An Interview with Herman Cappelen

PDF tilgjengelig

———-Illustrasjon: Oda Aurora Norlund———-

Herman Cappelen is a professor of philosophy at the University of Oslo with a part-time position at the University of St. Andrews. He has published influential books and papers on many topics, especially in the philosophy of language but also philosophical methodology, epistemology, metaphysics, and philosophy of mind. In particular, he has worked on the topic of relativism about truth, and, together with John Hawthorne, he wrote the book Relativism and Monadic Truth (2009). The book presents and argues against relativism about truth, while maintaining that truth is a monadic property: if something is true, it is true full stop. In this interview, Cappelen discusses his understanding of relativism about truth and the arguments for and against the view, together with how the debate relates to other questions in philosophy.

Interview

In general, what is relativism? And what are some examples of specific kinds of relativism we can find in contemporary philosophy?

I think it is important to distinguish two ways the term relativism is used in philosophy, including contemporary philosophy. The historical use of it had to do primarily with a very vague thought that is not as common today. The vague thought is that in some domain, or maybe very generally, truth is relativized to some kind of parameter. The easiest way to think about that is in the moral or normative domain: the claim something ought to be done is true or false only relative to something, e.g., your background, or your community, or your choices, or something like that.

This vague thought then gets cashed out or explained in two different ways in more contemporary work, and the weird thing that happens is that now one of the things that was called relativism in the past turns out to be the opponent of contemporary relativism. One way to spell out what I just said about relativization is that you say “well, when I say that it’s good to φ what I’m really saying is that it’s good for me to φ relative to…” and then you put in the parameter that you relativize to, into the claim made. And so when I say, just to make things very simple, that it’s right to φ, what I’m really saying is that it is right for Herman to φ, and then what you are saying is that it’s right for you to φ, and so on. So people make these claims that have the relativization built into it. That’s still a form of relativism in the old-fashioned traditional sense, since it is still going to end up with there being, in some sense, no objective truths about what you ought to do. There is the truth for Herman and there is a truth for other people, and those things can differ.

This move, where you build the relativization into the content of the claim, is what at least many of us today call contextualism. The relativization doesn’t have to do with truth itself it just has to do with the content of what you said. So, when I tried to explain this right now I said: what you said is that it is the right thing for you to do, what I said is that it is the right thing for me to do. So, the relativization becomes a part of the content. What you have done isn’t to fiddle with truth you just fiddled with the content of what is said. Now, in the past people used the term “relativism” to cover those kinds of views. Gilbert Harman, for example, in defending versions of relativism, would talk as if that was relativism. And there are people today who still talk in that old-fashioned way, so David Velleman published a book on relativism and he uses “relativism” in the way that I just described, where you are fiddling with the content of the claim, not anything having to do with truth. That is one way to use “relativism”, and it might be the way that it was used throughout much of its history.

But, then, today I think we’ve made some very significant progress, in that we’ve distinguished the view that I just described from a very different view, where it is not the content of what you say that has the relativization built into it, but the truth-evaluation of what is said. On this view, when you say that it is right to φ and I say that it is right to φ, we’re saying the same thing. You didn’t say that it was right for you and I didn’t say it was right for me. What was said was just the same thing, so in a sense we agree. If you say it is right to φ and I say it is not right to φ, then we disagree, because you’ve affirmed something that I denied. Recall that on the previous view you wouldn’t have said something that I denied, right, because you would’ve have just said that for you it is right to φ, and I would’ve said that for me it is not right to φ, and those are perfectly compatible. But in this new way, that I think of as the contemporary way, of using the term “relativism”, you and I have disagreed, because you affirmed what I denied.

So, where does the relativism come in? Well, that comes in at the level of how we evaluate claims as true or false, not in how we individuate claims. On this view, when you evaluate the one thought, or content, or proposition (people use different terms) “it is right to φ” or “one should φ”, you can say it’s true, I can say it’s false, and we can both be right. When I have been writing about relativism I’ve taken that to be the relevant sense of the term. It is a form of relativization that doesn’t build the parameter-relativity into the content but builds into the truth-assessment, whatever you think that is. I have now talked about normative claims, but, of course, you can be relativist in any number of domains, you could think that, to be super extreme, mathematical claims are only true or false relative to a certain type of parameter, or you could think the same for claims about knowledge, that it is relative whether someone knows something or not, and you could go through different domains and see where relativism applies and where it doesn’t. And, of course, the limit of that is you could be a global relativist where everything is relativistic.

Just one more thing about how different those two initial ways of using the term relativism are, the one that builds it into the content and the one that makes it to be about the truth-evaluation: You could be a relativist in the old-fashioned sense and build it into the content, so that when you utter the sentence “it is good to φ” or “you ought to φ”, then what you really said is that you ought to φ, you just talked about yourself. The person who denies the second kind of relativism could agree with that, and just say “yes, you built that into the content but the truth-assessment is now objective and universal”. So, the two views are very different, and the paradoxical and extremely unfortunate way about mixing these two ways of using “relativism” is that old-fashioned relativism is now sort of understood as the alternative, the opposing view, to contemporary relativism. So, the terminology gets confusing quickly. If you want to get into contemporary debates, the way to do it is to think in the second way where you are not fiddling with the content expressed, but just with how you assess truth and falsity.

You defended, together with John Hawthorne, in your book Relativism and Monadic Truth a non-relativistic understanding of truth. You call it the simple view where truth is a monadic property. Could you describe that view?

The simple description of that simple view is just that it is the denial of the second kind of relativism, let’s just call that relativism from now on. It is the view that when you assess something as true or false it is simply true or false. Another way to express the view is: it is true or false simpliciter, or it is true or false full stop. All those little extras at the end are just supposed to remind you that there isn’t anything more. It’s just true or false and there is no relativization.

In the book I wrote with John Hawthorne, we talked about this little package of views that we thought went well together: when you speak you express something we call propositions, or contents, and they are also the contents of beliefs. Propositions, then, serve two roles initially, they are the content of sentences and the sentences express what you believe. And those sentences are monadically true or false. This we thought was a package of views that go well together. They are, anyway, the traditional picture, we think.

How should we think about assessing for truth and falsity on this simple view?

The truth-predicate just applies to something in the following simple way: if you have a content or a proposition, it’s either true full stop or false full stop. Then there is this activity of trying to figure out which one it is, and, of course, in that activity we’ll be engaged in all kinds of complicated things and we’ll disagree and so on. Whether you will end up agreeing with me about whether it’s true or false will depend upon all sorts of things about you and all sorts of things about me. But the point is that that doesn’t affect whether it’s true or false. These activities of assessing are not constitutive of the property of being true or false, there is a super-important disconnect.

You have described the opposing view, relativism in the contemporary sense, where you have some content and the truth-value can vary due to some parameter. In the book you describe in more detail what you think the best version of that view is. Could you say something more about what you take relativism about truth to be? And what arguments people give in support of that kind of relativism?

It might help to give a little bit of history. The thought that relativism, in this contemporary sense, is true, had not been very popular among those thinking about truth and content and those kinds of things. The view just hadn’t been worked out very much. What had been worked out reasonably well was the thing that is now called contextualism, where you build it into the content, which had been worked out in all kinds of ways. The idea that you just have one content but the actual truth-value was relative to some kind of parameter, that view hadn’t been very well worked out. And then there was some, I think, groundbreaking work done: by Max Kölbel, who wrote a book, Peter Lasersohn, who wrote some papers, and then somewhat strange historically, a paper by my co-author John Hawthorne, Andy Egan and Brian Weatherson – maybe one of the first papers that tried to articulate this relativistic position in more detail. Then after that the person who ended up, I think, getting a lot of the credit for relativism and developing it throughout many papers and in a book was John MacFarlane.

The simple version of the view that they actually articulated was that, with respect to some particular terms – the examples they often went back to had to do with a certain kind of might-claims, “it might be the case that…”, they called it epistemic modals – they tried to find areas where they thought this kind of relativism is plausible and argue that there is evidence for it, even. Another case is what they call predicates of personal taste. An easy way to think about it might be something like “it is funny”, like in, “that movie was funny”. And the achievement, if you can call it that, of this tradition was to first develop a formal framework that included a truth-predicate that wasn’t monadic but relativized in a relevant sense. Then the hard work, so to speak, was to articulate and describe the way that framework explained a whole bunch of phenomena. The view itself, if you just put it without the thing that it’s supposed to do, is just: you have a formal system where all attributions of truth or falsity are indexed to what MacFarlane calls a context of assessment. So that is the view, and then the next question is why would you want to do that?

The driving idea, the core simple idea, that anyone can understand, is that if you say “that movie was fun” and I say “no, it wasn’t fun”, there is a very strong sense that we have disagreed with each other, and you have said something that I have denied. That’s data point number one. It says we disagreed, so you want to explain the sense that we disagreed, that there is a genuine disagreement, and, of course, if you had just said it was fun for you and I had said it is fun for me, then it looks like we haven’t had a disagreement. But if it’s just this proposition, “it was fun”, and you assert it, I deny it, that looks like a genuine disagreement.     But on the other hand there is a sense that you haven’t done anything wrong, and I haven’t done anything wrong. The second desideratum, then, is to respect the intuition that there is some kind of subjectivity in this domain, there isn’t an objective truth about what’s fun, that all depends on your sense of humor and so on. What they did to respect that was to say “well, we capture the disagreement by letting the content be these non-relativized things”, so that’s the disagreement bit, but then from your context of assessment there will be one standard of humor and from mine there will be another, and so since truth is always relativized in that way you get to be right from your point of view and I get to be right from my point of view, even though you say it’s true and I say it’s false.

So, they provided a structure for saying that we disagree but we can both be right. Some people, like MacFarlane, Max Kölbel, and Peter Lasersohn, likes to describe it as a form of faultless disagreement. There is disagreement but the disagreement involves no fault on behalf of one or the other participants. If you are looking for arguments, that’s argument number one, that’s like the data-driven argument.

Then there is another argument that Hawthorne and I talk quite a bit about, there is a whole chapter devoted to it, and it’s a bit more technical and a bit harder to get people to see. It’s an argument that somehow comes from David Lewis, it is found in some of the work of Jeff King, and you can find it in parts of MacFarlane’s writings, though he downplayed it a little bit when he published his book. Well, the way we describe it, it has twelve different premises and a conclusion, so I don’t think it would be very suitable for this interview, and the way Jeff King does it, it’s also super complicated and long. But to give just the spirit of it: In almost all formal systems for languages in formal semantics theorists tend to relativize truth to some parameter or other, so it actually looks like some form of relativism in these formal systems is the standard view. David Kaplan, for example, and this is a very important precedence for it, says, well, truth and falsity is relative to a world, it is true in this world but it could have been false, so it is false relative to another possible world. So people seem comfortable thinking that truth or falsity is relative to a possible world. And many others are comfortable with the idea that you relativize to times, a proposition could be true at one time and false at another, and Kaplan even included places as parameters. So you don’t have simply truth or falsity. But this was sort of independent of the original motivations of relativism, they were just formal moves that were made. And then MacFarlane, in particular, used to say “hey, so what’s weird about including standards of taste, or a sense of humor, or some body of evidence”, so you just add a parameter to something we’re completely comfortable with having parameters with respect to anyway.

So, that is two arguments that support relativism: the case of faultless disagreement and the fact that a lot formal theorizing in linguistics and logic seems to have added this relativity anyway by having parameters when you assess for truth-value. How do you respond to those arguments?

So, Hawthorne and I, in that book, we say, well, first it was a mistake to accept all of those other relativizations. Truth is monadic, across the board. When we talk about truth relative to a world that is a derivative notion, the basic notion is the notion of truth simpliciter. It was a mistake to include relativization to times and it was a mistake to include relativization to places. Now, that’s hard work, because now you have to show that you don’t need it in the case of modality, talking about what is possible and necessary, you don’t need it with respect to time, and you don’t need it with respect to place. And so we do a bit of that work throughout the book, showing how in each of those cases this relativized notion really is derivative and that the basic notion is a monadic one. So, it was hard work writing that book because you had to say something about modal logic and modality, say something about tense, you had to talk about all these different areas in which people have made relativizations and say, you know, that was a useful theoretical tool but it doesn’t cut at what is fundamental, it doesn’t cut at the basic structure of language and thought. That was one strategy of replies, go after that “look, we’re doing it many places already, so why not add them” and reply “no, you shouldn’t have gone down that road”. Basically, you misinterpret people if you go down that road.

As the second response to what I described as the first argument, the one from faultless disagreement, we tried to show that you can generate that sense of disagreement and the sense of faultlessness, without going relativistic. You could do that in two kinds of ways: you can explain those intuitions in better ways and you can show that the relativist predicts things that aren’t real, that relativism overgenerates phenomena, predicting that there should be phenomena that don’t really exist. In particular, we say there isn’t always that sense of disagreement, and we give a bunch of cases where one person says “this is fun” and another person says “that is not fun” and there is no sense of disagreement. If you think about very weird cases, like talking animals, it is very weird, when you realize how totally different from us they are, to think that there is a deep sense of disagreement. But the relativist would get us to think “no, there are these genuine deep disagreements in all these cases” and we show that, typically, that isn’t the case. And in the cases in which there is a sense of disagreement, there are many ways for the non-relativist to explain that. A natural case to think about is standards. You build the standards into the content, not into the truth-assessment, now, when you are talking about what’s funny and I talk about what’s funny, we try to generate a kind of common standard, and part of what we’re disagreeing about is what is funny or not relative to that communal standard. That’s a sense of disagreement but it doesn’t require that there are two separate truths – it’s in fact an effort to coordinate.

Since we wrote that book, which was quite a few years ago now, this literature has continued and it is a hard literature to get into. There are now literally hundreds of dissertations and papers written on little sub-parts of each of these issues. That’s great. This way it becomes more sophisticated. Through collective effort we now know massively more about of how to defend relativism and how to argue against it than we ever did in the history of philosophy. Which I think is a sign that we’re making incredible progress very, very fast. But it also means that if you were to try to get into to this now, it would take years and years of work just to look at all the explanatory models.

Have you seen any work defending relativism within a domain that you think is more persuasive than other work?

No.

So, still a global anti-relativist?

Well, I like the arguments we have in the book, they are pretty good arguments. I mean, the way I work I think about something for many years and then I write a book about it. Then we wrote, I think, ten replies to different leading relativists who were replying to us. And as John and I were writing up those replies, none of that made us change our minds. And then, I felt like I’ve made enough of a contribution to that field and I started working on something else, I think after that I started worrying about intuitions, and I kind of left studying relativism-topics, not behind, really, because I have students working on it and so on, but… yeah, maybe I’ll go back to it at some point and see what people have done.

You mentioned earlier contextualism about meaning.

Yeah, the parameter you want to relativize to gets built into the content of what you say. So, when you say “one ought to φ” what you’re saying is relative to your standards, and that’s actually part of the content. You didn’t say it out loud but it’s sort of hidden in the content there, the thing you asserted, the proposition expressed, has a reference to your standards in it. Or, if you say “it’s fun”, you said that by your standards it’s fun. So, let’s just try to speak in that way: I say “by Herman’s standards this is fun”. Now, that could be true for absolutely everyone, everywhere. It’s perfectly compatible for that sentence to be monadically true. You could say Herman expressed some proposition and it was the proposition that relativized the funness to his standards but that relativized claim itself is non-relatively true, That’s going to be true for you even though you disagree, you would say “it’s not fun”, because then you’re saying that by your standards.

Kind of the way that some traditional logicians wanted to deal with tense and place, for example, you would look to a fully specified proposition. Then when we utter something like “it’s raining” we’re really saying it is raining at that place at that date at that time, and that gives you a proposition that is invariantly true or false.

Exactly.

So, this view is compatible with the monadic understanding of truth. You have earlier argued against contextualism about meaning and some of those arguments that you used against contextualism can be used in support of relativism about truth. Could you say something more about the relationship between these two strands in you thinking?

So, the background is: I was right out of graduate school, it was a long time ago, and in the early 2000s, so almost twenty years ago, I wrote a book with Ernie Lepore called Insensitive Semantics. That book is an effort to argue in favor of something we call minimalism about semantic content. So, there we were in favor of a semantics that didn’t include much context sensitivity. However, in that book we also argued for the view that we need a notion of what was said that is very rich. We argued that semantics doesn’t exhaust what is said. There are many things said and one little part of that is the semantic content and that part doesn’t have all these relativizations built into it. So, what we argued against was a kind of contextualism about semantic content, not against contextualism about what was said, we’re in favor of contextualism about what was said.

Now, it should be said that there are some very interesting connections here, the way I see it. For example, a lot of MacFarlane’s early work just took the arguments from Insensitive Semantics – I mean, he didn’t steal them but he used the same kinds of arguments – and used them as a theory in favor of relativism about truth. So what he did with truth, to let truth vary with assessors, we did all that work with having what was said be much richer. That’s the history of it. Just after Insensitive Semantics came out MacFarlane published a reply where he sort of said “no, you guys really should have been relativists about truth”, and that’s an early MacFarlane paper where he says all these arguments are great arguments for relativism about truth. I still think that much of what MacFarlane wants to do with the relativization of the truth-predicate, can be done by being more pluralistic and rich about the notion of what was said.

I guess the distinctive thing about your view is that you accept a pluralism where, when you utter something, a lot of propositions are put into play at the same time, not just one.

Right. Another, even more radical part of my view is that one sentence can express different propositions for different people. So the view I have is that I utter a sentence, the sentence will express many propositions, one of them will be the semantic content. Relative to you the cluster of propositions could be C and relative to, say, Bjørn Ramberg it could be C2, and C and C2 need not be the same cluster. So, what I think is that I say each of the things in your cluster, so relative to you I will have said something that I didn’t say relative to Bjørn. This gets very tricky, I know it sounds very relativistic but it isn’t. You could actually correctly, truly, say that Herman said something true, Bjørn could say that Herman said something false, but that’s all compatible with monadic truth because one of the things I said relative to Bjørn is false but one of the things I said relative to you is true. If you have this picture, where there’s a cluster of content, you get something that, again, sounds a little bit relativistic. But I’m not worried about that because it doesn’t make the truth-predicate relative, it’s just a consequence of how what I said will depend in part on the interpreter. So, most of what is contextual I like to build this kind of relativization into what was said, say, what Herman said relative to an interpreter. Again, that is compatible with a monadic truth-predicate, because it only relativizes the saying-relation. This is what I call content relativism – and that’s a form of relativism that I endorse. It’s not about truth, but about content, i.e. about what is said.

Recently, you’ve been working on the topic of conceptual engineering, how concepts change and maybe improve. In some discussions of conceptual engineering, people talk of some kind of “relativism”, where depending on the concept we’re using, or the version of a concept we’re using, the truth-value of a claim might differ. So, one example is “fish”. Say that 400-500 years ago people just called anything living in the sea “fish”, so a whale would be a fish, but then on the modern understanding of the word there are much more stringent criteria, and the whale would not be counted standardly as a fish but as a mammal. The question is whether or not it was true that whale were fish when we had this concept and now it’s false once we’ve changed the concept? What do you think of this kind of seeming relativism?

It is important to keep track of what we mean by relativism here. This sort of phenomenon is not in any way related to the relativism that I talked about earlier. Here’s something that could happen quite easily, and, I think, happens a lot: You mean one thing by “fish” and then you utter the sentence “whales are fish”. By that you express a certain proposition, say, that the whale is an animal that swims in the ocean. That’s monadically true or false. Then I have a different meaning for “fish”, where it excludes mammals, for example, and then I say “whales are fish”. I will be expressing a different proposition from you, and mine might be false while yours is true. But given the way we set up things earlier that just means that we expressed different contents, it is the contents that have changed. Now, that’s the answer to the initial question.

Then there is a whole cluster of complications that look kind of relativistic, but if I have thought my way through it properly, they are really just versions of this content relativism that I’ve just described to you. Let use the example of “fish” again, but let’s make it a little bit different: At some point in the past people used “fish” in such a way that that little thing, one little thing, call that thing A, was a fish. Then I want to say that, well, concepts can change over time, so that things that once was in, is now out. Now we go a little bit further into the future, and A is no longer correctly described as a fish. Now, I just described why, so far, there is no form of relativism here. But there is a problem because I also want the following to be true: so I’m the person speaking now, I want to be able to say what the person in the past said and I want to do it, what we call, homophonically, I want to use the same word, that is, I want a kind of continuity of topic. So, I want the following to be true: that when I utter “you said that A is a fish” I’ve said something true. But, what I say when I say “A is a fish” is false, and when you said “A is a fish” in the past it was true, but at the same time I said what you said when I say “you said that A is a fish”. So, now it looks like we’ve both said something true and said something false relative to different times, but I don’t want that kind of relativism. What I really think is that what has happened is that what you said has changed over time. So, it’s a form of content relativism. These are complicated issues, they’re very fuzzy.

So, you want the content of the assertion of the original speaker to have changed at the subsequent time?

Yeah, but I also want it to be true that I can say what you said using the same sentence, the sentence that is now changed in meaning.

OK, because there is no relativism about truth if the concept change, but still we might want to say, at least many wants, to say that it is false that, for example, A is a fish.

It is false given what I mean by it. At the same time, it is also true for me to say that you said that A is a fish. But I know that you meant something different by it, when you said it, it was true. So, there is a clash. What you said was true because of what you meant but at the same time you said that A is a fish and that’s false. A lot of work needs to be done to resolve that tension. But it’s not a view that is the kind of relativism we talked about at the beginning.