 I want to begin by just going over where we'd got to last time before looking at questions about whether a folder's approach can possibly be right. So the general approach here is what we're trying to do is explain how it is that we have reference, that we have truth and falsity in the language at all. And we're trying to do that by talking about the causal connections that things stand in to the signs of the language. That's a general project, so it's explaining reference in terms of causation. And then the disjunction problem is if you have x's and y's both causing the sign x to be produced, but x's are what the sign refers to and y's are mere look-alikes, then with what right can we say that the causal connection between x's and the sign x is the one that fixes reference and the causal connection between y's and the sign x is just something that's producing illusions. y is one rather than the other, the causal connection that's fixing reference, thus the disjunction problem. So we have these two causal pathways. Horses cause, I mean let's just stick with Fodor's toy example. As came up last time in the discussion, this is really a simplification. Lots of language is much more complexly interwoven and you don't always get nice discrete packages like this. But let's just work with this for the moment. Suppose we have this nice discrete package. Horses cause production of the word horse. Cows cause production of the word horse. But only one of these is the reference fixing relation. Why is that? Why does one rather than the other what makes that relation hold? Well, the answer is this causal connection, the causal connection here, if this causal connection between horses and cows went away, if that had not existed, would there be that causal connection between cows and the use of the word horse? Oh, that's the right answer. Look, there you go. So if this relation went away, this one would go away. But now suppose class that the causal connection between cows and the word horse went away. Would the causal connection between horses and the word horse go away? Oh, look, it just stays there, right? So that's the difference between the two causal connections. So the one on the left does not depend on the one on the right. If the one on the right went away, the one on the left would still be there. But the one on the right does depend on the one on the left. If the one on the left went away, the one on the right would go away too. Yes? That's asymmetric dependence. So the causal connection between cows and the word horse is asymmetrically dependent on the causal connection between horses and the word horse. So that terminology at this point you're very comfortable with that. Dependent, this one's, the one on the right is dependent on this one in the sense that if that one went away, that one would go away too. It's asymmetric in that it doesn't go around the other way. Yep, okay. So the analysis is if you've got that asymmetric dependence, the dependent causation is causation of mistakes. The independent causation is what's fixing the reference. And just let's do that exercise. If D stood for being a horse or a cow, then you'd have a situation in which horses cause D and cows cause D. But these two causal connections would be dependent on one another. If one went away, they'd both go away. And that works either way around. Yep, yep. That's right, yeah. That's exactly it. That's a good example, yeah. That's right. Well, there are two things here. One is the contrast between a genuine umbrella term like mammal that isn't defined as either a horse or a cow or something. And one that really is defined as either or. That's a good distinction. But I think D here could be a term of either sort. If cows just went away, then how should I say, there wouldn't be any cows around to cause the use of the term. That's right. But it would still be true that the causal connection between cows were there any and the use of the term would be in place. You wouldn't have disconnected the connection. You'd just have taken away the stuff in one end of it, if I can put it that way. Okay, so if you've got that symmetric dependence, then both of the causal connections matter for the meaning of the term. Yeah, that's the idea. So if it's symmetric dependence, then both of the causal connections matter for the meaning of the term. But if you've got ambiguity, so suppose that a photo pointed to a horse, a cow and said horse, and said, of course, I was using the word horse there in the sense of cow, then you could say, well, the word horse in that understanding over is systematically ambiguous. There's two just quite different meanings. In one sense, it means horse. In another sense, it means cow. Then you've got two causal connections from horses to the use of the word horse and from cows to the use of the word horse. But they don't matter for each other. If you took one away, the other one would still be there. And that works both ways around. So if you've got an independence that's symmetric, then you've just got an ambiguous term, right? So if you've got dependence, then that tells you that there's only one sense of the word here. It's non-ambiguous. And if it's asymmetric, that tells you that one of these connections is the one that's fixing the reference and the other one isn't. If that's all plain as day, then you've really got it. Got the theory, okay. So asymmetrically dependent causation is causation of mistakes. So just one last thing here. I just did that by making the arrows go away, but just to be fully explicit, what we're saying is if there's a counterfactual, this is about what's going on in other possible worlds. If horses didn't cause horse to be produced, then cows would not cause horse to be produced. Even if cows didn't cause horse to be produced, horses would still cause horse to be produced. Is these points about what's going on in other possible worlds that are telling you which cause or connection fixes reference and which one doesn't? Okay. Are you absolutely at ease with that? Yep. Oh, sorry. Excuse me. Okay. The thing is here we have a really serious attempt to give an explanation of how it comes about that there are standards of right and wrong for language. It's saying there are standards of right and wrong for sentences in virtue of and then what's coming after that is something from the natural sciences, something that doesn't presuppose, talk about meaning at all. Yeah. And when we're talking about these causal connections and this asymmetric dependence, the attempt is to get something that's very austere from which talk about meaning has been eliminated. So you can understand this base independently of knowing anything about meaning. A Martian that didn't know anything about human languages should be able to come to Earth and spot that asymmetric dependence and the causal connections and say, aha, so they use the term horse to refer to horses but sometimes they get taken in by cows. If you didn't know anything about meaning, you should be able to derive the facts about meaning from these more basic points about asymmetric dependencies. The trouble is, and this, it took me a little while to see this but many of the questions last time are actually homing in. I think, I mean, that's my picture of it, but many of the questions last time are actually homing in in the comment I want to make now. Here's how Fodor explains the asymmetric dependence and it's actually very like the way I was explaining it last time. Misidentifying a cow as a horse wouldn't have led me to say horse except that there was independently a semantic relation between horse tokens and horses. So the key point is that meaning relation, that reference relation, the semantic relation is there but for the fact that the word horse expresses the property of being a horse but for the fact that there's that semantic fact about horse it would not have been that word that taking a cow to be a horse would have caused me to utter. So it's very clear when he's explaining this that this actually appeals to facts about semantics. When you're explaining what the asymmetric dependence is you actually are talking about something that has to do with semantic relations. His point is the semantic relation between horse and horses holds independently of the cow's horse relation. Well that's okay but I mean that's true but that's the fact we were trying to explain. We're trying to characterize what that means that there's a semantic relation. We're trying to explain in naturalistic terms how there can be such a thing. The thing is it seems like I think that's a real giveaway this paragraph because it seems like what's going on is and as I say if you recall some of the discussion last time that people were really pressing on this point what's happening here is that given the meaning of the word horse since the word horse has the meaning it does it's a consequence of that that if horses didn't cause the word horse to be produced cows would not cause the word horse to be produced because cows are coming in only as leukolites exploiting our understanding of horse. So it's the facts about meaning that explain the existence of these causal relations. In fact you can always say what the causal connection is here by appealing to these points about meaning. If we're really going to be getting a deep explanation here it should be that the facts about causation are explaining the existence of the facts about meaning. But this is round the other way the facts about meaning explain the existence of asymmetric dependence and given the meaning of the word horse if cows didn't cause horse to be produced horses would still cause horse to be produced that's to say the way you now understand the word horse you could keep that intact even if you weren't being fooled even if you'd improved your discrimination capacities so what that means is again that it's the facts about meaning that explain the existence of the asymmetric dependence so the asymmetric dependencies maybe they exist at any rate in these simplified models of language but they don't themselves explain anything about meaning they are consequences of the facts about meaning. Naturalism was trying to explain how there can be standards of right and wrong for sentences without appealing to notions like meaning. We were trying to explain how it comes about that in the world described by physics there is such a thing as going right or wrong on your use of sentences but all that's happening here is that we're saying given that you're in a world where there are facts about how you go right or wrong on your use of sentences as a consequence of that there will be these causal relations. The idea is if the word horse means horse and you know what's going on that is you know that the word horse means horse then these things the horses are going to cause production of your term horse the causal connection here is a result of the fact about meaning and anything you can't differentiate from a horse will also cause production of your term horse this is all that's happening here you take the facts about meaning as basic then you remark given that someone knows these facts about meaning given that someone understands the word then there's going to be that causal connection between horses and their use of the word and they're going to be taken in by lookalikes but they can improve the discrimination against the lookalikes so you get the asymmetric dependence but that doesn't explain anything about how there comes to be such a thing as meaning in the first place a slightly different way to put this point is to say we are after all looking at counterfactuals here if cows didn't cause horse if horses didn't cause horse it's obviously not naturalistically acceptable to say and we held constant the meaning of the word horse but with these questions about zebras and so on last time what people were saying is well which other world are we talking about when we say if cows didn't cause horse or if horses didn't cause horse are we talking about for example a world in which zebras didn't cause the meaning of the word horse didn't cause the use of the word horse and I said well what we've got here is in the nearest worlds in which horses don't cause the use of the word horse cows don't cause the use of the word horse and really saying nearest is not much of a steer as to which world we're going to be looking at what we really want to be able to say is these worlds we want those worlds in which horses don't cause horse and those worlds cows don't cause horse what we want is these are worlds in which the word horse has lost its meaning we also say in the nearest worlds in which cows don't cause horse horses still cause horse we want it to be that these are worlds in which the word horse has still kept its meaning it's very hard to know how we could possibly specify exactly which nearby worlds we're talking about unless we're using the notion of meaning we can't even say which counterfactuals matter here and I kept saying when I was trying to make it intuitive what the asymmetric dependence is well if horses didn't cause the use of the word horse anymore that could only be because the word horse had lost its meaning and then that drives your senses to which worlds you're talking about but you're having to use the notion of meaning to explain even what asymmetric dependence is and you couldn't use the notion of meaning here it's not even obvious why you diagram the situation this way diagramming the situation this way starts out with two causal pathways and says well one of them matters for reference but if you really said let's forget about meaning for the moment let's just look at the hard facts let's just look at the basis for our talk about meaning what would you have you just have one causal pathway by which both horses and cows cause the use of the word horse so the whole thing is round the wrong way there's a causal asymmetry alright but it can't explain the facts about meaning rather the facts about meaning explain the existence of the causal asymmetry so this way of explaining how there can be meaning in a natural world is simply hopeless how there can be meaning I think that's fair enough if you read it that way if you say I'm going to take for granted that there is such a thing as meaning, truth, reference and so on and I just want to point out here's a kind of diagnostic you could use that would be helpful in thinking about suppose you're facing a foreign language to translate and you're wondering are these are these remarks by the natives are they, by the indigenous population are they expressions of mistakes are they getting it right then you could appeal to causal asymmetry as a kind of diagnostic taking for granted the facts about meaning but that's really not the project here was supposed to be a lot more ambitious than that you have to take my word for it but if you look at the photo text you'll see that right from the start GSI is this correct? I'd allow you to keep me honest it seems to me right from the word go the problem he's trying to address there is we live in a world that is entirely a physical world how did this happen? that we got language, that we got representation think about Dretschke in Dretschke it's the same project and given that all is going on is a whole bunch of biology where did the standards of right and wrong come from? the idea is to get right down to brass tacks and explain the existence of right and wrong but you're right that's one reason this is an important idea it's important partly because it's such how should I say a bald headed attempt to go right at the problem and you forgive him all the simplifications if you can get the simple case right here it would be so great to get the simple case right but even if it's not right it's likely to be a valuable idea anyhow a good piece to have in thinking about these problems just the kind of way you're suggesting but that's a lot more limited than the exciting original idea okay? it seems to me though in fact that this problem actually generalizes it's partly because Fodor's account is so gung ho you know he really is going for it he really is trying to do the thing here of explaining meaning in terms that don't presuppose meaning and when you see how difficult it is it actually starts to look like the problem here is quite general I mean suppose Fodor solved the problem suppose you could say in entirely naturalistic terms what this asymmetric dependence is suppose you could do it without talking about meaning suppose you would just give him about the objection I just raised there's no good right? so then you could say how cows cause horse depends on how horses cause you have then a kind of imitation you have a kind of lookalike for I mean as the last commenter said you've got this in fact when you have meaning as a consequence of that you have this asymmetric dependence so when you know you're dealing with facts about meaning you can look for this asymmetric dependence to clue you in as to just how the facts about meaning lie you know which utterances express mistakes and which are correct so you'd have a kind of facsimile of that if you stated causally symmetry in terms that don't require meaning you'd have described something that looks quite like the situation you do find in languages that do have meaning I suppose for example it turned out that there's a high temperature pathway I suppose the visual system has got a low temperature pathway and a high temperature pathway there are no vision scientists here take my word for it one of them is very very hot and one of them is very very cold you don't really notice it because of the fantastic insulation in the brain so suppose what goes on is that there's a high temperature pathway by which horses cause the use and the low temperature pathway cows cause the use of the word horse I have to say you have to be suspending disbelief a bit here because partly the point of the example is that ringers for anything are actually using the same causal pathway I mean the whole thing about being able to be taken in by a ringer is that once you set up a vision so it can detect horses you have thereby set it up so it can be taken in by a ringer do you see what I mean because vision is not infallible so this can't be quite right you're not going to get too distinct pathways in that easy way but let's suppose that that did happen suppose that the illusions all came in by the low temperature pathway and the real thing all came in by the high temperature pathway and suppose it turned out that for reasons that are very technical I don't really have the jargon to explain it to you properly but suppose it turns out that when you look at the high temperature pathway in the brain and the low temperature pathway in the brain it turns out that they're asymmetrically dependent let's suppose that the high temperature pathway is really what keeps everything running and the low temperature pathway derives all its energy it taps off the energy it needs to keep going from the high temperature pathway so here we got just to not to overwhelm you with technical detail but here we've got the high temperature pathway here we've got the low temperature pathway the low temperature pathway is getting its energy from the high temperature pathway so if the high temperature pathway got knocked out? Would the low-temperature pathway keep running? No. But if the low-temperature pathway got knocked out, would the high-temperature pathway keep running? Yes. So do we have a dependence here of one pathway on the other? Yes? Is it an asymmetric dependence? Yes. Okay, look at that. So now we've got, that's purely naturalistic, right? I didn't talk about meaning at all there. I just talked about temperatures and one drawing energy off from the other. So then you've got a kind of facsimile, a kind of imitation of that asymmetric dependence, this kind of diagnostic of meaning and illusion in ordinary talk about thought and meaning. So you've got these two pathways and the cow-horse pathway, the low-temperature pathway, is asymmetrically dependent on the horse-horse pathway. But what's happening here is that these simple engineering facts, I mean, well, actually I shouldn't, they're very complex engineering facts, but they are, as you don't mind me putting it like this, they are just engineering facts. They are mere engineering facts. How could that be the same thing as the existence of standards of right or wrong for your use of language? Just having that kind of structure and visual pathways in the brain wouldn't of itself mean that you had right and wrong here. So there's this kind of asymmetric dependence that does exist that is, as someone said, diagnostic of what's going on when we have meaning of the distinction between illusions and mistakes, but between illusions and getting it right. But merely finding a facsimile of that because of some engineering fact that couldn't explain the existence of standards of rightness or wrongness. And it would be perfectly consistent with having a low temperature pathway and a high temperature pathway linked like that, that they're both getting it right or that they're both getting it wrong. Asymmetric dependence of itself can't explain the existence of standards of right or wrong. So if the facsimile is described in entirely naturalistic terms, then it can't explain the existence of standards of rightness or wrongness. You only look like you're getting standards of rightness or wrongness here because the intuitive understanding of what's going on with this causal asymmetry presupposes talk about meaning. But when you really do the thing of abstracting from that and saying let's suppose we've got just some basic engineering fact here, then the thing seems obviously completely hopeless. You couldn't explain right and wrong in those terms. But then once you get that point, it seems like that kind of point is going to apply to any attempt to explain the existence of standards of rightness and wrongness in purely physical terms. I mean how could, how could, how could you possibly, I mean, if this engineering thing doesn't work, right? If you say, well, I mean, can you put your hand up your hand? I don't, I hope this is, it seems perfectly obvious to me, but what seems perfectly obvious to me is not always perfectly obvious. If that sometimes is wrong, but can you just put your hand up if you agree with me that this high temperature, low temperature thing, that doesn't explain the existence of standards of right or wrong. Yeah? Is it? Yeah. Well, then what more could do it? I mean, suppose you said, ah, it's not temperature, that's the important thing. It's electrical spiking activity. It's synchronization of neuron firing. Well, what do you want? What do you want? I mean, what could possibly hit the spot there? But if nothing there could possibly hit the spot, this kind of engineering thing isn't explaining the existence of standards of rightness or wrongness. We're not able, in these terms, to explain what we're trying to explain. Now that's, I mean, one of the things that is so good about Fodor's discussion is that unless you see just how stark the problem is, and he gets close enough to giving you an intuitive explanation of the phenomenon of right and wrong, that you can see suddenly how very hard it's going to be to explain why there is such a phenomenon as reference, how it can be that there's such a phenomenon as reference or such a phenomenon as truth or falsity in a physical world. And in fact, once you grasp this point about the high temperature and low temperature pathways, it seems like the thing is actually completely impossible. And the problem is quite general. Naturalism just can't happen. You can't give a naturalistic explanation of the existence of reference or signs of the rightness or wrong. You can't explain how there's meaning in a natural world. So I'm sorry to have taken you this far. It's been fun, right? Okay. I'm actually ending a little bit significantly earlier than I intended to, but perhaps this is a melancholy enough note. Well, I guess what we hope now is that Wittgenstein will show the right way to think about standards of right and wrong. What's happening in Wittgenstein is that he is trying to characterise what it means to be going right or wrong in your use of a sign. He seems to want to explain what's going on here in a way that is not naturalistic, is not in terms of causal dependencies between pathways and the brain. So it's a quite different take on the same problem, and that's what we'll do on the next couple of weeks. Okay. Okay. Thanks.