Ruth G. Millikan, “Beyond Concepts: Unicepts, Language, and Natural Information” (Oxford UP, 2018)

I love wearing clothing that's both comfortable and elevated. Outfits I can wear on a walk through the park or in a meeting with a client. Quints has become my go-to, with fabrics that are incredibly soft, clean, and versatile. This spring, I refresh my wardrobe with quints. I especially love their Pima cotton tees and bamboo jersey lounge shorts. Surprisingly soft and breathable, with a quality level you'd expect to pay a lot more for. If you're looking for new clothes this spring, I highly recommend checking out their Italian swim trunks. I love swimming, but can never find swimwear that feels comfortable and looks good. Quints' swimwear is the best I've ever owned. I can't emphasize enough how affordable quints is for the quality you get. Check out their incredible deals and offerings, especially if you're looking for clothes that feel good and look great, whether you're at the office or the beach. Refresh your everyday with luxury you'll actually use. Head to quints.com/newbooks for free shipping on your order and 365 day returns. That's QUINCE.com/newbooks for free shipping and 365 day returns. Quints.com/newbooks Hello and welcome to New Books and Philosophy, a podcast channel with the New Books Network. I'm Carrie Figdor, Associate Professor of Philosophy at the University of Iowa. I'm co-host of the channel along with Robert Toulise, Professor of Philosophy at Vanderbilt University. Together we bring you conversations with philosophers about their new books, drawing from a wide range of areas of contemporary philosophical inquiry. Today's interview is with Ruth Garrett Millican, Distinguished Professor Emerita at the University of Connecticut. Her new book, Beyond Concepts, Uniceps Language and Natural Information, is just out from Oxford University Press. Kant famously asked the question, "How is knowledge possible?" In her new book, Millican responds to this question from her characteristic naturalistic and specifically evolutionary perspective. Millican has long been a leading figure in theorizing about language and thought. Her latest work considers the clumpy world that organisms confront and the problem of how we recognize the same objects and properties again, as well as their kinds and categories. Our cognizing machinery includes unit trackers, whose job it is to track these atoms and channel information of the same item to one place, which she calls a unicep. And while each of us has distinct unit trackers and uniceps, they can be attached to the same word in a public language and words themselves are seen as lineages of reproduced signs. There's plenty to talk about in the book, and there's plenty we talked about in the interview. So let's turn to the interview. Hello, Ruth Millican. Welcome to New Books and Philosophy. Hello, Carrie. Nice to be here. I'm very honored and pleased to be talking to you about your or with you about your new book beyond concepts. Before we get to the book itself, I thought, you know, because many of our listeners will be to some extent, you know, certainly familiar with with who you are, to some extent also familiar with your work, at least in broad outline. We like to start with a question about, you know, the interviewer themselves about, you know, how they got to philosophy and how they came to write this book. And, you know, despite the fact that many people will be familiar with you and with your work, I think it would still be an interesting way to begin to get some idea of, you know, how you became attracted to philosophy and how you came to work on the particular topics that you have. Well, I guess I didn't even know there was such a thing as philosophy until I went to college. I went to Oberlin College. And there it was required. You had to have a course in philosophy. And luckily, I took it in my sophomore year, because then I could major in it. If I'd taken it later, I probably couldn't have changed. But it was the teacher I had, I think, who really brought me into philosophy. His name was Paul Schmidt. He had, he was a Yale graduate. And he ran his classes completely by a psychotic method, so that you were brought into thinking about problems, ancient problems and modern problems immediately. And he always made you feel as though you actually might come up with the answer. Anyway, I was absolutely fascinated by the topics I had never heard anybody talk about topics like that before. So I actually skipped the infirmary one, jumped out of the infirmary one day, snuck out when I was sick to code it was class. I was so fascinated. So, but so I majored in philosophy. But then I was a couple of years, for a couple of years after, I was just doing various things. You should teach you a little kiss to believe I held in mountain Berkeley. Things like that, taking photographs. Anyway, my friend Ernie Hook, his father was Sydney Hook. Or you might have heard of is a very, very well-known logical philosopher. Anyway, Ernie and I had taken logic together. We have the only two people that took symbolic together at Oberlin. And so we knew each other quite well. And he was out in Berkeley. And he just simply arrived at my house one day with his portable typewriter under his arm and an application blank for Yale. He said, "You're going to apply to Yale Graduate School." So he wrote it all down and I applied to Yale Graduate School. And so that's how I began. And then you came to be writing about philosophy of mind. I mean, you changed the field in a way. I mean, in a fundamental way. When I was at Yale, I thought I was going to write a dissertation on what I called attitudes or what attitudes are, which are, which are going to be related, I suppose, to work people doing all any motions. And somehow or other, it turned out that I had to think about some more basic kind of epistemological problems before I can even think about that. So I've never got back to the thesis problem. I'm still trying to do the underpinnings. So I mean, you mentioned in the book towards the beginning that the broad theme is the contiant question of how knowledge is possible. And you approach it as many listeners will know from a naturalistic and specifically evolutionary perspective on how we obtain and store and generate concepts. Or in this case, uniceps, we'll get into that in a second. But you start the book with a certain vision of what the world is like that. We think about that. We may have knowledge about. And you describe it as a clumpy or messy world that presents the organism with a lot of problem distinguishing which clumps kind of cohere, which are different clumps from other clumps and which are the same clump over time and from different perspectives. So can you say something about the world that, as you see, that we are, that the organism that wants to know the world, what that world is like, that they're confronting in terms of the clumps and then the categories and kinds that are out there as well. So the thought is that if the world are not structured the way it is, we really couldn't have such a thing as thought. That thought is dependent upon the fact that the world organizes itself in a certain way. And the basic idea here is that the world organizes itself into clumps. We can think first of individuals. Clumps of individuals that are like one another in a whole lot of different ways. Many of them, then some of them will not be, you know, they will have. They won't all have exactly the same characteristic, but so they won't be a definition for them, but they will have similar overlapping characteristics. And there will be a reason for this. The main, this is very important that there will be a reason for this. So the paradigm, example, I would be a species. So one dog is like another dog for a good reason, right? There are causal reasons that connecting one dog with another. Anyway, the fact that the world comes in clips like this means that it's possible to use induction. So if you discover that there's a clumped dog, then it becomes a reasonable to suppose that a characteristic you see in one dog may well show up in another. So I think, for instance, of a veterinarian, where they're investigating, to think about dogs, they experiment on two or three dogs, and they figure that this might be true of all dogs. If it's, so it's, you know, we were constantly generalizing really from just a few examples. And the only way that we could do this is because the world comes in, comes in kind of clumps like that. And so the spaces between these clumps, there isn't really anything. If you tried to sort of think of possible creatures running between dogs and cats, even, you wouldn't find any way to sort of make one just like the next, almost just like the next, almost just like the next to get from cats to dogs, where there wouldn't be, as a matter of fact, you'd find there was nothing whatsoever somewhere in there. Now, be a big space somewhere in there. Things don't fade into each other, except of course, within clumps, they do. I mean, they're also are, you know, they're, the clumps of these clumps are pretty vague, you know, exactly what is going to count as a member of the clump is, is not clear. So my, anyway, my claim is that the, the thought is only possible because there are such clumps. We can latch onto them, and that's just why inductions are possible. And that's why we could have knowledge. Okay. So, you know, in traditional, I mean, there's, there's a very helpful glossary, I should say, at the, at the back of the book, which, you know, listeners can, can look at at their leisure. But, so the vocabulary I used to ask questions will not necessarily be your vocabulary, but we'll, you know, yeah. So, you know, traditionally, we think about, you know, and this, you know, count as well, but we think of, you know, we have concepts or ideas, you know, to use the empiricist term, and we get sensory input, we store and organize this information some way about the world in concepts or ideas. And that's all very, very, you know, vague, but I mean, the vocabulary and some sort of mental entity of some sort is, is clearly what, what people are trying to capture there. You however, you know, introduce, I mean, as the subtitle shows, the concept of a, of a uniceped, and then the associated concept of a unit tracker. And since, you know, these two, these two concepts are central to the, to the book, maybe you can tell us a bit about unicepes and about unitrackers and, and, you know, how the unicepes are distinct from concepts. And that will, that can get us started, I think. I take it the most fundamental job that cognition has for any animal. That's difficult job, really. Is to be able to recognize the same thing again through many, many, many different sensory inputs, right? I mean, a child learns to recognize a dog by its bark, by its, what it looks like for many angles, by what it feels like, by what, by, by, eventually, learn to recognize information about dogs, by reading the word dog. That, anyway, I take it that, that really a very, very fundamental thing that any animal has to do, I think higher animal has to do is to learn to recognize the same thing again. If you don't, you can't recognize the same thing about, again, the knowledge you have of it is going to be useless, right? You can't apply it. If you can't recognize the same thing again. So, all right, then the claim is also that, as I said, that the world, the world comes in clumps. And so what one is really trying to do is to recognize these clumps. So the dog is a clump, a hole, a bunch of animals that are pretty much like each other, maybe some of the fringes that are different in some basic ways, but kind of a clump. You have to learn to recognize that, an instance of that clump, so that you can apply knowledge that you learned before of dogs or whatever. So this is true, not just, of course, of things like species and kinds. I thought that through properties, perceptual constancy, mechanisms are designed precisely to do this, to learn to recognize these the same color and they're many different lighting conditions, the same shape and they're many different perspectives, etc. Anyway, let's take this to be a very fundamental job. And what a unit tracker, I take to be simply, you can think of it as a neural apparatus, whose job is to learn over time to recognize the same thing again and to channel information that's gained about that thing into sort of one place so that those various pieces of information can interact with one another. Yeah, so that everything you know about some substance will be in a form that you know that it's about the same thing and can use it as about the same thing. So if it has five characteristics, you know that they're all characteristics of the same thing and you might be able to infer something from that and see. So that's what a unit tracker is, it's a top kind of neural net, I suppose, very complicated. A unit set then is simply what the unit tracker feeds into that manages in somewhere or another to store information that's been acquired by the unit tracker, by tracking the same thing over time. So the unit set then is what basically figures and say a propositional knowledge about something, practical knowledge about something. So you can think of it as initially at least as a folder people have been talking now since I've just lost to be honest really about thinking of a folder in which you put information, which is what the only information about Johnny or all the information about dogs, let me put in a folder. But I think that there's only one thing that's bad about that and that is that model, things that you know about Johnny or about Fido, say that the Fido is brown. If you've got the brown and the Fido folder, if you've got brown also in the spotty and overfolder, the problem is that you have to somehow rather recognize that brown is the same and the folder analogy won't give you that. So I'm thinking in terms of a much more intensive network, which is a matter of fact, the brain is an incredibly intensive network and which also properties are recognized by an individual kind of place of storage, which is connected to all of the things that have that property, one knows has that property in the brain. So a unit tracker then is something I have my unit tracker for dogs. You have yours. A unit tracker is an individual thing. It's not like a concept because we share concepts. You don't share unit trackers. You've got yours for dog and I've got mine. And we may have somewhat different ways of recognizing dogs. That can recognize dogs maybe in some ways that I can't. Helen Keller recognized dogs in a very different way, probably from a way you and I do. She certainly had a unit set for dogs. A unit set and a unit tracker for dogs. But it was very, very different from the ones that you were I have. So the unit tracker is individual. You have yours for dog and I have mine for dog and what they have in common is that they both are focused on dogs. That's out in the world. So it's their connections with the world that makes them alike, not necessarily anything that they have sort of psychologically in common. Let's see where maybe we do have quite different methods of doing the same thing here. So it's very unlike a concept in that regard. Also unlike concept in another way. I take it that different people will have different sets of ways of recognizing the same thing and that there will be no way of recognizing that's a definition of dog. So you don't find out, the way of which words are connected to you, this is another story. What is not true is that your idea of a dog is you and I both think about dogs that what we have in common is that we would identify. the same thing as dogs, say in other possible worlds or etc etc. So the method of recognition is not what defines what I'm thinking of. It's what I'm recognizing by that method that defines what I'm thinking of. So that's very different from thinking of a classical concept where usually you think of, you know, people talk about our concept of a universe, our concept of this. There's no such thing as our concept of, right? No such thing as our unicep, we all have different ones of the same things. So what they're all like on the end that they are of the same things. All right, does that help? Sorry. Yeah. No. So, so let me just, you know, since you mentioned, you know, words, I mean, one of the questions that comes up with a sort of a radical, radically individualist, you might say, I mean, I think that's the way I'm talking about it. The way I'm talking about it. identity of kind, the identity of properties is aware of the word. So that's to represent identity by similarity of that sort. It's the way language works. So if I said John did this, and then a few minutes later I said John did that. You know, it's the same things being talked about because I used the same. A duplicate, I used a duplicate word John in another sentence, right? There's another way to recognize identity or in other ways to more fundamental. In fact, way to represent identity. And that's by using it. The very same individual thing again to represent. Well, I'll give you an example. You're making a map. And you have one little mark there that's for Pittsburgh. And you show the relation of Pittsburgh to Boston, to San Francisco, to Philadelphia. You know, all with the same thought. You don't duplicate. You don't say Pittsburgh is this and Pittsburgh is that and Pittsburgh is something else like in language. Pittsburgh is that thought. And it's connected to all of those other things, but it's it itself is standing, you know, is standing. That's the kind of move the straws have made, which has developed into the current talk about folders. As folders as being, but concepts are. But I mean, that is the that's the important thing that straws and did in his move. What's exactly to say? You know, sameness is recognized by identity, not by similarity. So I just I'm just pushing that further. I want to say that that's the way identities recognize this by sameness. Actual numerical identity. So I want to show that that's a fundamental way of recognizing identity. What did you ask me? Yeah, the language. Yeah, the language stuff. Oh, all right. So, all right. So one of the things that's wrong with the language of thought is it's like language that uses duplication. Instead of identity to represent identity. All right. All right. There's another thing about the language of thought that's very important. But it's interesting because it's not Jerry voter. The language of thought idea was developed by people like. Well, well, the sellers was probably primary primary here. And the thought there was that what made a belief say about something. Was the fact that it was connected by inferential dispositions. To a lot of other things that the actual but the thing was itself connected to. So what made a dog, what made your thought of dog about dogs. Was it having thought of a dog you think four legs and then you think it barks and then you know you have all of these inferences that you will make. Given and also input inferences thing in a way that you will infer that something is a dog by knowing something else about it. So those inferential connections. Were what made the thought of dog about dogs. So it was the connection the network, the connection of of these ideas to one another that gave them their aboutness. So in sellers, for instance, you find him saying explicitly that you can't have a thought of one thing without having a thought of a whole lot of things. So that's the language of thought picture, although it's interesting because Jerry Foder himself of course argues against holism and on I think his arguments are wonderful. He was absolutely right, but this kind of holism really can't stand up. But but anyway that thought that somehow there you find out what a comfort about a concept what you find you record watch out you describe a concept by describing its inferential connections to other things. That is usually bound up with the language of thought view has has been and that I reject completely. Oh, that you are inferentialism I reject completely. So let me let me I mean we've been talking a lot you know sort of touching on the idea of you know language and and once we get to. And that that we kind of proceed to the to the second part of the book where you talk about infosigns and intentional signs and and that sort of a thing. So we kind of build from the uniceps to the what I saw as you know at least. A more straightforward extension of your you know previous work in language thought and you know other biological categories where you're directly talking about you know intentions and language and things like that. So there's a number of different different topics here and maybe we should start about you know the concept of a of a info sign or a sign that carries natural information intentional signs as a you know a particular type of info sign. And then also you know how are these sorts of signs related to the to the uniceps. Alright so uniceps is trying to collect information over time and perhaps from from various perspectives about the same thing it's trying to recognize the same thing through various different signs of them. You know think of what's hitting the retina what's hitting the eardrums as signs of the external world. Alright so this is something that I hadn't talked about at all in my previous work and that is how do you how do you think of these signs what is the relation of these what is the related to what hits your eardrums to the thing that you recognize as a trumpet you know what is the relation here. Alright so I begin then by trying to give a description of what I call info signs think of it roughly as what people used to call natural signs. Back clouds mean rain the rings on the tree show how old it is so etc. Those of trying to give a write a description of natural signs of what I call info signs signs signs signs signs signs in the natural world that carry information on trying to give a description of what kind of what it is for us there to be a sign that carries information in the natural world. So that's the first project then the second project is to say how something like language the signs of language are related to that. And the rough claim is is is all you know the rough claim is rather striking maybe what I want to claim is that signs like language signs thoughts signs etc. When they are working properly when they are true when they are correct. And when they got to be correct by in a normal way you know not just by accident accidentally have beliefs that are true but when they got to be true in a normal way that they actually carry information in the same way that black clouds they carry information about rain. So that basically understanding language has a surprising similarity is simply recognizing that it's a trumpet when you hear it. So words that words actually I want to argue that words actually carry information. When the sentences when when they are in true sentences and those true sentences were arrived at normally in the way which historically is accounted for the proliferation of the mechanisms that do this stuff. But in those cases that of the language carries natural is it let the language is a natural sign of what it's about. So all right so that's the connection so then the question is all right so what makes language different. And the answer is that language has been purposefully designed to carry information by purposefully I simply mean by natural selection by learning by it all various analogs of natural selection that all the mechanisms that produce and understand language that produce and understand any intentional sign are mechanisms that have been designed to do that. So it's not by accident when you speak a true sentence to a friend it's not by accident that it carries natural information right language is designed to carry that natural information and that's what makes it different from just ordinary natural information. It's it's designed natural information information. So we don't charge again. Yeah so let me let me just you know speak of conventional signs and in this part of the book in particular I was I was thinking. thinking of the pragmatic semantics. And I noticed that you had a couple of references like to Robin Carsten or Francois Reconnaity. And people-- can you say something about your view of meaning of these conventional signs? What one of the debates here is the extent to which the meaning of a linguistic item is electrically encoded or core or context independent. I mean, there's a number of different ways of referring to roughly the same phenomenon. And of course, the pragmatic semantics say there's not very much that is electrically encoded, you might say, in the meaning. Where do you stand on that? And it seems like you would be much more of a-- well, I mean, you just-- your description of uniceps those being particulars. It seems to blend itself to a fairly radical indeterminacy of meaning and a very pragmatic semantic e-type view. So could you say something about less? I think the symptoms you were seeing are not symptoms of the-- oh, it's just suggesting. Let me see if I can-- because yeah, one of the things I try to do-- people talk about this is the literal meaning. And then you have this complicity. And then you get different meanings. What is the question I would ask is, all right, tell me really what constitutes that this is the literal meaning? And everybody sort of assumes, oh, well, there's the literal meaning, right? And then there's compositionality. And then we turn out meanings of soul sentences. The question is, what constitutes that this is the literal meaning? And my claim here is that if we can look at language in a very interesting way, you can look at language as actually having an evolutionary history in this sense. Not that somehow or other-- the word dog is somehow rather a better sound than anything else to stand for dogs. But that evolutionary history in the sense that evolution is very interested in symbiosis and the way on which, say, different animals can actually develop traits that are coordinated, such that they can work together on something. So what languages is, of course, the development of coordination between speakers and hearers. All right, so my claim then is that you can look at-- you can ask yourself, what is the function of a term in language? And what you can mean by that, interestingly, is what is it that accounts for it continuing to be used by speakers in roughly the same way and understood by hearers in roughly the same way? What is it that accounts for its proliferation? Or you can go back-- thinking, you can't entirely enter-- what accounts for its survival? What accounts for its survival without changing its meaning? What is it doing that accounts for its survival? So I take it that the very central question to be asked about any part of any idiom, whatever, is what's its survival value? What's it doing? Now, what that gives you is a very different picture of language than perhaps as usual. You have to think in terms of what some linguists call constructions. The construction is not-- two things are not the same construction because they sound alike, or look alike. Two things are the same construction, what the often said-- because not only do they sound alike, but they have the same meaning. So if you have hominins, they're different constructions. But my claim is that what really corresponds to a construction is-- you could think of it as a lineage. It's a string, a reproduced cases. Dog, dog, dog, dog, reproduce, reproduce, reproduce, reproduce, reproduce, reproduce. I'm going to ask why does it continue to be reproduced? What's it doing? All right, so that's where you would look for literal meaning. What is it? Let's look for a lineage. Let's look for something because maybe reproduce, reproduce, reproduce, and reproduce, to do this particular job. But if you do that, then it's going to turn out that-- what shall I say? Not only that there are hominins, but that there are also words and phrases that are related to each other, but that may as a matter of fact proliferate quite independently. And people who study language have recognized this. People learn a huge number of phrases of-- well, people call them a constructions, right? That are copied and copied and copied and copied. And that's what's learned rather than learning a lot of individual words and then putting them together every time you make a sentence that's just not how it works. Psychologically, big chunks are copied. And that means that big chunks can have a literal semantic meaning without taking them apart, without deriving that from the internal structure. And it also means that there can be a lot more hominins as it were in the language than you might think. That is independent lineages that are proliferating the same sound with somewhat different uses. But I use it this way because other people used it that way. So what this does is just suggests there are a lot more literal meanings than you would have thought. And the most, I think, a huge proportion of these of what has been taken to be-- to require pragmatics to explain-- it's actually explained as a literal meaning. It's been reproduced and reproduced and reproduced to mean that. So one of-- I mean, there's a couple of problems that you address in terms of having how infosigns relate to the things that they signify. And there's two in particular that you discuss at some length. One is the strength of correlation problem and the other is the reference class problem. And I was just wondering if you could say a little bit about these two issues that arise. I take it that there's-- even though there is a lot of-- you have these lineages. And there's a lot more. Sounds very polysemic in a way. There's still stability there. And so could you say something about both of those problems, the strength of correlation problem, and then the reference class problem? All right, the discussion of the strength of correlation and the reference problem was in the attempt to define infosigns. Infosigns, as you remember, are something prior to intentional signs. So we're talking about something that's more fun about the amount of language here. We were talking about language when language is functioning-- well, I say it's functioning properly. That is, functioning in the way that has accounted for its survival in not in ways that-- not in ways that are not going to continue and be reproduced, but in ways that-- OK. So one project that is just defining natural signs-- that's one way of putting it. No, I call them-- infosigns, signs of carry information. One classic way of doing that-- and at least modern classical way of doing that-- is by saying that a sign is-- the relation between a sign and what it signifies is correlation. But there's a correlation between these two so that are put it this way that when the sign occurs, there's a better chance of what it's signified occurring than if the sign hadn't occurred. There's a correlation between the sign and the probability of the signified. So all right, so I could take off from that. I say the problem here is people often don't-- people often simply talk about, what's the probability of having happened? There's no such thing as a probability of something having happened without a reference class. Do 18-year-olds have more accidents than older people, right? 18-year-olds where, in what culture, at what time? You know, whenever you do statistics, you have a ring around a reference class. And it turns out that if you think very hard about this, you'll see that the problem here is that you would have to say what reference class was involved in order to say, "Oh, it's a matter of correlations." So, so, I'll say, this is not an easy issue, but roughly my conclusion here is that it turns out that being a natural sign of something that is not a natural sign. Really, it's like, what shall I say? Being a sign for an animal is like being food. Whether or not something is food depends on the animal who wants to eat it. Similarly, whether or not something that's going to count as information, as carrying information is going to be relative to an animal. That's going to be relative to the capacities of that animal, relative to the, relative to this place in which the animal is lives and moves. And then you have to take a account of this. Anyway, that has a lot of interesting results that I can't go into here, but, but all right. So, you can see that that's that's part of how context then enters the matter here. Right context is what reference class are we in? And that sense is very important, but there's another way in which context is very important in which I think you may have in mind, Carrie, I'm not sure. What I've also claimed that in the case of language, very often pieces of the environment are actually part of a sentence. In other words, to think that language is only of utterances is confusion. That is the matter of fact utterances placed in juxtaposition with objects. The entire configuration is often what's the sign. And so this is why I give an analysis of indexicals, for instance, which which are out of demonstrators. In which the thing demonstrated is actually a piece of the sign like a sign that says. No lories on a bridge bridge. Part of the sign is the bridge. You have the slightest idea what that means unless you see that it's all a bridge. The bridge is part of the sign. It stands for itself. So actually the theme of the themes that there are things that in certain contexts can sort of stand for themselves is also a theme that. Weaves in and out of the book here with various effects. So that that that that sounds a bit like, you know, the extended mind thesis, you know, Andy Clark and David Chalmers, I suppose, are you are you endorsing that sort of a few. No, no, no, no, why not? No, I mean, there's not well, I don't think that any mental signs are composed partly of things in the world. No, I think that's quite a different, quite a different question. What shall I say? I think the question for the mind is partly external. It's unfortunately largely a verbal problem. I don't think that there's very much substance there, but I'd have to back that back that up when I. Let me, there's you also introduced this concept of a reproductively established families. You know, it plays an important role. I don't know if you would want to go into that to flesh out what you've already said or. That was a term that was defined back in language, another biological categories. And I'm not going to, I won't define it. I want to try to go through that here, but just the way it applies to language is fairly straightforward. And that is what I was talking about, lineages, the word dog being copied and copied and reproduced and reproduced and perhaps phrases being reproduced and reproduced and reproduced and reproduced and reproduced they have lineages. Those are those families where where one thing is reproduced from another from another from another, those are reproductively established families. You know, well, the broader definition, we don't need to play with here. I think reproductively established family. What would you say are the main, main differences or advances or modifications that you introduce in this book as opposed to, you know, the work that it has built on. It's done prior work or work of others by on prior work. The explanation of uniceps the. I talk quite a lot about empirical concepts, right starting with language, another biological categories and and going through a clear and confused ideas. A lot about this. What I was working toward in my mind, I guess, was the notion of of a unicep, a uter-trucker and the combination of a uter-trucker with its unicep or a unicep with its unicep. And so there are a lot of things that I said earlier about empirical concepts. I would now say that was a very bad, very bad terminology. And if you look at it as a discussion of uniceps, it may make a little more sense, I think. So I'm definitely carrying that on, but also changing it by having discovered, I guess, how very, very different in so many respects, the notion of a unicep or a uter-trucker is the classical notion of a concept. So that's one thing in this change. But also the whole basis of talking about information theory and the last half of the book and also spelling out something about spelling out the notion of the literal semantic meaning. And also some things at the end of the book. I mean, I talk about some specific problems people have had about language talking about, well, talking about indexicals and demonstrators, for instance, I hadn't done that before about intentional context. I have more to say, but also tackling problems like, what do you do with, definitely descriptions that aren't definite, the dog. Well, I saw the dog. So, you know, talking about some ways of language works that you could look at language somewhat differently from through these eyes and some of the results you get. So, well, one of the things you do at the end is to draw certain parallels to talk a little bit about perception, right? And perceptual content. Could you, could you, you know, say something about, you know, the perceptual content and, you know, sign content? Yeah, that's another huge issue. And, yeah, I think what I do is to try to back up a claim that I think most people have found quite bizarre. And that is that in understanding, that understanding language is actually a variety of something like perception. That, yeah, the processing, the processing and understanding language is like perceptual processing. So, this requires looking at perception in rather a different way. So, one thing, one thing that I do here with perception, I would like to understand perception or try to argue that you can understand perception. As interpretation of signs, you know, the classic classical view that what you do is just, and say, in visual perception, just to try to reconstruct what could have caused the, could have caused the visual sensations, etc. It is you sort of reversing the, you have to somehow rather reverse the process by which the object hits your eye in. So, I want to claim instead that no, it's like recognizing a sign and recognizing the sign for something I do not have to have the slightest idea how the sign got to be there. And what, by what process the sign got to be there if I could recognize the sign, I may not have the slightest idea why those signs are around. So, recognizing a sign is very different from understanding anything about the process by which the sign was produced. So, anyway, it, yeah, it turns out that that makes quite a lot of difference. So, I think if you look at perception that way, there's some Gibsonian tinges here with, you know. So, we are just about out of time and I like to ask what, um, What's on the immediate horizon for you? Are you working on, well, I don't know, maybe perception or perceptual content or what's your next step at this point? Philosophically. - All right, all right, so I'm working on a couple of things. Let me come again. One of them is, I said that I wanted to understand the meaning of us. I wanted to understand linguistic meaning, by trying to understand what accounts for the reproduction of the words and citatic structures, et cetera. What are they doing that accounts for their reproduction? And what I have an essay now that I'm trying to, I never had looked at all of the things that language does besides representing. It has many other functions than representing. So anyway, so I have a working on essay at the moment that it's a more careful explanation. I hope of what it means to say that you're looking for the actual function of a linguistic term. And then a number of tentative suggestions about puzzling words about how probability words about probable words about existence, identity, I don't know, they're a bunch of them, that are sort of classic philosophical. If then, a number of classic questions that I think a lot of can be cast on, but looking at function instead of looking for linguistic conditions. - Okay, well, anyway, that's one. But the other is that I'm also working a little on the current context, trying to make very much more simple and obvious the notion of representation and what makes a representation count as correct or incorrect, trying to make this more straightforward, more obvious than perhaps I had before. 'Cause I think there's been some strange controversies about whether they're a mental representations, right? - Right. - Okay, well, I'm sorry, but we are out of time. But I appreciate you're taking the time to talk with us about your new book. It's been a real pleasure on my part. - Oh, thank you. - Thank you, thank you. - You're giving me the opportunity to do this, yeah. - You've been listening to my interview with Ruth Garrett Milligan. We've been talking about her new book, Beyond Concepts, Uniceps, Language and Natural Information, which is just out from Oxford University Press. This is New Books in Philosophy, I'm Carrie Fickdor, I hope you enjoyed the podcast and thank you for listening. (upbeat music)