Epistemology of Testimony Explained | Stephen Wright | Oxford Philosophy

[00:00:02]okay so let's begin by imagining a situation imagine you've just arrived in Edinburgh for the first time you haven't been before and you want to know first things first where your hotel is you step off the train and want to find the right way assume also that it's 1993 so you don't have a phone with the GPS to point you in the right way how are we going to find out how are you gonna find out where your hotel is well it looks like the obvious thing to do is to ask somebody that seems to be not only the obvious thing to do but it seems to be the thing that you're supposed to do asking people and listening to what they say and believing what they say is a way of finding things out the epistemology of testimony which I'm interested in is the study of how it is that we come to find things out by asking people and believing what they say it looks first of all like a lot of what we take ourselves to know or found out does depend on believing other people somehow I mean in my case I take it that I know that the Battle of Agincourt was in 1415 but I wasn't there to see it particularly I can't remember it happening my memory doesn't stretch back that far I take it that I know the capital of Uganda is Kampala but I've never actually been to Uganda I take it that I know this because well someone told me and I take it that Neptune has 14 moons but I haven't seen any of them I haven't even seen Nettie I haven't got that kind of visual apparatus or equipment to help me along so not all of our knowledge from testimony comes in the form that we might imagine of us going up to people and asking them questions and they tell us something some of it's written down some of it takes the form of diary entries we overhear people saying things and come to learn like that but nonetheless it seems like if we take out the stuff that we know from testimony it's hard to see how we know very much at all testimony is a really important way of finding things out we might think think about the case where you got off the train in Edinburgh I mean how you might wonder do you know that you're in Edinburgh assuming it's your first time being there well you might look at because the sign said so but the the form of testimony it might be because you heard an announcement saying we're arriving into Edinburgh just beforehand announcements are a fairly obvious form of testimony you might think testimony gets everywhere in our ability to find things out and this is why it's really important to understand how it works the puzzle with testimony I guess is the idea that well at least sometimes we do find things out by believing what people say all of those things that I mentioned before are cases where testimony allows me or us to find things out that we wouldn't otherwise have been able to find out we're successful in finding these things out but we don't always find things out by believing what people say sometimes we might believe what people say and thereby able to find things out sometimes people lie to us it's in their interests to have us believe false things so they tell us things that are false and when we believe them we don't find out anything at all we're deceived other times people don't intend to deceive us but they don't really know what they're talking about they just speculate and if we believe them then that's not really a way of finding things out either if I did guess and the number of moons that Neptune hasn't tell you that doesn't seem like a way for you to find it out even if I get right that's not really a good way of finding these things out so we don't want to believe just anybody in these kinds of cases we want to work out the cases in which testimony is a good way of finding things out and compare them with the ways that testimonies not a good way of finding things out and that allows let us understand how testimony works that'll let us understand how we get to find things out by believing testimony it looks like a lot of our beliefs based on what other people say amount to more than just guesswork it doesn't seem like even if I don't know how many moons Neptune has it's not just because someone told me and that I haven't seen for myself testimony is a respectable epistemic source it's a respectable way of forming beliefs and the way that we're going to try and work our way through and standing how testimony is a respectable epistemic source or a way of finding things out involves having a think about some cases some other cases - the one that we mentioned getting off the train and thinking about vertically what we want to say about them whether or not they involve with people finding things out in a good way or not and what exactly that shows us about a theory and an account of how believing other people gets us to find things out about the world so without further ado let's launch into our first kind of a case our first two kinds of cases we might meaningfully compare these two types of cases so I imagine a doctor performs thorough medical tests on you and informs you at the end that you have a particular medical condition it seems like in this situation everything's going right everything's going as it ought to go you can find out by believing the doctor that you have that particular medical condition everyone seems to agree on that point compare that with the second kind of a case imagine a doctor appears to you to perform a thorough medical test and examination to find out whether or not you have a particular medical condition but but unbeknownst to you your other doctor you can't tell the doctors actually just guessing the test is unreliable and doesn't work but nonetheless at the end the doctor tells you you have this particular medical condition and you've seen them do what appears to be a test it looks like it's less good in an important sense but you don't know that it's any less good okay so now we have two instances of belief form based forms on the basis of testimony we have two instances of beliefs in what doctors say the question is in each of these cases are we equally well off in terms of finding out what kinds of medical conditions we have or you have well according to our the first kind of theory of testimony we're going to look at the answers yes in terminus theories of testimony take it that's what matters for your finding out about your but what matters for your finding out facts about the world is the reasons that you have that support the belief that you form so in the case of the doctors what makes it the case that you find out in that first case in that good case is because there are some things that you can say when someone says well why do you think you have this condition we can say well I saw someone who looked like a doctor two million they did some tests and those tests look pretty good and then they told me that I have this condition and that seems to mean that I have this condition reasons here are things that you can bring to bear in an argument when someone says why do you believe this the idea is that your reasons are what comes out in response and it's these reasons that support your belief so the question is whether or not this in terminus theory is right the internalist theory just says a listeners belief is supported by her reasons for thinking that the speaker's testimony is true that's what justifies belief according to this kind of internalist story the case of the two doctors comes from Frederick Schmidt and he argues that look intuitively or it appears to us that in terms of justification in any interesting sense these people are alike in the good case and in the case where the test isn't real both of them can bring to bear the same points the same kinds of considerations because things appear similar to them in both cases it appears to them like there was a doctor and the doctor did a test okay in one case that's not actually true but nonetheless the listener doesn't know that so according to internalist theories yes these two cases of doctors and patients are both alike one of the nice things internalist theories is they can capture a particular thought that we might think is really quite important for the epistemology of testimony it seems like we can't just go about believing anybody we need we need some reason to think that people are saying true things before we can go about believing them we don't find out things just by looking in the right direction and having people say things at us and this is really important we might think and internalists have a nice story of why this is the case in the case where you got off the train at the beginning it seems like the thing you're supposed to do is go and after somebody but not just anybody not someone who's walking around and looking as confused as you are about everything you're supposed to find somebody who looks at least to you like they know what they're talking about someone who's are relatively reliable looking source of authority on these kinds of matters and the internalist theory has a really nice account of why it is that's gullibility just believing whatever people say when you don't have any reasons for doing these things isn't the way of finding things out according to the internalist story you find things out because you have reasons for thinking that what people say are true in the case of gullibility you believe what someone says without any reasons for thinking that what they say it's true and this means that no case of gullibility is ever going to be justified by internalist standards because internalists say you need these reasons and gullibility is just where you lack these reasons so this is a nice point for internalist theories they can explain how it is that's the apparently obvious epistemic crime of gullibility is actually an epistemic crime they can see why gullibility is incompatible with forming justified beliefs okay let's leave internal ism just for the moment and think about another pair of cases ever saw think about a case in which first of all an individual looks at a thermometer and thermometer says it's 20 degrees and the individual comes to believe they've looked at the thermometer that it's 20 degrees that's the first of these cases the second and individual confronts another individual a speaker and after what the temperature is and the speaker says it's 20 degrees and the listener comes to believe that it's 20 degrees both of these it's worth being clear look like perfectly respectable ways of finding things out if you want to know the temperature you should either consult a thermometer or consult someone that you think might not on these kinds of things that seems fair enough in both of these cases it seems like there are epistemic ly good things going on someone in trying to find something out is doing what they're supposed to do or at least something that they could reasonably be thinking is what they're supposed to do okay so now we have two people doing what they're supposed to do but are they in all interesting respects we might wonder well one type of theory says yes they are reliable asphere ease theories that think what matters for whether or not you've successfully found something else is how reliably you formed your belief how easily you might have gone wrong in your inquiry very roughly and theories that think that's what justifies a listeners belief or what supports our listeners belief what makes a listeners belief somehow good is the fact that what believing that speaker is a reliable way of forming beliefs think that in the case where you rely on an instrument and you rely on a speaker epistemic ly speaking things are just to say it might be of course that one's more reliable than the other one but let's leave that detail aside for the moment let's suppose that the speaker and the instruments are equally reliable they're both likely to go rights as often as each other of course they won't go right for the same reasons the thermometer might go right because it's in some sense sensitive to the ambient temperature the speaker might go right because when they've looked at the thermometer or something like that but assuming they're both equally reliable equally likely to go right and wrong so say reliable as theories there's no difference in the kind of thing that's supporting the beliefs in each of these cases so the idea is that what happens here is then the first case an individual goes up to the thermometer and sees what it says and presumes that it will be reliable importantly they presume rather than figuring out for themselves that it will be reliable in the second case they go up to a speaker and presume that the speaker's saying something reliable when they tell them the temperature so the idea is that's there's a presumption here of reliability a presumptions important in this kind of case it's different to what the internalist thinks goes on in these cases according to the internalist the idea was that well in cases of testimony you tell me something and then I consider my background reasons and I think that well given that you said it yes it's likely to be true or something like that and I construct my own story of why it is that you're saying something means that it's likely to be true that kind of figuring out and constructing work doesn't go on on this kind of reliable a story that we're looking at here rather which kind of taken as a background assumption that the speaker or in this case the instrument is reliable so the idea is that in an important sense there's less work going on rather than needing to tell an entire story about these things and why the speaker or instrument should be reliable reliable theories of this kind just take it for granted that it is and the idea is that this comes to connect up beliefs to the reliability of the processes involved in the formation of the belief in the case of testimony that's believing a speaker's testimony okay so there's a reliable it story about these cases of believing instruments and believing speakers and it seems like the reliable story says we'll look in the important essential details these cases are the same you rely on something in one case it's an instrument in other cases a speaker and your beliefs supported by reliability of the instrument or the speaker the fine details come apart but the basic epistemic shape of things is more or less the same thinking about those two cases though some theories think that's well know that two cases aren't alike in an important sense or at least they don't have to be alike in any sense at least I don't have to be alike in the most important sense thinking about transmission theories in the epistemology of testimony these are a different type of theory all together their idea is that when a listener comes to believe a speaker the listeners belief can be supported by the fact that the speaker knows the speaker has something that supports her belief and that whatever it is comes to support the listeners belief so the idea is that if I see that the time is 4:30 and I tell you that the time is 4:30 I have some justification for thinking that it's 4:30 I saw it and you also have that same justification according to transmission theories at least that's something they think can happen and the idea for transmission theories is that there's an interesting kind of presumption going on here like reliable lists they're interested in presumptions rather than usually the idea that the listener has to figure out and tell their own story about why our speaker's likely to be reliable or in this case knowledgeable the idea is that for transmission theorists to listen and I presume that the speaker's knowledgeable rather than merely reliable they might think what this speaker knows what she's talking about as a background assumption and believe what she says based on that and in these cases we get the transmission of justification rather than a belief that's supported by reliability so this gives us our three accounts of how it is that beliefs based on testimony come to be justified how it is that we come to find things out according to one way we go around in the world the internalist way and look at people and figure out what they're saying true things based on our previous experience of people or based on our reasons for thinking they might be telling the truth versus lying to us or mistaken or speculating or any of those sorts of things the idea is that we we very much work it out for ourselves as listeners we find some speakers who tell us some things and then we work out with it's true for ourselves maybe using the fact that they said it but we do the working out according to the reliable Astari we walk around in the world relying on people and when they're reliably saying true things then we come to find out things about the world and according to transmission theories we we don't rely on each other so much but we take each other to be knowledgeable or presume each other to be knowledgeable maybe we trust each other something we'll come back to and the idea is that when we do that we come to know things or find things out about the world because the knowledge the speaker has or the fact the speaker's found out about it gets passed on to us as listeners so this gives us our three accounts of a sort and they say competing things when we have our three accounts it's worth seeing what might be said both for and against each of them so let's think about our first case again internal ISM that was driven in a sense by thinking about the kind of to dr. cases the idea was supposed to be that well you have this case in which someone thinks that their doctors doing a normal regular procedure on them and indeed they are another case in which it looks all the same to a different listener but the doctors not actually doing any kind of procedure that's any good and the internalist thought was that well these listeners are the same with respect to justification in any interesting sense these listeners are alike because they're aware of the same things though they have the same perspective on the world and this means that they were like with respect to the stuff that they found out or their justification for their beliefs that's supposed to be obvious says the internalist to my mind it's entirely far from obvious indeed I find it obvious that the reverse is true because one test is good and another test is not it strikes me as entirely obvious that one listener finds out in ways that the other one doesn't and that the interesting kind of thing to focus on here is what's actually going on rather than the two listeners and their perspectives on these things so I think we should be entirely cautious in treating this internalist observation that these cases are clearly the same as as obvious as it sounds or as obvious as the internalist wants it to sound of course the fact that to me the reverse is obvious isn't any more evidence that's the internalists wrong then their intuitions are that they're right but nonetheless even though I don't think that I have any particular god-given insight working out what the right thing to say about cases is I do think that there's enough there to cast doubt on the internalist story after all you might think if they later found out that they were in different situations they might then take it that this is evidence that they would never alike so suppose you're in the case where you find out that the test that the doctor was doing on you is in fact not a good one it seems to me like you might actually think well I thought I found something out about my condition but it turns out I hadn't it was all just a giant cop but note the way of thinking there it's not just I thought I found something out and now I haven't found something out it's that I thought I'd find something out and in fact I never had found anything else it was all just a con and in fact there's an external storage we told there means I think we should tread carefully with this kind of thing that's not decisive by any means but it's worth being clear that sometimes we might think the intuition that the internalist is offering here is one that could be negotiable the important thing for internalist is very much the idea that facts about the speaker whether or not speakers knowledgeable or reliable or doing a proper test or any of those sorts of things don't matter to a listeners epistemic state they don't determine whether or not a listeners found something else about the world unless the listeners aware of those facts so unless the listener knows that the speaker's reliable or is aware that the speaker's reliable the question of whether or not the speaker's reliable is a side issue for internalists it might be relevant to other things but it's not relevant to a question of whether or not the listeners found something else this gets brought out in a case from Jennifer Lackey who says look imagine your friend tells you that another person is a reliable authority with respect to birds and then the other person tells you something about birds but it's it's entirely unreliable and it's false in fact she tells you that she saw an albatross but she didn't see an albatross it seems like in this case you have very good reasons for believing her according to the internalist things are going well here after all your friend who you know and trust and have reasons to think a reliable told you that this was a reliable speaker but in fact she's not and Lackey thinks that in this kind of case we really shouldn't think that the listeners found anything out that's interesting there's an important sense in which the fact that speaker doesn't know what she's talking about and is unreliable matters for a listener whether or not they're aware of it and if these observations are right it looks like there's more going on in the case of testimony then the internalist can allow for it looks like there's more going on than just the individuals reasons whether or not you found something else doesn't just depend on whether you have good reasons for thinking that you found something else of course this opens a question of what should fill this kind of gap but nonetheless it seems that even when things are going well by internalist standards there might be reasons for thinking that on balance things are not going well we might think that there are more ways of things going right and wrong than the internalist has written to accommodate so there's reliabilism the idea that beliefs were justified because of the reliability of the speaker fare any better well we might think that in an important sense reliable ist's right people can rely on each other in the way that they rely on instruments but we might think that's not all there is in the epistemology of testimony we don't just rely on each other in the way that we rely on instruments we take it but speakers know things in ways that instruments don't ok sometimes we say that instruments do know things we say that the supermarket doors know when there's a person in front of them and they open appropriately but when we talk about them knowing there's an important sense in which we're not talking about the same thing as when we're talking about one of us and knowing a person knowing or something like that it seems like we take different attitudes towards people as well we can make presumptions about people that we can't about instruments and back to the literature on trust that I mentioned briefly earlier it seems like when people go wrong we blame them and feel resentful towards them in ways that we don't with instruments if your car breaks down well you might feel resentful towards it in a sense but you don't feel the same kind of resentment and betrayal if you did if someone offered to give you a lift but then didn't turn up or something like that so we might think that there's a worry here for reliabilism and the idea is that reliabilism is just too unsubtle and too narrow with its weight account of the ways in which we treat each other it says that look we just go around relying on each other and that's all is to it our beliefs are connected to reliability because we rely on people and make presumptions about their reliability but we might think there's more going on than that we actually make presumptions about people's knowledge ability and that's not just relying on them that's not the kind of attitude we take towards a thermometer say we take it that the person knows that it's 20 degrees not just that they're relying on someone to believe that it's 20 degrees at the right time and then say the right thing so we might think that if the reason that our beliefs are connected to reliability is because we make a presumption of reliability reliabilism can't give us the full picture here it's not all about reliability we might think this was where transmission theorists come in so do transmission theories get on any better can we say okay well it's not all about reliability it is about finding knowledgeable people presuming them to be reliable and picking up their knowledge by transmission or their justification for what they say by transmission this of course does catch the idea that we make presumptions about people that we don't about instruments on the other hand and that much is fine but we might think there are also worries about transmission think about a case where we have someone who is systematically confused about the world they think that all triangular things are square but they also lie all the time so we're never there thinking that something's Square they'll tell you that it's triangular and the idea is that this person will reliably say true things they'll reliably tell you that things are triangular when they are in fact triangular it's just that they'll do it in a weird way they'll see something triangular believe that it's square and tell you that it's triangular assuming that all of these processes are reliable we have reliable testimony and reliable testimony sounds like a good thing sounds like finding someone that's reliable should be a way of finding things out the trouble is it's not the way of finding things out the transmission terrorists can make any sense of because in a case where someone sees something triangular believes its square and tells you it's triangular importantly they don't know what they're telling you their justification for what they're telling you is non-existent I mean they think the objects square so the idea is that in this kind of a case we can't say a speaker's justification for what she says is transmitted to a listener it's not the speaker doesn't have any justification for what she says and it's entirely obvious that where a speaker doesn't have any justification for what she says a listener can't get any justification by transmission so there's a problem here for transmission theories transmission theories on the face of it should want to say that someone who reliably says true things believing them is a good way of finding things out believing reliable speakers is a way of finding things out about the world but this kind of reliability can't we put in terms of transmission so we have a worry here for transmission theorists okay so let's take a step back for the moment before we head towards what I'd like to offer as a solution we have three theories we have internal ISM we have reliable ISM and we have transmission and we have objections to all of them sadly it seems like they all fail in their own particular ways so I'm thinking for a moment about what we want a theory of testimony to do and I think we want it to do two things firstly and most obviously we want it to say the right things about which beliefs are justified in which beliefs are not in which cases does someone find something out and in which cases doesn't someone find something out but we don't just want that we don't just want something that will tell us which cases and knowing and which cases are not knowing in which cases are justified in which cases are not justified which cases are finding things out in which cases are not finding things out we want to find some principled way of doing this we want to find some way of explaining this that brings all these things together it isn't just pointing at this case and going that seems justified and that doesn't we want something that all in some way help us to understand what's going on what's in common between all of these things and with those two adequacy conditions on a theory in mind I'd like to edge towards the idea that we might combine them somehow so according to the kind of new I'd like to give the idea is that there's there's truth in all of these theories the internalist theory goes right in important ways the reliability theory goes right in important ways and the transmission theory goes right in important ways each of these three theories is indispensable to a complete account of testimony we might think and the idea is to try and bring them all together so think about the case that was a problem for internalist theories I mean we might think it was the dr. cases or we might think it was just a case where you have some reasons but on the face of it things don't seem to be going as well as you thought they were insofar as that's possible that the internalist theory could be wrong because things could go well by internalist lights but yet you still fail to find things out this only shows you that internal ISM is incomplete as an account I mean it doesn't tell you that's the internalist reasons are no part of the story just shows you that they're not all of the story it shows you that it's clearly more to what's going on here than just the things internalist theories are interested in likewise the idea for reliable Asteras was that there are particular presumptions that we make about each other that are not presumptions of reliability the idea was that there are presumptions that we make concerning the idea that people know things and reliable theories can't catch this idea because they think that we're connected up to reliability because we rely on each other that much is fine we might think and there's a problem for reliable as theories but it doesn't show that we never rely on each other I mean the fact that we sometimes make presumptions about knowledge ability doesn't show that we never just make a presumption about reliability we can and we might think just rely on people in the way that we rely on instruments and it's hard to see what might be wrong with that so the problem with reliabilism at least as I was offering was not so much that it can't capture anything that's right it wasn't it can't capture everything that's right like the internalist theory it falls short because it fails to give us a complete account moving on to transmission theorists well transmission theorists had a problem in that case where someone reliably searched through things but doesn't know anything at all because they don't even have the right beliefs but that doesn't show us that transmission is never the right thing to think about a case that just shows us again that it's not always the right thing to think about a case and the idea here is that by appealing to the idea that there might be truth in each of these views we might be able to use each of these views to plug the gaps that the others miss so the idea is that the kind of case where transmission theorists struggled will be explained nicely in terms of reliability and the kind of cases where a reliability struggle can be explained in terms of transmission but if we want to go for this kind of approach an approach that brings together all of these three things we need an account of when is it that each of these things are relevant sometimes people will have reasons but to their reasons matter sometimes a speaker will be reliable but does that matter other times there'll be a knowledgeable speaker but is that important here why and when or each of these considerations relevant and this is important because this is the bit where we start getting past the idea of just pointing at things and saying oh that's justified for that reason and that's justified for that reason and that's justified for that reason what we need here is the principle bit any theory could tell us which cases were justified in which cases were not the hard bit you might think is the principle bit a bit where we try and work out exactly what's relevant when the way that I suggest we might try and do this is by identifying different ways of responding to what a speaker says I mean at the core of all of these theories appears to me to be different ways of thinking about how we respond to people so according to internalists we respond to people at least we're supposed to respond to people by thinking about what they say and reflecting on the reasons that we have for thinking that what they say is true and then coming to believe them or not believe them according to whether or not we have enough reasons for thinking that they're saying something likely to be true according to reliable astaire is we rely on each other by making presumptions about reliability in the way that we do about instruments and according to transmission theories we presume other people to be knowledgeable or speaking from a position of justified belief so the idea here is that we might exploit these kind of conversational differences differences in the way that we respond to people to underpin epistemic differences and we might say that well when you make the kind of presumption of knowledge ability what's relevant to your justification is the question of whether or not the speaker is knowledgeable we have an explanation in terms of transmission but on the other hand if you don't make that presumption if you presume that the speak is reliable what's relevant to your justification is the reliability of the processes involved in believing that speaker whatever they may be and lastly if you decide well I'm not going to presume that the speaker's knowledgeable or reliable but I'm going to presume nothing I'm gonna do the figuring out for myself I'll take my beliefs in my own hands and I'll use my reasons to justify these beliefs well these are when beliefs are justified or otherwise by internalist standards and the internalized explanation goes right so I say the idea here is that we can get all three parts of a theory and we can fit them together based on the idea that what determines your justification is a listener is how you respond to the speaker and in doing that I think we can sort out how there's truth in each of these views and fit them together in a way that isn't just saying well sometimes it's reliability and sometimes it's transmission and sometimes its internal ISM the idea is that it's reliability in these cases these cases where someone is relying on someone else its internal ISM in these cases where a listeners taking her belief in our own hands and figuring things out for herself and it's transmission where the speaker's presumed to be knowledgeable so to wrap things up it's worth bringing ourselves back to that initial thought that we had that the point of a theory of testimony is to explain as a justified belief as possible I mean I started out by pointing out that look really quite a lot of our beliefs seem to depend on testimony in some sense and that if testimony is off the table as a way of finding things out we really don't know very much at all since that was how I decided to motivate these things you might think an important part of a theory needs to be explaining quite a lot of this stuff it's no use saying okay well loads of my beliefs depend on testimony but I'm only going to explain this really small bit that wouldn't make for a very interesting theory nonetheless I think that the theory here can do the kind of explanatory work I set out for it it can explain how it is the lot of our beliefs depend on it can explain how there are really diverse ways of believing what people say and the idea is that in appealing to these kinds of diverse ways of believing each other we can find a theory of testimony that will explain a lot of our beliefs it will bring together different types of explanations based on testimony and it'll do that in a way that finds a common theme to them all it'll appeal to conversational differences and use those to support epistemic differences the thought is that if we take all of these things will somehow get ourselves towards a complete epistemology of testimony