 Welcome all to this new Alaphos seminar. We have today the pleasure to receive Alexander Byrne. Most of us in this group, we know Alexander through his metaphysics writing because of our interest locally, but he also wrote extensively about many other subjects and today he will present a very challenging or maybe provoking argument and after one hour we'll do a short pause. Pilar, one of us, will make the comments and after that there will be a general discussion. So the floor is yours. Thank you, Alexander. Ladies and gentlemen, it is your pleasure for me to meet you here at the Vananen-Vranco. Unfortunately, I have to speak in English. I think you don't have enough patience for my French from now on. Just so I can, I believe you have to. Okay, so I'm going to be talking about empiricism and its relationship to epistemological internalism. So there might have been another title at some point, it's about empiricism and internalism. So what we need to be arguing for is this, empiricism plus epistemological internalism is incompatible with scientific realism. So this is possibly the important claim, a more controversial one. Empiricism without epistemological internalism is just unmotivated. If you're an internalist, you have a reason to be an empiricist. But if you're not an internalist, you're an externalist, you've got no reason to be an empiricist. So the consequence of these two together is that either you should be, you want to be an empiricist, you can be an empiricist, I have an internalist and an anti-realist. That's a coherent package. But the alternative is, if you want to be a realist, you shouldn't be an internalist and you shouldn't be an empiricist. And this is clearly controversial because I think a lot of scientific realists would say, no, I'm a scientific realist but I'm also an empiricist. So that's why this is controversial law for provoking. So what do I mean by empiricism? Really I'm thinking of empiricism as any philosophical view, cosmological view that gives a central role to perception, sensual experience, sense experience, sense state, that stuff, whatever you call it. You could need empiricism more broadly. If you're an empiricist, I don't believe in the synthetic API. I believe that substantive knowledge requires some kind of interaction, maybe causal, with the world. That's not my target. It's the one that focuses on perceptional sense experience in particular. But there's a lot of empiricism of that kind around. So here's Quine for example. As an empiricist, I continue to think of the conceptual scheme of science as a tool ultimately for predicting future experience in the light of past experience. If you're a philosopher of science, you think that observation and evidence are important epistemologically. So the following views turn out to be counters empiricists. Patrick Mahler, the proposition p is amongst asses evidence even though it's ass knows p directly by experience. Or Hashtag, a colleague of Cambridge. Observation is an act of gathering information. Observation is an act of gathering information by human sensation without optional interpretations. As I said, a lot of realist philosophers of science are empiricists. So Anjan Chakravarty, a realist, has a view of observation that's much the same as Hashtag's. I think of as an anti-realist although you might dispute that. Anjan says, there are things that one can, under favorable circumstances, perceive with one's unaided senses. Let us call them observables. All that sort of stuff that you read in standard philosophy of science. It's all empiricists because it links observation, key epistemological concept in philosophy of science, with perception. Clearly, almost obviously, Batsman Frassen, a realist philosopher of science, links observation to perception. Okay, so these are just the things that I've just quoted. So actually, on the following slides, I will be writing observable subscript P for the term observable as these guys use the term observation. Why do I do that? I do that because that's not how scientists use the term observation. Look at scientific papers. Look at the titles of scientific papers. They talk about things like observation of gravitational waves resulting from the collision of two black holes. They think perfectly sensible to talk about the observation of a gravitational wave. Utterly acceptable. You'll see that from not just the gravitational waves, but observations of pie mesons under certain circumstances, dot, dot, dot, so forth. So scientists don't use observation in this way, linking it to perception. But philosophers of science and epistemologists often do, and that's why I'm putting that P there to be clear that that's the sense of observation that we're working with when we are. I think that the scientists' usage is the better one and I think there's a good reason for it, but I'm not depending merely on an ordinary language argument that scientists use it this way, therefore philosophers must as well. I'm not going to assume that. I might include that, but I'm not going to assume it. So here's my plotted history of philosophy, which is really dangerous territory for me. But what I think of the historical debate between empiricism and rationalism as a debate about, basically, about foundations for knowledge. And I take both empiricists and rationalists in the 17th century to be looking, particularly as it's time of the scientific revolution, post-Renaissance, you're looking for foundations for science, for new thought, in certainty. And the difference between empiricism and rationalism is about where one can find this. And the empiricist is to say, well, one finds it in and exclusively in experience, set impressions if you're human. Whereas the rationalist thinks that there's a fact of rational intuition that delivers self-evident propositions. But from my point of view, what's interesting is what they share in common, this search for certain foundations. And I think that this search for certain foundations is motivated by epistemological internals. So let's turn to inter-physicism. Actually, this shouldn't really be an epistemology, this should be just an omnius, but I shouldn't change that. Okay, so what I'm saying here is, this is what I take internalism to be. Look, there's all sorts of debates about what the right word characterizing internalism is and how the different characterizations do relate to one another. But here, this will do for our purposes, I hope. And it's the idea that if some thinker is justified in her belief, then that's only because, or it has to be the case that, only if she can become aware by reflection of all the facts relevant to her being justified. Here's a condition on justification, what makes the subject justified is ax internally accessible to her, accessible by reflection. So it can't be the case that I'm justified in my belief, but I'm not in a position to see that I'm justified or why I'm justified. So contrasts, for example, with a simple externalist view like this, which I'm calling simple reliableism, that says that a subject is justified in her belief when her belief is the result of a reliable process of belief formation. And that's externalist, not internist, because on this view, as long as the process is reliable that links the world and my belief, that's sufficient for justification. I do not have to be in a position to, by reflection, see that this link is in fact real. The causal theories of knowledge are also externalist. So that's the stuff I'm sure is familiar to everyone. But this will be my contrast, my contrast can't see. There are reasons for thinking that simple reliableism is false, but I'm using it as a, as it were, my template externalism. Because I think we argue, look, it's sort of simple as this, right? If you are a reliableist, if you think this is a good account of justification, then you've got no reason to be an empiricist, right? Because nothing in this says, or, you know, there's no good reason thinking that the only reliable belief processes of belief formation are ones that involve sense perception. Indeed, some of the reasons that some people give for rejecting reliableism find that implausible. But from my perspective, the point is, if you believe this, you don't have to be an empiricist. So one of the test cases for reliableism is Mr. True-Temp, this individual who's connected to a thermometer and has accurate beliefs about what the temperature is. The question is, yeah, if reliableism says that this person is justified and can have knowledge of what the temperature is, even if they're unaware of this, why they have that belief, and that is reliable. There's similar arguments about clairvoyance. Some people think, I can understand why, that these are reasons for rejecting simple reliableism, but my point is that the True-Temp case shows that if you believed that, you wouldn't have to be an empiricist. Because you just have the beliefs, right? Who says that they have to come through since experience? If you get wired up to the world in the right way, and it doesn't go through sense experience, that's good enough. Okay. Okay, so let's get back to some philosophy of science. Okay, so this is something I call the pessimistic deduction. It's a lot of pessimistic induction, the pessimistic deduction. And it goes like this. The evidence of science concerns only what is observable with this subscript P. So this is the common thought here. The evidence of science is found in observation, brackets, philosophers, the analysed observations, the exception. However, the conclusions of science often concern what is unobservable. And here's the controversial third premise. No conclusion concerning what is unobservable can reasonably be drawn from evidence solely concerning the observable. From which we can deduce the conclusion we can't have reasonable belief in the conclusions of science. Of course this doesn't have to say every conclusion of science, but yeah, a lot of conclusions, many of conclusions of whatever it is of science. So for those conclusions we can't have reasonable beliefs in them. So I think that this is a standard form of an anti-vehilist argument. It's the sort of thing that's very implicit in Van Fraat, but it's not actually in Van Fraat. It's the sort of thing that sets out clearly what I think is at issue in many debates between realists and anti-vehilists. And it's no surprise that it's the third premise here is where a lot of the action is in the debates between realists and anti-vehilists. Because realists will come along and say, inference is the best explanation. That shows us as well, right? Inference is the best explanation so that you can have a reasonable conclusion regarding unobservable stuff from premises concerning only observable stuff. And the activity will say, no, no, we don't like inference the best explanation for whatever reason. So that's where a lot of the debate lies between realists and anti-vehilists. So the debate between what I call the optimistic empiricists, those who think P3 is false, the pessimistic empiricists who think P3 is true. Why do I call them both empiricists? Well, that's because both sides tend to accept this premise. That premise is a statement of empiricism in philosophy of science. And the fact that they are arguing over this debate shows that they're all empiricists, whether they're optimistic or not, realists or anti-vehilists, whether that's what makes the fact. Whereas I, what I want to do is to give reasons for rejecting that. Oh, yeah, am I going to... I could talk about this for a while. As an interest of time, I might whiz over this bit, perhaps. Of course, there are some people who want to resist this claim that conclusions of science confirm what's unobservable. Strictly speaking, Van Frassen is one of those. Because Van Frassen doesn't actually think that we believe the conclusions of science. So in some sense, he accepts that the conclusions of science in the sense of what you should believe don't include what's unobservable. He thinks we should accept our best theories. But acceptance is, roughly speaking, thinking that our best theories are empirically adequate. So he will reject this. I've got a couple of slides on why Van Frassen is the stupidest philosopher No, he's not stupid, he's not stupid. He's probably even taunting you. I just can't get my head around constructive empiricism. If you want me to tell you why, then please ask me again. I have a rant. I start frothing at the math and it's probably not for me to do that. Because it just seems inconsistent with the way that modern science is done. Anyway, I'll move on. Let's come back to that. I'm very happy to do that later. So look, here's my strategy. I'm going to be considering three skeptical problems that are problematic, yeah, become problems if you're an empiricist. If it's the best explanation, if that's the familiar testimony and it's the question here, all these will probably be familiar. What I need to do is to put them all into a common format. And the common format confirms the kind of response that you can make to these skeptical arguments. And I think that there are three kinds of response. First of all, you just accept a skeptical argument. Be an anti-realist. You can reject it. And I've said that the standard way of rejecting these arguments is just to be an externalist for some description. Or you can reject the argument, reject the skeptical argument and you can try and look for some a priori solution. Of course, this thing is unlikely to appeal to most comparisons because it looks a bit rationalist. But I also think that in none of the cases this is really likely to be a good answer. These are your two. The two reasonable responses are either accept to be a skeptic and an anti-realist or reject the argument for externalist reasons. This tends not to work. Okay, and what I need to do is I'm going to use as my model because it's perhaps more familiar. Hume's problem of induction. And show that how these are the three standard responses to your Hume's problem. Then I'll have to move on from that so I'll be here then testing the instruments. Okay, so what's that slide there? I'm sure there's a reason for that slide. Oh yeah, I guess it's because as I said, the argument here is typically about inference is the best explanation and whether we can make inferences about what is observable from what is un-observed but inferences about what is un-observable from premises about what is observable. Hume's problem is different one. It's about making inferences about what is un-observed from what has been observed. Okay, so it's not what was from that. But yeah, that's the problem. Can we from our past experience predict future experience or I'm not going to insult anybody by rehearsing Hume's problem. So I'm just going to go to the three solutions. Look, you can have these answers. Be an inductive skeptic, Hume bang on! Popper, right? You can't pop a inductive skeptic. Here's another response that people can make and is made widely by the literature including by the people who talk to me like David Pattenhill, Humano and others. If they just go externals and they said that when you analyze Hume's problem carefully you'll see that Hume's problem requires making internalist assumptions. It's it requires for me to be justified in thinking that just because for as long as I've known him Alexander can speak French that in five minutes time he'll still be able to speak French. What do I need in order for that belief to be justified? Well, it's not simply that I just have the evidence of his being a competent French speaker in the past but I'm able to justify my inference procedure. And then we get into the whole problem and it's all being circular. So it's internalist because my justification, my being justified requires that I not only have some evidence but that I'm able to show that that evidence supports that conclusion. And so the externalist response says hey that's internalism but if you're an externalist like the simple reliable list, you don't have to do that. It's sufficient that you've got the evidence and in fact induction is a reliable way of doing things. So that's the standard move the next externalist will make in response to the Hume's problem reduction. But of course there is another anti-sceptical response which is look for some a prior justification of induction you find this incantant you find it in other authors you find it in straws there are other number of people you find incantant obviously. Attempt to find a legal justification of induction my own view is that Goodman's New Riddle of Induction shows that this isn't possible. But I might go into that. But arguably you can try something along this line I think that's implausible for my argument today in this case it's not essential that I reject that although I will reject the analogue of this in the other cases. So that's Popper that's standard externalists Van Cleave she in Melon in the football. So simple and reliable is that inductive justification is possible. And furthermore this isn't definitely important but it's interesting they also go on to say that actually you can if you want to although this is not required you can actually justify the rule of induction itself by a circular enormously circular argument and rule circular inductive argument so you can show that inductive justification is actual but this is a additional the point about externalism is this is a voluntary additional move you're not obliged to make for this bit to be true. So 3 kinds of response to use problem oh yeah this is I think new rule of induction shows that that's not going to work but we can come back to that if you want let's move to in front of the best explanation more germane to fossilary science arguably particularly that P3 we were looking at so the question is is it the case that better explanations better potential explanations do we put out some hypotheses they all look explanatory of the evidence at some look to be simpler more elegant explanations or have some other virtues things that make them good or in Peter Lipton's terms lovely explanations are explanations that are lovelier better more likely to be true and this is what Peter Lipton calls Voltaire's problem but it's just really the skeptical problem of inference to the best explanation we can think of there being 3 kinds of response you might think that there is an analogue of huge problem for inference to the best explanation after all if you think huge problem is a problem you think that there should be something analogous going on for inference to the best explanation and you might think we just have to accept that we can't get justification or knowledge from inference to the best explanation you could reject it on externalist grounds or you could try and find you could reject the skeptical you could reject this worry by trying to find some a through or i justification a through or i link between explanatory loveliness explanatory goodness and truth people think that this is you know this is this isn't going to get you very far but we can argue about that but none of all people think that an a through or i justification of inference to the best explanation it doesn't look like a very hopeful project so we can either accept this and that takes us back to the standard philosophy of science anti-realist or we we can be rejected and in fact this is the this is the move that's made by the sort of upgraded version of the No Miracles argument that we find in Dick Boyd and Staphis Silos Staphis Silos is sort of a big book on scientific realism he adopts an externalism to justify to show that you can justify belief using inference to the best explanation but it turns out that his argument the argument that he and Dick Boyd make is in fact the analog of the the additional step that you can make in response to human's problem which is if you are an arbolist you could not only have justified belief in your inductive conclusion you can actually do the thing of saying well actually my evidence is that induction is reliable and that via that will be in itself a justifying inductive argument but since it's rule circular not premise circular that's supposed to be okay and in fact you can do the same the Boyd Silos argument is pretty much the doing the same thing but in the case of inference for the best explanation we could talk a little bit you know I've just waved my hands and said that's what they're doing I haven't you've gone into any detail in the interest of time but we could talk about that okay so as I said I didn't think that that finding the rule right justification for instance the best explanation that doesn't seem to be a plausible option okay and as it looked do the same thing again but with respect to testimony and with respect to instruments so here is here's what's known as reductionism in in the foresight of testimony we've done it to think would pretty theoretically think that some of the stuff have a justified belief in stuff by you getting other people telling you too but how why should you believe something on the basis of somebody else telling you something well this is what we find in Hume standard reductionist some people call it the empiricist account of testimony somebody asserts something so that's your point of view you didn't say it your second premise is that they are reliable and therefore you infer what they say is true but an important premise is this has been here that the person you are hearing this from is reliable so that's reductionism and the reason why reductionism is important in our context is if you're an empiricist you've got to do something like this right yeah and as long as you say something to me like the one time that trains out to Brussels tomorrow from an empiricist point of view all I've heard is all I know is Alexandra said these words right that on its own doesn't justify believing that content so there what I need is to be to be able to justify believing content is to know that Alexandra is reliable about train times and that's fine because look here I've known him and he's tended to tell truth in the past or something I can justify this second premise so that's how that works the problem is that the scale of the division of labour and science renders this it's very difficult to see how we can actually do this for science because where do we get our information from in science almost information in science is gained from journals we gain some from our lecturers but even then you know how do your students in the science lecture they seem perfectly reliable how do you know that they really are we're talking about lunch today sometimes aren't you some of these are professors of medicine so you read stuff in a journal there are all sorts of ways you could try and think of generating some kind of word reason for thinking that what's produced in a reputable journal is in fact reliable but if you're an empiricist that reasoning thinking that has to fall down to your own experience but how could you have enough experience to really know that all the science that you are consuming typically give them the large scale division of labour and science the amount of information you're consuming from the vast number of individuals scientists who contributed to that you could possibly from your own experience have enough experience to justify and believe in the reliability of your scientific sources so I think if you are a this reductionist approach you might work for small scale stuff like getting information from people that you know reasonably well whose reliability you are able to vouch for your own experience but that just doesn't work for science so then on the prime face we have a skeptical argument in science based on testimony so if you reject the skeptical argument and again there's three things you can do when you face it with this skeptical argument you say yeah you can accept it Locke has said something about testimony that seems to suggest that Locke was a skeptic about scientific testimony this is all regarding about this but he does say stuff that seems to suggest that he is a skeptic about scientific testimony and it was sort of part of the rhetoric of the royal society in the late 17th century the experimental method made you being there able to see experiments done for yourself possibly in a small society like the royal society where the other members were gentlemen whom you knew because you're all gentlemen and aristocrats and you knew them personally perhaps if you didn't attend a meeting you could expect you could accept their testimony these are people you knew personally from your right social class you could trust them but his idea the basic idea that you had to be either experienced stuff or stuff or be pretty close to someone who had was a thing there you could reject the skeptical argument on externalist grounds look just go back the requirement that you are able to justify this premise that's internalism for you the fact that the person called you is reliable and simple and should be enough now the Sanford Goldberg has a an externalist response a defensive testimony but his reliable is more sophisticated than the simple reliableism that I was advertising about in the beginning and correspondingly more plausible he says that here I can gain some knowledge when three conditions are met the speaker is reliable secondly the hearer basically you gathers the correct proposition from what's being said and believes it so that sounds fancy but basically I understand what you're saying and here's a third condition so this is where things are an addicting simple reliableism the hearer can and does reliably discriminate amongst speakers for those who are reliable and what's important about this third thing here is is that internalism coming back is this one is this saying the hearer must know that Alexandra is reliable but Charles Charles well no I'm able to say why well no all this is requiring I'm I'm able to not that I know from my own experience who is reliable who's not just that I am able to make a distinction between them well this is the person who is being utterly gullible right the utterly gullible people who will believe anything anybody says to them are not in a good position to gain knowledge by testimony but if in fact I'm fairly reserved and I I I will I trust say people are, yeah, Alexander but not Charles because he had too many meetings today so I'm a bit I just make that discrimination then even if I can't say why I think Alexander is reliable as long as I make a accurate discrimination I limit myself to the reliable then that's yeah that's good enough in the science case it might be a matter of trusting journals that are well known and being less inclined to trust journals who's citation in the disease of rubbish or yeah you've never heard of we know that there are all these dodgy scientific journals out there that will publish everything for money you know, as long as you think I'm not going to trust them but I will trust the ones I've heard of then that's enough for this condition to be met so the important thing is that this is not a intervism coming in by the backdrop which is not intended to so let's say look you've got this externalist approach now of course in fact there are people who think that you can test the money on grounds that are acceptable to internist by a priori methods so Tyler Bird says we are entitled to rely on other things being equal on perception memory deductive and inductive reasoning and the word of others so he thinks actually that we have this default entitlement and he gets in fact we have a priori justification of all these things including induction and that includes testimony and the word of others and he gives an argument for this an a priori argument which is roughly if you can make sense of what they're saying then that shows that they must be rational but rationality is a guide to truth there's this link between rationality and truth and therefore if there's defeating reasons you've got a reason to believe that what someone is saying to you is true I mean that just seems like a pretty weak argument to me just because I know that you're rational because I can make sense of what you say the evidence of you're saying something comprehensible is that you're rational that doesn't give me much reason to think that you're likely to tell me the truth it might be used to give me reason thinking that might be telling me truth or you might be telling me a complete lie because you're a rational person who wants to deceive me for whatever reason rationality your rationality is a sign that you could tell me the truth if you try to tell me the truth that raises the chance of what you tell me being true but it doesn't that your rational gives me no reason to think that you want to tell me the truth so I think that Burge's argument is very weak and we can look at science yeah research is sort of interesting because you might come up with some arguments as to why the structure of science where other people get rewarded for success in science so other people are motivated to come up with the truth by the reward and centre structures of science that might be a reason for thinking that people in science are more likely to use their rationality to tell the truth than to tell to tell the false hurts or to get things wrong it's true but actually the intensive structures of science also motivate quite a lot of people to cut corners, to engage in questionable research practices or even outright fraud as over the border on whichever direction is north of the area, in Dietrich Staples wonderful, I admire this forwarding psychology it's just brilliantly done the quality of this stuff in science is such that you could point to it and say the incentive structures of science are set up actually for in a way that doesn't encourage truth to tell it always I mean a couple of things but even if it was halfway decent the actual facts about science you don't necessarily go along with this the examples of questionable research practices and even outright frauds just otherwise so I think the only options with respect to testimony in science are skepticism or an externalist rejection of the skeptical argument and possibly gain the interests of time we've come up to an hour but we started 10 minutes later so I'll just um so what should I do look I was going to talk about instruments but look, I think you can see where I'm going here right inductive justification of what I'm going to say that was this is the the testimony thing look, science uses instruments a lot but we've got a similar issue we've got an instrument that delivers some output propositional output reading on a dial but these days more likely you know a database why should I believe what the instrument tells me well because there's arguments right but I get to use it because I write this the instrument tells me that the instrument is reliable that way I believe that they um but there are familiar this is the reductious argument with respect to instruments there's familiar arguments as to why how do I get to know that my instrument my experimental technique will generally is reliable look, we can talk about what this gets to go on and so on but it's I mean one way of looking at it one way of quickly getting there is just simply to piggyback on the testimony because so many of our instruments are made by other people right and we borrow it off the shelf or we get sent to the Vice-Bio-Scientific Instrument Bakers um in other cases they're laden with theory and that may not be theory that I'm able to justify myself um particularly because I'm using this instrument and it's doing something clever designed by Phidius I'm a material scientist but I'm an archeologist you know that sort of um what I'm interested in is you know the age of this rock or bone or tooth um and I've got something clever that tells me what that age is but I'm not in a position to justify that theory because I'm not a physicist so all sorts of reasons for thinking that the reductionist is talking about this one doesn't work it's the same thing, get a sketch of an argument which we could either accept or we could reject on external aspects, you know that's fine so long as the instrument actually is reliable then we'll give you justification and knowledge even is it possibly even more obvious that an a priori justification of an instrument isn't going to get you very far so these are the two of them alternatives okay um look some people might say I think this is an important thing I think that often the fossils of scientists are quite sympathetic to interdism because they say no but scientists should understand the methods they use they should not be good scientists and just accept a method without understanding it at all I think there's something right about that first of all I think that scientists what we expect as scientists when scientists expect to be each other is that they should have some understanding of the methods that they use but not a sufficiently strong understanding of justification by internalist standards you know so if I'm an archaeologist using radiometric dating using isotopes to date this rock I should have a rough idea about you know how isotopes you know how there's exponential decay and different rates of different things I guess I'm understanding that but if I were able to give a full internalist justification I probably wouldn't be able to do that and first of all I think that we can we have reasons for expecting to be able to justify their methods there are independent of interdism basically the argument is this look sometimes if we want some process to achieve a goal we can sometimes but not always we can be expected to need to know whether our process is in fact reliable then you might want and especially if implementing this process is expensive then it's make sensible to inquire about which process really is going to deliver the goods but in other cases it may not be so important that you do that now in the case of science you know that we're developing methods all the time and we're going to be putting a lot of intellectual and other resources into into our science it makes sense for us to want to know which processes, methods techniques, instruments are reliable because we care about getting to the truth and which one we choose is going to be expensive to do so it's going to work but this kind of justification is independent of internalism so it's a bit like this it's okay so I want I want not to get carried away I have my pad but back when it was really dangerous how do I so I want to get vaccinated since I really care about not getting COVID and it's going to make a difference in my life whether this vaccine works or not I'm going to take some interest in which of the vaccines are good or not yeah and I will AstraZeneca developed in Oxford made in Belgium but I don't know I'll have that one Spotmeg, Putin's one um um sorry but that's that is not that's not motivated by tournaments it's just motivated by these kinds of consideration and the fact that the only way of knowing which vaccine is working is to investigate the vaccines themselves okay um so let's move on last couple of slides anyway look look look someone is going to say I'm empiricist but I'm not an internist and here's my reason and but perception um doesn't deliver certainty so if you this is an intersubjective notion of perception um but it does provide high degree of reliability and further you get away from perception less reliable stuff becomes um you know my perception I really know that this is a project I have in my hand stuff I found out with instruments my temperature was the moment I thought that's less reliable yeah I can't be so confident because I don't know if this is reliable sometimes these things go wrong um actually um so this would be a reason for being empiricist because you think that perception is not certain that it's better than anything else we've got actually I don't think that's right um the whole point about science is perception is rubbish and using instruments is a whole lot better um think about oh it's a big thing for medicine temperature right, thermoception you know we can tell using since a thermoception which things are hot and which things are cold but it's not desperately reliable it depends upon the state of our bodies you know the whole thing about putting hands in cold water and hot water can mess that up our real real thermoception isn't so good and it's also not very precise you could say oh it's hot, really hot quite cold nothing precise enough for science that's why we go to thermometers, okay mercury thermometer it's much better than thermoception um so it's over here still problems, parallax errors and the rest of it well you might move to a resistance thermometer a resistance thermometer and you have a digital readout or even better you will have a method of continuous temperature monitoring linked to a database where that database is you go through some analysis and that gives you your temperature so I think that even this sort of sense that this justification for empiricism doesn't get us any more for the rest of the game it was a quote from Jim Bowen that says what else is it okay I'm just going to so I think that, oh look yes, this is what we've got to look forward to if you particularly over next term, post human unconscious information up in the future this is giving a slightly wacky bit of a talk but we won't need to read books let's say none of us read books anymore we'll talk about this later but look why shouldn't stuff be uploaded into our hippocampus why couldn't that be done no experience we just find ourselves believing in stuff I need to upload into your hippocampus the period at table and then the next day you just have to know which element has which time of number and so forth we'll have to have beliefs about it well if I'm doing a good job of uploading that stuff into your brain why shouldn't that be a way of knowing stuff okay so let's finish conclusion so I started off with this pessimistic induction debate I've said that there's been too much focus on this premise whereas I think this is the one that we ought to focus on as well or instead I looked at three skeptical arguments inference-based explanation testing instruments three kinds of responses you could make well, only two of which I think are really plausible either being a skeptic and accept these skeptical arguments or reject on externalist grounds so maybe you're alright but we didn't really like that in any of these cases so your options are being an empiricist accept the skeptical arguments so you're an anti-heroist because of your internalism you did alternatively you can be an externalist you reject these skeptical arguments so you're not an empiricist and at least allows you to be a realist and I said possible realist doesn't force you to be a realist but allows you to be a realist and releases you from the anti-heroist argument now obviously if someone places it in what I'm saying that I prefer to be over here right but in the sense that the real argument is what I would take away if you believe anything if you believe anything these are the reasonable alternatives these things go together for a long time to say 5 minute breaks as usual and after that it's time for the formal comment before the general discussion and I'm very happy to give the floor to Pilate Reis, who is a postdoc at Lucille Valle thank you very much and thanks for this opportunity it's been a pleasure to read and prepare the comments I would like to start by saying that I approach this paper with a lot of interest because of the suggestive title and provocative title against empiricism my initial intuition was like but if not empiricism for science then what what do we have, what do we rely on and then reading the paper is when I realized that my initial maybe naive intuitions were wrong and needed to be need to be more precise so it was a great opportunity to think about what empiricism means with justification, perception experience and even what's the distinction between scientists and the scientific community also my comments will probably go in the direction of developing or precessifying a bit more all these intuitions so the first comment is motivated by so in the paper we are presented with in the talk we are presented with different dichotomies in particular we have to see that we have internalism and form of externalism liabilism empiricism and non empiricism and skepticism and realism this dichotomies my first reaction my first intuition was that they were about a particular scientist and then at some point in the paper I was thinking no it's not about a particular scientist it's about the whole community of scientists and then I was trying to put some order on this and then I thought okay maybe internalism externalism is more oriented into the individual level of each scientist but not necessarily empiricism and rationalism or non empiricism can go in the two different levels I guess and skepticism and realism can be justified for a particular individual and for the community so there's I guess whether we choose one or the other level we will be situated in two different debates one is epistemology and the other is philosophy of science which of course will have connection between them but in epistemology when we talk about unobservables maybe we talk about other things to what we talk about in the scientific community so my first comment and question is about where to situate this dichotomies at which level and in particular I was thinking of whether it's possible to have different responses depending on where we are in one level or another for instance we could say that a particular scientist should be externalist because he should rely on what other people said reliability and realist however the whole community of scientists should be empiricist because someone somewhere has to have the particular experience for one particular justification for something look that's really perceptive and that's quite right and to try to identify the social and the individual and that's something that in the book that relates to this talk I do go into more details so you're actually right that's an important point that said my view about the community is actually the community is itself very much like an individual and that what one says about individual epistemology can be carried over to the community as a whole but I do in fact have a section on whether exactly what you said whether in some senses empiricism can be true for the community of course there are not two ways of looking at that one is whether you think empiricism might be true for the community because some individuals must have had the relevant experiences what you said but another thing you could also say is there a community analog for perception and I think that might be an interesting avenue to explore but let's talk about what you suggested which is as a community as a whole is it necessary that some individuals have had some experiences well I'm not even sure that's true unless you think of our community as this historically extended thing and that's part so it doesn't mean so I don't think any I certainly think it's not true but many current individuals have to have any experiences or recent individuals possibly you could construct an argument that we couldn't be where we are in science without past individuals having experiences and then they develop the first knowledge upon which subsequent science has built but one of the bit I left out about Van Frassen emphasises the fact that perception plays a minimal role in much modern science so just think about Large Hadron Collider what do the scientists who discovered the Higgs boson, what did they see they saw their computer screens right well is that important it doesn't seem to me that in the epistemology of science needs to have a special place for looking at your computer so that's why on the other hand it might be that doesn't work if you think of the community historically extended there is an important place in L.A having his experiences whether that's of falling candles or through a telescope I think there might be a version of your argument that says some past historical experiences were essential but I don't think there are now I think that the next comments maybe they are a bit related maybe they are guided by the same intuition they are more focused on the pessimistic deduction and on the particular terms that are used in the pessimistic deduction in particular so premise so the problem is that we have premise one premise two in which we use observable and then we cannot go from or we have problems in justifying how we go from observable to unobservable in a very particular sense of observable so my first reaction about the one comment they have about this is about the distinction between perception and instruments because we use the notion of observable using the notion of perception and at some point in the paper there is an intuition that I completely shared and I think it's very interesting that says that senses are just one among many of the instruments we have so there is no privilege sense the privilege role in our senses this made me think well if that's the case we have a clear distinction between what is observable and what is observable is it observable something so there are clearly situations in which things are observable I can observe the computer however if I'm very mild or I have very strong problems in my and then I get some surgery this surgery has modified my perception of what I can see that I can put glasses I can refine more the way I see or the way I feel in my skin in any sense given that it can be difficult to have a sharp distinction between two I was wondering whether this could put a problem in the distinction between observable and observable so I guess my question is to know a bit more about this distinction right yeah so the sorts of things that you're talking about are indeed the sort of things that have in philosophy of science people have been arguing about but they argue about it precisely the area that you're discussing which is in connection with perception or what's the limits of perception a telescope can't or all you have is surgery to improve your eyesight you know I actually agree that all these things put pressure on the unobservable observable distinction is used here with the subscript P and I suppose one of the things that I'm interested in saying is since I reject the first premise I'm not so interested in that question still that's not to say that I haven't got a distinction I just located something completely different because I distinguish between the observable and the unobservable because it is part of the job of science to help us find out about things that we regard currently as unobservable but so in the book I have try to give an account of what I mean by observable and it's roughly using the techniques of science to generate knowledge that will be evidence in a relevant area of inquiry so observation is a matter of evidence generation generating knowledge or evidence and knowledge are the same and it's not related to perception obviously so roughly speaking it's not a global distinction general distinction between inferred or non-inferred that's one way you could draw it here the observable observation is a matter of what you know without inference then you get to make inferences to the unobservable it's something like that but as a word it's relativised to fields of inquiry because in fact making observations of gravitational waves involves tons of inference but that inference is all highly reliable and it's built into the process it's not speculative inference it's inference that as it were we make in a highly reliable way to produce evidence which then will be used in as it was slightly more speculative inferences possibly what might start out of speculative inferences about we can use gravitational waves to make inferences about the truth of or possible not truth of general relativity that's why these things are observations because we regard the process of thinking of this observation utterly reliable as highly reliable as ways of generating evidence if it was unreliable or only speculative then we wouldn't regard it as generating evidence that we could use for some other purpose that's different so that's not a very articulate answer but that's where I locate the difference the edges of reliability within what's reliable producing evidence to make inferences where we're being more speculative I wonder I think that my second and third comment were very connected so I don't know if you already answered so I'm going to slightly change and I will mix my third and fourth comment it's more general about how would not philosophers of science but scientists think about these things and in particular about this argument if we forget about the particular notion of observable that relies on perception and given your last comments in the paper and you also commented this at the beginning of the paper of how scientists talk about observation of waves, observation of these observations of that things that we consider unobservable and in the paper it's claimed that this is a rejection of premise one however I was thinking that also a rejection of premise two in light of the scientists they would say we start with observations and we conclude other observations you're good yes there are results that is relatable I've just said so I think that there's something right about that and I think as it were that the conclusions do concern observations when it's the case that those conclusions have been brought about by processes that we regard as entirely reliable so that what we get out at the end we're confident about using as evidence for the next process of science so something will change from being an unobservable to an observable depending upon the reliability with which we can generate the information about it that's exactly right so once upon a time electrons and atoms were observable because you've given the evidence that we had at the time it was a little bit speculative but once we were utterly convinced that we were doing what we were doing then we observed these things all the time and that's why we can use looked on microscope I've got one I think I can infer from your first from your answer to the last first comment I'll answer it but I want to ask it anyway because one argument in the vicinity that I wonder how you would reply to and you'll catch this floating around the philosophy of biology and literature sometimes I'll see it in sort of uncritical forms occasionally from people like deniter Dawkins who give this sort of empiricist argument for an empiricist reliable where essentially things bottom out in natural selection right so you go look we're in body brains in hunts of meat what could guarantee the reliability of anything winding up inside of brain it's got to be something like repeat connection with the external world refined by natural selection or something on this word because that's the only kind of process that we know of that's at work in the world around us to do that kind of refining over time so at the end of the day we're justified in something that feels like empiricism on those kinds of bases not predict that you're going to disagree with the definition of empiricism in play but I just wonder what you because it's a neighbouring argument and I at least I've never known what to do with this argument that's why I'm interested in asking what happens if I say look I quite like that argument look I can imagine look I can imagine there's some responses to it that someone who might not be quite so marvelous some skeptical responses on that well just have to be good enough in the context and that doesn't necessarily mean truth something but less than truth might get us you know get us survived but ok but what's it got to do with empiricism that's the question so I think that would be a good argument for thinking perceptual approaches senses that generally deliver the truth at least in normal circumstances revolutionally like circumstances circumstances like the ones under which these capacities were evolved ok so what does that tell us about empiricism as a philosophical thesis and does it it doesn't follow from that that we have to ground everything in experience it might tend towards something like what Piran and I ended up discussing that historically as a community perhaps experience plays a whole historically and this is helpful for that view that could be a kind of empiricism that might come out of our discussion that would be supported by this kind of argument but as it were it doesn't tell us that I mean if you think this is part of my argument towards the end I regard some of the stuff I said about this in the paper I regard some realists it's not a game of topic that some regard instruments and so forth there's ways of extending our senses I think like a telescope helps like being closer to the thing extending our senses I think that's actually the telescope is a poor example because actually most instruments in science don't extend our senses or they do something our senses never could do in the first place like a gas meter for detecting magnetic fields I mean some animals actually can detect but we can't I don't think so why did I mention that so where does the David Dawkins get us there what did at the beginning it tells us that they're pretty reliable for some constrained circumstances but the point about science is we want to get beyond those circumstances we want to get into realms where that stuff has never evolved to deal with so we need to do something different and that's why we have instruments and fantasy theories and the rest of it and that's to replace many cases or do stuff that our senses never did so I was surprised about always your argument about the not about the argument but about the third option because what you're asking here is for an a priori justification so in a certain way to be able to a priori prove something about the external world so to produce knowledge about the external world and no rationalist would argue that it's possible today so most rationalists would say we have some cognitive structure that is innate and these are necessarily acting when we acquire knowledge but it's not obvious that we can acquire knowledge about this structure think about general grammar and linguistic or things like that so there's a structure cognitive structure that is there but it's not strong enough to produce knowledge by itself so you need some input from the empirical world so it's not incompatible with their second option so it's not empiricist it could be realist but it's not obvious to me that there's not an internalist position there about this kind of rationalist that they could be internalist you think I don't know I'm asking you right, okay, right right, right, good I think they're not empiricist in your sense that's clear that's absolutely clear and they could not be they are reliable in the spirit but they cannot produce knowledge based on this because this cognitive structure is there it's not necessarily that you can know with a priori it's like a neo-kentium it's just frame your experience but what we can do is link it with the the staff the people produced by evolution and so that's the reason for thinking that in fact that when I guess I believe that these structures are evolved ones and one reason why they're evolved is so that actually small amounts of information can multiply them but the distinction with the argument is that we can produce scientific knowledge on this structure using this structure and a neo-kentium would say how do you know that how could you do this circular well see that's when they're being internalized I think if you didn't make that move I think you could be right with next okay my temptation is to think that as I said there's some as a liberalized version of empiricism with which I don't have an argument which is at least I'm not arguing here which is you have to be connected to the world interacted with the world to know stuff that's the way that's the way that probably somebody like darkens would try to push back like at least show you that the right kind of sense perceptions have to be in the causal network somewhere somehow for stuff to have entered my brain and then so that's enough to suffice as empiricism maybe that evacuates empiricism so much that there's not much interesting left on the table I don't know but even if you're born in other countries you might get into here's where rationalism is by it's left behind as it were the origins that I was talking about with the certainty but there's innate structures that in something encapsulate knowledge or at least enable us to generate knowledge from small quantities of empirical input more than can be deduced from those inputs on their own and then since I think that that doesn't there's a successor to rationalism and there's a parallelist successor to empiricism there's some ways of describing both of them that make neither of them interness but it depends in which case I'm finding both of them and they can fight it out they can come to some they can come to some kind of accommodation which I sort of suspect is what evolutionary theory does for us so yeah that's interesting but yeah it's interesting to have some of these things thank you thank you very much for your talk I've got a very large tough question it's about the internalist externalist distinction because I think that's it's often with internalism so you ask that the person himself has got all arguments with you in order to justify his or her beliefs and justify justify them it means deductively justify them you will need sufficient reasons for having these beliefs which is surely to require but on the other hand externalism just makes the person not having the reasons in order to justify his or her beliefs well I think if I see you in front of me that's a quite good reason to think that you are there even though it's not probably a deductive sufficient reason but you know it motivates my belief it provides good reasons but not probably sufficient reasons and the same way like induction induction is not deductive and valid but of course having lots of experiments going the same way provides you good reasons not sufficient reasons to have this belief so of course if we want internalism to have deductive justification it's too requiring and then we would like to go to internalism but is it not just because it's too requiring to have deductive justification in the case of internalism? I think that's right and I think that that's why there is this close connection between internalism and skepticism because it is too demanding however I think you're also right in applying this what you're saying that simple internalism isn't demanding enough so what we see as a result is that and Samford Goldberg's response to the case of testimony is an example of this what we see is relabilists sophisticated relabilists build in a little bit more to their versions of relabilism that make it little bit more demanding than the simple version for example that in his case it was I can I have some reliable way again this might not be something which I'm aware but I do in fact distinguish between reliable and unreliable sources and so I think you put your finger on it it's the question how much should we require and my view is that in science we talked about this at one point and in other areas we require more than the simple relabilist but not enough to satisfy the internalist so I think that what you say is I think that's right I have no answer I've seen that quite a lot of time we are very demanding for internalists I agree I think that part of our knowledge doesn't justify does not have to I agree but that's why I think that but it's funny though isn't it so if you think about children and animals your children and animals favourite externalists love animals and children we like children and animals because children and animals aren't reflective but we want to say that children and animals know stuff and if you think and a kind of notion of justification with which the externalist likes to work is one that is if you know something you're justified so there's a sense in which children are justified in believing things even though they haven't really thought about them and they certainly haven't gone through any eternal reflection on their producing processes nor are they in a position to do so so it's interesting that when it comes to adults and in particular to scientists that we do get more demanding than we think that scientists should know stuff about our methods but again for the reasons you said not so requiring, not so demanding that they get as far as the internalists would require them to be but that's why internalism if you follow it through will tend to lead to skepticism unless you can be super duper clever and come up with a priori that are deductive justification of induction and all these other things but what I've been arguing is that all those views are a bit hopeless really I mean I agree with you but I think that there's a sense in which from the externalist perspective a whole bunch of evidence is sufficient without a deduction that's what the externalist is saying if you've got enough evidence and you reason inductively that's sufficient, that's enough I mean it's not deductively sufficient but it is sufficient for you to get to know or to be justified and could not not be possible for an internalist not to be deductive to claim that he can reason on the basis of some evidence and infer some consequences while knowing that there are not deductive consequences he has got justification for some of this or her conclusions on which he or her she responds but without the same kind of deductive justification that leads to absolute certainty I think some people have tried this in different ways a little strawson tries it in the sort of typical analytic philosophers kind of way he says that's just what it means to be justified or be reasonable being reasonable just means believing something when but only when you've got a decent body of evidence right and there are moves that made like that I think that's missing the point because as of what this skeptical challenge is being made is I don't care whether my conclusion is called reasonable what I want is what I want is have some reason to think it's likely to be true and that's the notion of justification I'm interested in is one which is correlated with being called reasonable it's just a word this is getting slightly off the topic but I think there are attempts but then there are also the recantian ways of doing things a little bit more sophisticated than that the neo-agilean like sellers position yourself in the space of reason it's internalist you don't have access to the reason you cannot position yourself in the space of reasons and to provide reasons you're externalist but you're right they don't claim to be realist yeah that's right yeah it's kind of related but I was wondering about this reliability strategy that you mentioned by Colbert you seem to be quite sympathetic towards it but I was wondering how you can avoid problems with accidental discrimination capacities like if there is no good reason why you are able to discriminate you happen to be able to discriminate you happen to be successful in discriminating just by pure accident or by some neural network that has not because of the real content but because of it has the right so imagine a neural network working on results of scientists and if they have a nice kind of grammar that successful scientist uses and I don't know they look sophisticated and maybe that's a good maybe that correlates with reliability but actually this is just pure coincidence at some point somebody who doesn't have this nice sophisticated language could exactly have the right sort of insights and the sophisticated person could be a fraud so it seems that there might be deeper reasons for reliability maybe I'm too internalist than just to be able to discriminate I mean you have to be able to discriminate but also for a reason it has to be intentional discrimination and not just like extensionally making the right that's my intuition good, good that's a good question but they're all good questions so I think that the notion of reliability here has to have some modal element to it it's not simply that the criterion of discrimination I'm using in fact distinguishes the reliable from the unreliable but you would do so in nearby circumstances so it can't simply be I trust only people that don't have beers and I don't trust you and it turns out that for purely spurious reasons in the community where I engage the people, the bidded people are unreliable and the screen-shaving are but yeah clearly if you shaved it wouldn't turn you into the all of us you're a little bit more so that's right there has to be something slightly modal about it nearby possible in the world as well so the question is but even then you could say look it could turn out you might be able to run a version of your argument even even so I'm not sure what to say I mean I think at some point I just say look that as long as you've got this modal bit in place that's good enough but I can see but I can see how the temptation is to go well and all to be able to give us some explanation of why they make this discrimination and why it's a good one to make but I think at this point I'm saying that's just too eternist and I'm not so we might find that this is an area where we end up seeing where the eternist and the next eternist disagree but at that point our intuitions about being quite difficult to just but I see exactly what you're saying but the modal I think it's important to find to have the modal bit specified exactly and not just left to some sort because that's what philosophers always do you need the modal thing and you need to make it intentional and so on but then what that exactly comes to is not very certain and very clear and I mean it seems to like what kind of modality would be using then sort of physical modality or what I don't know I mean there's all of kind of problems that always show up when you're invoking modality that make it not so obvious and I think once you can settle when you have a good approach to sort of that a good kind of modality that does the work then I might find it also quite attractive but it's not trivial it's worth it needs to be done I absolutely agree I absolutely agree with that but I think it can be done I guess one test for this whether you think whether you are sympathetic to what it can be done is you know the whole chicken sex thing do you want to talk about chicken sexing? I'm not sure I get this exactly right but it's one of these things that's been bit garbled by fossils and that will include me but people are trained to distinguish chicks that are male from female but they're virtually, they're indistinguishable they seem to be indistinguishable but this is important because you don't want to allow the male chicks to you don't want to feed them because you want only the female chicks to go to hens and legs and so forth so but apparently you can train people to do this so this is the male one, this is the female one and then you get them to guess and then they guess you randomly begin with and after a while they get quite reliable about this and in fact for some time I wasn't knowing quite why they could do this but apparently it turns out that they are detecting something in the smell but they don't know that right so there are several levels of Julia's looking very doubtful about probably talking rubbish now but if it's not true there's a possible world to end right and then the question is whether this looks like this is the kind of reliable discrimination I'm talking about here people can discriminate between the female and male chicks they can't tell you how they do it but you want to say they don't know which is which I'm glad to say they know which is which that helps yeah that makes it ok thanks should we ask do we say ok I thought it's a very good empiricist internalist non-realist yeah right so here's why but I sort of so this is where I was catching the slides so you in the way that this premise is intended we don't believe the conclusions of science we merely accept them we take them to be believe that they are empirically adequate so since the conclusions require what we believe don't concern with some observable so he might just get off this whole thing there ok so I gave the example of the Large Hadron Collider in the ATLAS experiments there what does the scientist actually look at when she's doing her science what does she see she's looking at a cathode ray tube like her sort of predecessors would have have been nor even at a photograph of a cloud chamber she the Large Hadron Collider the beams of particles which are whacked together with your you want to start up behind you with particles these beams are this machine is constructed right and you turn a switch on and stuff happens and there are these an amazing array of instruments for picking up the decay particles decay products of these collisions or at least that's what the realists would be be saying and ok and in fact there's so much data produced and this is all electronic this is all electronic and there's so much data produced apparently that if you were to run these ATLAS experiments for 6 days and collect all the data it would fill up the whole internet right and obviously they can't keep all that data so this is they've got highly theory laden means of in real time analyzing it by the way most of it hanging on just the data that's important or the conclusions of certain bits of analysis this all ends up in huge great database which the scientists then interrogate the database using fancy statistical software run on there in the laptop right ok so so we're saying the perceptual experience of the scientists is limited to what she sees on her book and the much is limited to what she sees on her screen ok so what how should you understand what's going on if you're Van Frassen so um that for that matter if you're quiet given that quote about predicting future experience in light of past experience um so really what for both of them the aim of science is to produce theories um we don't believe those theories but they are they're taken to be empirically adequate which means that they correctly predict our experiences that's to say they will correctly predict what the scientist sees when she opens her laptop and types in certain order instructions in and then some stuff will happen we need to think hang on we've just spent billions billions in order to um get to to be able to know the truth the certain sort of conditionals of the form of if you were to build an extremely expensive thing to this specification attach your laptop to the open up this software and type these or type these commands you would see a screen that looks like this that sounds like I'm being cautious trivializing it but actually I can't really see what else the constructive empiricist thinks is going on in this highly computerized modern science I had a different example in this paper on regarding radio astronomy but it's much the same thing right you've got this sophisticated piece of equipment that you generate data and you interrogate that data using software and there's nothing like looking through a telescope let alone seeing a sample with your own eyes it's just anyway I think it reduces science it's difficult to see that it makes any science it makes a science plausible that's the argument that's the argument it's a little bit late I think you believe that Koin and Van Trezen are instrumentalists which they are not so they maybe have a little bit more sophisticated answer but I think we are at the end of the time and we could finish this discussion around a beer thank you very much for a best place to discuss thank you very much Alexander