if you say, well, banks can mean either river banks or financial institutions, yes? And river banks stopped causing us to use the word bank. Would banks still cause us to use the word bank? Yes, of course they would, yeah? I mean, there is two different ways in which the sign is used, and the fact that one of them was lost wouldn't mean that the other one was lost too. Yeah? If there are three different bills, then any one of them might cause you to use the name bill, right? But if you find out that one of them is an imposter and his name is not bill after all, will the other two still cause you to use the word bill? Yes, of course, right? Okay, so in that case, you get the two causal connections. Does either of them depend on the other one? No, they're independent of each other. This one's independent of that one, and that one's independent of that one. So they're symmetrically independent. Yeah, some other questions? Okay. So if you have an asymmetric dependence, if you've got a sign X where horses cause X and cows cause X, but the dependence is not symmetric, well, then if it's not symmetric, you know that the meaning is not disjunctive, and if there is a dependence, you know that it's not ambiguous. So there's only one answer. It must be a term that only stands for one of these things, not for both of them, and it must be the one that is independent and the other one is depending on. Isn't that great? So there you are, causal asymmetry theory. If you get this, you get everything. If you don't get this, you should really put your hand up and say, please, what is going on? Okay, can you put your hand up and you're comfortable with this? Okay, very good. Not quite all, but just about. You can't have it in this situation. That's right. It doesn't mean both because if it did mean both, then there wouldn't be an asymmetric dependence. There would be a symmetric dependence. Yeah. Do you want me to talk through why? Oh, sure. If X meant both horses and cows, then if you took away horses causing cow, horses causing the use of the sign X, that would mean that X didn't mean what it always meant. It would just have changed its meaning, lost its meaning. So cows wouldn't cause the use of the sign either. Yeah. And if cows didn't cause the use of the sign anymore, that would be because the sign was now being used in a different meaning if it was being used at all. So horses wouldn't cause its use either. So you'd have a symmetric dependence. Yeah. So you know it doesn't mean both. What about horse and cow? That's meant to be an example like that. Oh, you mean a word that covers two different kinds of things? There is a juncture thing. Well, take horse. There was someone in the class who knows about horses. There are two different kinds of horses, right? There are big horses and small horses. Well, I mean let's... It's small. No, I don't see that. At the moment we use horse indifferently as a word for small horses and big horses, right? So small horses stopped causing us to use the word horse. Why would that be? That would only be because we'd started using the word in some quite different way. Nothing to do with horses. That could happen. Yeah. Well, when you start pressing this, right? I think you're getting back to the question which possible word exactly are we looking at? Yeah. So all I've said so far is the nearest ones. Yeah. But if you say, well, which one exactly is the nearest one? Is it one where the sign just has a different meaning altogether? Or is it one where the sign narrows in meaning? Yeah. I think that's fair enough. I want to go on, actually, in just a moment to say some more about that problem. Is it picking up on what you had in mind? Yeah. So it's not exactly you can't think of cases. It's that you have to specify very explicitly just which counterfactual situations, which possible worlds you're talking about when you're talking about the dependence here. But the general line here is dependent causation is causation of mistakes. Yes. No. Disjunction is symmetrically dependent. I hope I'm keeping track of things here. But in the case where D is disjunctive, so it means either horse or cow. Sorry? Well, yes. So we're saying if you're going to sign X, and X has this asymmetric dependence condition, what's another question of whatever? No, I'm saying this is how you diagnose. There's an error. The theory says suppose you've got these two causal connections and you find that there's an asymmetric dependence between the two causal connections. Then you can conclude from that, from the fact of the asymmetric dependence, that is not disjunctive. Because if it was disjunctive, it would be a symmetric dependence. And if it was ambiguous, there wouldn't be dependence at all. So it must mean one rather than the other. So the cow's thing being the dependent one must be an error generating arm. This is the whole thing. I think it's a beautiful theory. I mean, there is really something very intuitive about this. I think it's about force, how things work, how they come to fight back. Right, yeah. And it's a young state's reference. Right. So what do you say about this? Well, some of them think it's disjunctive. Is that right? Have I picked it up right? Well, what do they argue about? Some of them think it refers to horses and some of them think it refers to cows. Okay. It's going to have to be a pretty fancy case of this. I mean, the simplest case would be it's just ambiguous, right? Like, I'm not managing to think of an example of it off my head, but there are cases in British English where a word means something different than what it means in American English. Yeah. But it would really, it would not be intelligent to come to blows or have a big row about this. You see what I mean? What you need is a case like, what you need is a term like democracy or something like that where you're talking to some political leader in Iran and they're saying, look, by democracy I mean this. You think that's not a democracy. That's what I mean by a democracy. Yeah, this is what a democracy is. And that can be a real row. Yeah, that can be a real argument. Yeah, actually you could have a row between a Democrat and a Republican about what you mean by democracy. You know, what it takes to be democracy. But in those cases, there's more fixing the meaning than just the causal connection. Yeah, a notion like democracy is very complex and there's a lot going on with that. It's not like folks are going into a field and saying, horse, you know, it's not that you just land in a foreign country and you say, ah, democracy! You see what I mean? These are much more complex terms. So I think the kind of situation you're describing does our eyes have more complex terms, but it doesn't really... It's hard to see how it could really come up with these more basic uses of language. Yeah. That's right. Or it's like British English and American English. You could say that, but I think in the case of democracy, you wouldn't because everybody cares about the term. Everybody wants to seize hold of the term and make it their own. You see what I mean? That doesn't mean there's a lot going on with it. You're not going to just give up the term and say, all right, you convinced me. I hold to my Republican ideals and I give up my belief in democracy. Nobody's going to give up the word like that. You see what I mean? It means too much. Okay. I've been going to proceed to blister and critique, but I don't know. We just have three minutes to go. So what about that as a general point? That's what Fodor's theory is. Dependent causation is causation of mistakes. Asymmetrically dependent causation is causation of mistakes. So far we've looked at the theory stated and we'll go on to critique next time, but any last questions on just what the theory is? Later today? Completely comfortable with that? Yeah. Yeah. I do know what you mean. Right. I was making fun of Fodor with that simple example of horses and cow, but there's a really serious point here that the example is a simplification and there are many uses of language that are really that simple. And it is true, I think, that bank is a very special case because you've got two meanings that just seem so similar, I mean just seem so clearly distinct. I once read on the back of a potato chip packet that the most ambiguous word in the English language is set, S-E-T, set. And so I went and looked it up and there were something like 57 definitions, you know, entries for set. But if I ask myself, are they clearly distinct meanings? You should try this, try this new favorite dictionary. Look up the meanings and ask, are they really 56 distinct meanings? Ambiguity in ordinary language is really a much more complex phenomenon than having clearly one meaning or clearly another. These things, intuitively these things just seem to shade off into one another in a way that's very distressing for the theorist of meaning.