What Kind of a Fact is a Flying Pig for Kant? and Things like That
Interview by Richard Marshall.
‘For Kant, when you make a judgment, you combine representations – the constituents of the judgment – together in thought. There are different ways the mind can combine these constituents, different ‘functions of unity’, resulting in judgments of different logical forms. For example, (very roughly) I can combine the concept dog and the concept mammal in a positive way to form the judgment that dogs are mammals, or in a negative way to form the judgment that dogs are not mammals. At the heart of Kant’s theory of judgment is a table of these different functions of unity (there are twelve).‘
‘Kant thinks that there is a correlation between objectivity and necessary unity. For example, what differentiates an objective experience, e.g. of a ship sailing downstream, from a subjective state, say a daydream about a ship sailing downstream, is that the events in the objective experience are necessarily ordered – the bit with the ship upstream has to occur before the bit with the ship downstream – whereas the events in the subjective state are only contingently ordered – I can jumble up the movement of the ship in my mind.’
‘I think there are good reasons to think that the laws of logic are laws of thought. I think this view helps us to explain why it is that the laws of logic seem to have a particularly close relationship to our thoughts, for example, that it seems really hard – indeed, impossible – to rationally doubt basic logical laws.‘
Jessica Leech is a philosopher of modality. She is interested in contemporary issues in the metaphysics of modality, such as the notion of essence, and the relationships between different kinds of necessity, has explored what Kant had to say about modality, and issues arising from that and is currently working on a book which attempts to draw out Kant’s view on modality, applying them to contemporary issues in the metaphysics of modality. Here she discusses modality, flying pigs, the metaphysics of modality, Kant and judgment, the distinction between logical and real modality, the relativity of necessity, whether laws of logic are laws of thought, the logocentric predicament, and issues arising from hyperreliable belief forming methods and why the a priori/a posteriori isn’t slicing at the epistemological joints. Groovy…
3:AM: What made you become a philosopher?
Jessica Leech: I suppose I was always interested in asking difficult questions. I remember asking my mum when I was quite young why some sounds were meaningful and some not, for example. I was so happy when this finally came up in Philosophy of Language! I also had a great RE (Religion) teacher at school. His lessons were much more philosophy than anything else, and I really enjoyed the class debates we would have about issues such as divorce, abortion and God. I was bored a lot at school, and I was happy to finally find something that was interesting. That got me into doing Religious Studies A-Level (which was effectively a philosophy course), and I didn’t really look back from there.
3:AM: You’re an expert in the notion of modality. This is a term of art in philosophy and may not be clear to people outside philosophy what it is. So perhaps to start with you could say something about what modality is and what the stakes are? Why is it important?
JL: Modality is a catch-all word for notions like possibility and necessity. We use modal terms all the time, words such as ‘must’, ‘can’, ‘might’, ‘could’ etc., so it’s important to understand what we’re talking about. But I’m also interested in modality because it presents some distinctive puzzles.
It might seem quite simple to explain non-modal facts or truths. For example, it is true that grass is green because grass is green – and we have ways to check that, perhaps by looking at some grass. But so-called modal truths or modal facts are more puzzling.
Suppose it’s a fact that there could have been flying pigs, although there actually aren’t any. What kind of fact is this? What could make it true that there could have been flying pigs? There are no flying pigs around – so what else could explain this kind of possibility claim? Philosophers have made lots of different suggestions, all of which have advantages and disadvantages. For example, you might think that it is something about pigs – the actual pigs – that they have a potential to fly (even if they never actually realise that potential), or you might think that there exist merely possible pigs that do fly, or merely possible worlds in which pigs fly, or that it is just that the meaning of the word ‘pig’ doesn’t rule out flying. So possibility is puzzling.
Consider also things that are necessary. Most philosophers agree that it is necessary that 2+2=4. It really couldn’t have been that 2 things and 2 things added together came to anything other than 4 things. Now, just like the case of other things that are true, we can verify that 2+2=4 is true. But it isn’t merely true, it also couldn’t have been otherwise. What can explain why it is necessarily true, as well as true?
In sum, as soon as you realise how often we make modal claims – even if we’re not aware that that is what we’re doing – it becomes important to solve these kinds of puzzles. We should be interested in understanding what kinds of claims we’re actually making, and what it would take for those claims to be correct.
3:AM: You’re interested in the metaphysics of modality rather than just modality – is this the territory that Tim Williamson has been writing about recently?
JL: Yes, although his approach and his views are rather different to mine. His recent book, Modal Logic as Metaphysics, addresses questions in the metaphysics of modality, but he takes modal logic as his leading light. As I understand him, he thinks that these metaphysical questions are inseparable from our choice and development of a particular logic for modality.
3:AM: Kant’s a looming figure in your work isn’t he. One way he discusses modality is in terms of judgment. What are these modal forms of judgment and how do they fit in to his wider system?
JL: That’s a big question! I can’t do it justice here, but I’ll try to give the short version. For Kant, when you make a judgment, you combine representations – the constituents of the judgment – together in thought. There are different ways the mind can combine these constituents, different ‘functions of unity’, resulting in judgments of different logical forms. For example, (very roughly) I can combine the concept dog and the concept mammal in a positive way to form the judgment that dogs are mammals, or in a negative way to form the judgment that dogs are not mammals. At the heart of Kant’s theory of judgment is a table of these different functions of unity (there are twelve).
Three of these functions of unity are modal. But they have a different role to play to the others. Whereas the other nine are different ways to unify the constituents of a judgment, resulting in different logical forms, the modal functions concern instead the way the whole judgment is related to the subject and her other judgments. I think the best way to understand this is in terms of the location of the judgment in a course of reasoning of the subject. If the judgment is made as a conclusion to an inference, then it is apodictic (necessary). Given the premises, it must be true. If the judgment is made as a premise, then it is assertoric (actual). It is assumed to be true for the purposes of argument. If the judgment is a constituent of another judgment, and so neither taken to be true, nor to follow as a conclusion, then it is problematic (possible). For example, if one judges assertorically that either P or Q, for Kant, the judgment that P is problematic, because it is not itself assumed to be true (only the whole disjunction, P or Q, is assumed to be true, not P on its own).
How does this fit into his system? In several ways. In brief, Kant’s theory of judgment is an important part of his theory of experience and knowledge. Hence, understanding his theory of judgment is important for understanding the rest. Kant claims that there are a priori concepts – categories – that we must possess and use in order for us to be able to have experience of an objective world. They provide a familiar structure to the world we experience: physical objects, with various different properties, standing in causal relations to one another, subject to laws of nature. The twelve forms of judgment play a central role in Kant’s explanation of these categories, so again, we need to understand the forms of judgment in order to understand the categories.
More particularly, I think that the modal forms of judgment play an important role in Kant’s account of the unity of the thinking subject. In simple terms, we can think of the unity of a thinking subject in terms of necessary connections between all of that subject’s mental states. The modal forms of judgment involve necessary connections between a subject’s judgments: every judgment is part of a course of reasoning, as premise, conclusion, or part of a premise. So the modal forms of judgment can help to explain the unity of that thinking subject.
Finally, how I understand the role of the modal forms of judgment is not uncontroversial. I have argued for these claims in detail elsewhere, but my explanation above is by no means agreed upon amongst Kantians.
3:AM: Kant also talks about modality in terms of his table of categories. So what are the contents of these modal categories and again how do they fit in with the rest of his system?
JL: In general, the categories are pure concepts – concepts that are not derived from sense experience, but rather play a role in the very possibility of experience at all. Associated with each category or group of categories are principles that, for Kant, are necessary conditions of the possibility of experience. So, for example, one of the categories is causation, and the principle of causality is that every event must have a cause. Kant argues that fundamental aspects of objective experience depend upon such principles.
Amongst the categories (there are twelve), there are three modal categories: possibility-impossibility, existence-non-existence, and necessity-contingency. And there are also three modal principles:
1. Whatever agrees with the formal conditions of experience (in accordance with intuitions and concepts) is possible.
2. That which is connected with the material conditions of experience (of sensation) is actual.
3. That whose connection with the actual is determined in accordance with general conditions of experience is (exists) necessarily.
For Kant, there are formal conditions on how minds like ours are able to be presented with objects and how they are able to think and experience those objects. He argues that we are presented with objects in the forms of space and time, and that we think about them in accordance with the categories and their principles. Together these thus comprise the formal conditions of experience. So, the principle of possibility states that whatever agrees with these conditions is possible. For example, things that are spatiotemporal, and events that have causes, conform to these conditions, so they are possible.
But it is not enough to have formal apparatus in play. We also need to be given – presented with – objects via the senses (what Kant calls intuition). The material conditions of experience are thus that we need to be presented with objects to have experience of them – we can’t, for example, create objects just by thinking about them. If an object is given to us in intuition, via sensation, then it is actual. Contrast this with, for example, the concept of a human that is 9 foot tall, which seems to be compatible with the formal conditions of experience – so is possible – but is never given to us in the matter of experience, and so is not actual.
Finally, neither the formal nor the material conditions of experience concern particular things. They simply provide a general form that experience must take, and the general principle that objects must be given to us. Putting these together, then, we get the general conditions of experience. Because these are general conditions of experience, they lay out what must necessarily be the case in the world we experience. However, because they are general, they cannot determine that anything in particular exists necessarily – they only say how things in experience must be in general. In the principle, then, Kant doesn’t bother to define when something exists necessarily, but rather a kind of hypothetical necessity: given what is already actual, what must also be the case, according to the general conditions of experience. This final modal principle, then, concerns necessary connections between actual things.
This is a brief summary of how I understand the content of the modal categories. How do they fit in with his system? Again, my answer to this is not uncontroversial, and rather complicated. In essence, I think that the principle of possibility is intended to play a role in restricting the domain of application of the categories to the world of objective experience. Kant’s explication of the principle mostly consists of him providing several examples of a priori representations being assured application to objects of experience via the principle of possibility. The content of pure concepts is not derived from experience itself, and so there is a requirement for some additional mechanism to ensure they have a content that is applicable to experience.
I daren’t comment on the principle of actuality. This is mixed up with Kant’s Refutation of Idealism, which is a large and thorny topic, about which far better philosophers than I have had a lot to say.
Finally, I think that the principle of necessity plays a crucial role in the unity of experience. Kant thinks that there is a correlation between objectivity and necessary unity. For example, what differentiates an objective experience, e.g. of a ship sailing downstream, from a subjective state, say a daydream about a ship sailing downstream, is that the events in the objective experience are necessarily ordered – the bit with the ship upstream has to occur before the bit with the ship downstream – whereas the events in the subjective state are only contingently ordered – I can jumble up the movement of the ship in my mind. It seems to me that these necessary connections are precisely what are described in the principle of necessity.
3:AM: Kant distinguishes between real and logical modality doesn’t he. What’s this distinction about and why is it important – and how does it distinguish Kant from his predecessors?
JL: Yes, he does. Logical modality concerns the consistency of concepts and thoughts. For example, so long as it is not part of the meaning of the concept of an event that it is caused, it is logically possible that there be an uncaused event. But it is not, for example, logically possible for there to be an unmarried bachelor – that’s a contradiction in terms. By contrast, real modality concerns things – what objects of experience there can or must be. Hence, to be really possible something must conform to conditions of experience. So, for example, because it is a necessary condition of the possibility of experience that every event must have a cause (as I noted above), it is not really possible that there be an uncaused event.
One way this is important, is that we cannot fall back on logical, analytic or conceptual necessity to give a global account of the possibility and necessity of things. We have to do more work to understand real possibility and necessity. Kant’s predecessors (I’m thinking broadly of the Rationalists here) tended to hold the contrary view that the possibility and necessity of things just is logical modality. Today, many philosophers of modality would agree with Kant on this at least. For example, Kripke famously gave us lots of examples of basic necessities that are not a priori, and certainly not conceptual or logical. Hence, we can see Kant’s departure from his predecessors as an important – perhaps even revolutionary – step on the road towards our contemporary understanding of modality.
Nick Stang has much more to say about the historical development of Kant’s views on modality, both in relation to other philosophers, and in his own works. I would definitely recommend his book Kant’s Modal Metaphysics to anyone who is interested in this.
3:AM: A contemporary question you’ve faced is how to explain the idea of necessity being relative. There seems to be a paradoxical air to this notion doesn’t there. So how should we understand it – and how many different kinds of relative modality are there?
JL: I suppose I’ve never thought of relative necessity as being paradoxical. It’s just the idea that, in lots of cases, we qualify our claims of necessity. So, for example, I might agree that it is necessary, given the laws of physics, that nothing can travel faster than light, but concede that it isn’t necessary without qualification. Or, I might agree that it is possible, given general facts about human beings, for me to run 100 metres in less than 20 seconds, but be fairly confident that, relative to my own levels of fitness and motivation, it is impossible.
In general, I think we can understand relative necessities in terms of what follows logically from some other propositions (e.g. the laws of physics). In lots of cases, those propositions are themselves pretty interesting, and so we end up with kinds of (relative) necessity that are themselves interesting. For example, the laws of physics are a special class of propositions, and so physical necessity, defined as what follows from the laws of physics, is also special. In general, there are as many kinds of relative necessity as there are propositions to which necessity could be relativized. This doesn’t worry me. As I said, we can still pick out special and interesting kinds of necessity by being able to pick out special and interesting kinds of proposition.
There is interesting work on how many subtly different modalities we often use in natural language. Angelika Kratzer has some leading work on this, for example, her paper “What ‘can’ and ‘must’ must and can mean” (Linguistics and Philosophy 1, 1977, 337-355). There is also a fun passage in William Lycan’s book Modality and Meaning where he takes a page or two from a pulp fiction novel and systematically extracts all of the different modalities in it. In general, there is quite a lot of linguistic evidence that we use relative modalities in our language. That’s not a decisive argument for taking there to really be all of those relative modalities, but it would at least be a fairly simple explanation of what it is we’re talking about.
3:AM: Are laws of logic laws of thought and should we take this question to be a normative or constitutive one, or both?
WL: I think there are good reasons to think that the laws of logic are laws of thought. I think this view helps us to explain why it is that the laws of logic seem to have a particularly close relationship to our thoughts, for example, that it seems really hard – indeed, impossible – to rationally doubt basic logical laws.
There are different ways to understand how a law relates to what it is a law of, and so different ways to understand how the laws of logic relate to thought. Laws are constitutive for Xs if they tell us what is and what is not an X. If the laws of logic were constitutive laws of thought, then illogical thought would not be thought at all. But I think this is deeply implausible. If illogical thoughts are not thoughts, then it is too difficult to explain how we can come to recognise our logical mistakes and rationally correct them. Laws are normative for Xs if they tell us what is and what is not a good X. If the laws of logic were normative laws of thought, then illogical thought would still be thought, but it would be bad thought, and logical thought would be good thought. I think this is closer to the truth, but it still doesn’t explain why it is so difficult to doubt logical laws.
I therefore tend to favour a view according to which the laws of logic are constitutive norms for thought. The idea is that a constitutive norm for Xs tells us what is and what is not an X, not in terms of whether or not something conforms to the law, but in terms of whether something is evaluable in light of the law. So, to say that the laws of logic are laws of thought in this sense, is to say that something is only a thought if it counts as good or bad in light of the laws of logic. So on this view, an illogical thought can still be a thought, as long as it counts as bad. And a logical thought is only a thought if it counts as good. We can explain the close relationship between thinking and logic, because the laws of logic are, in some sense, constitutive of thinking. But we can make room for genuine illogical thought by treating the laws of logic as constitutive norms.
3:AM: One of the issues that all this bumps up against is similar to the ‘logocentric predicament’ – where we try and give an account of logic but are bound to use logic to do so. How does your approach tackle this issue? Does recourse to some notion of indubitable logical principles help, and if they do, where do they come from?
JL: This isn’t something I have a ready answer to. I’ve thought about how we can make sense of the fact that we can’t rationally doubt certain principles. That would seem to bear an important relation to the question of how we might go about trying to justify those principles. When we try to rationally justify some logical principle, we need to use some logical reasoning to do that, and so we find ourselves in this logocentric predicament. I haven’t yet worked out what I think about this, although I’m aware there are some very difficult issues lurking there.
For example, if we really can’t help but think a certain way, one might worry that this doesn’t guarantee that that is the correct way to think. If we can’t rationally doubt certain principles, then there is a certain kind of sceptical worry that we are stuck being unable to doubt false principles. In the paper I’ve written on this, I tried to address this issue of the truth of indubitable logical principles. There I appeal to Kant’s idea that conditions on the possibility of thought and experience are themselves also conditions on the objects of thought and experience. Hence, if certain logical principles are necessary conditions of thought of objects, then those objects are guaranteed to conform to the principles. Hopefully some of the discussion above will give readers a sense of how this is supposed to work.
3:AM: So can you sketch out your approach to how best understand logic and the laws of thought – and what advantages does your approach have over its rivals?
JL: As above, I have suggested that the laws of logic are constitutive norms for thought. According to this view, something is only a thought if it counts as good or bad in light of the laws of logic. A purported illogical thought is a thought, as long as it counts as bad. And a purported logical thought is only a thought if it counts as good.
Some alternative accounts take the laws of logic to have their source or basis in worldly things utterly unconnected to the mind. For example, one might think that there are abstract logical entities the essential natures of which ground logical laws. I think such views are unable to explain the fact that we cannot rationally doubt logical laws. On such a view, there is no principled relation between these logical entities and our thinking. My preferred view also does not slide too far in the opposite direction, into an unacceptable psychologism. It is important that the view does not concern how we do think, but norms for thinking.
3:AM: What’s so problematic about the contingent a priori example presented by Williamson that you discuss and Hawthorne’s issue with hyperreliable belief forming methods?
JL: I don’t think his example is problematic per se, just that it raises some interesting issues. Philosophers have tended to think that the contingent a priori is always connected to the use of some indexical, a term that has its meaning on an occasion of use fixed by some features of that occasion of use. A good example is ‘I am here now’. It is contingent that I, Jessica, am here, in Sheffield, now, at 5pm. But I can know a priori that I am here now, in the sense that, whoever I am, and wherever and whenever I am speaking, the way these words work ensures that I am speaking here and now. Williamson’s example is intended to show that there can be cases of the contingent a priori that do not involve the use of indexicals. I don’t dispute this, but I think it’s interesting exactly how his example works.
Williamson’s example is ‘there is at least one believer’. He argues that we can know this a priori via belief-forming method M:
M Given a valid deduction from the premise that someone believes that P to the conclusion that P, believe that P.
It follows from the premise that someone believes that there is at least one believer that there is at least one believer. So according to M, we should believe that there is at least one believer. The belief thereby formed is true. Moreover, Williamson argues that M is a hyperreliable belief-forming method – necessarily, any belief formed using M is true. Hence, the belief is knowledge. But this didn’t require empirical experience of a believer – M provides a priori knowledge of the contingent, non-indexical, truth that there is at least one believer.
I have argued that what is curious about M is that when we use it to form the belief that there is at least one believer, we thereby guarantee the existence of the fact that there is at least one believer (one forms the belief, and hence is a believer, hence there is at least one believer). This seems to be an importantly different epistemic achievement to coming to know a priori a fact that exists independently of one’s epistemic activities in general, and one’s forming of this belief in particular.
In the course of my discussion of this example, I defend Williamson against some worries that John Hawthorne has raised concerning hyperreliable belief-forming methods, as well as trying to answer some questions raised. In particular, Hawthorne argues that simply necessarily providing true beliefs (hyperreliability) is not sufficient for a belief-forming method to count as knowledge-conferring. There must be a suitable relation between the belief-forming method, and the kind of fact about which true beliefs are formed. My account of how M works can meet this challenge: the belief-forming method creates the relevant fact (or at least, guarantees the existence of it). By using M, one forms a belief, creating the fact that one is a believer, which in turn guarantees that there is at least one believer.
3:AM: How do you approach this and why do you conclude that the a priori / a posteriori isn’t slicing at the epistemological joints. Why is all this significant?
JL: As I understand Williamson’s example, it involves a belief-forming method that guarantees the existence of a fact corresponding to the belief formed. This seems to be a very different kind of epistemic achievement to coming to know some fact independent of particular empirical evidence of that fact, where the fact exists completely independently of one’s epistemic activities. Hence, I suggested that maybe `a priori’ is too broad a classification, if it counts these together.
3:AM: And finally, are there five books you could recommend that will help readers here at 3:AM get further into your philosophical world?
Bob Hale (2013) Necessary Beings: An Essay on Ontology, Modality, & the Relations Between Them. Oxford: Oxford University Press.
Bob was one of my PhD supervisors, and continues to be a great collaborator and friend. His work has been enormously educational for me. We don’t always see eye to eye philosophically, but his recent book is a really excellent treatise on modality, logic, and metaphysics.
Immanuel Kant, I. (1998) Critique of Pure Reason. New York: Cambridge University Press. Translated and edited by Paul Guyer and Allen W. Wood.
Recommending the Critique of Pure Reason is probably a bit much, but it’s the source of most of the ideas in Kant that I’m really interested in. I don’t think it’s as difficult to read as its reputation suggests. But if you’re new to Kant, it’s best to read it alongside an introductory companion, of which there are many excellent examples.
Saul Kripke (1980) Naming and Necessity. Oxford: Basil Blackwell.
This series of lectures has set many of the terms of the contemporary debate in the metaphysics of modality. It’s an absolute classic in analytic philosophy.
David Lewis (1986) On the Plurality of Worlds. Blackwell.
Another contemporary classic. Here David Lewis sets out a theory of modality (and more) in terms of the existence of concrete possible worlds. I don’t agree with a lot of Lewis’s view, but this book is incredibly well-written, and is another prerequisite for contemporary metaphysics of modality.
Penelope Mackie (2006) How Things Might Have Been: Individuals, Kinds, and Essential Properties. Oxford: Clarendon Press.
I love this book. In it, Penelope Mackie carefully and thoroughly argues against many contemporary essentialist theses. I’m not a great fan of essentialism, so I was really happy to find so many excellent arguments in this book to support my own suspicions. I very much admire Penelope as a philosopher, and this is a great book.
ABOUT THE INTERVIEWER
Richard Marshall is still biding his time.
Buy his book here to keep him biding!
First published in 3:AM Magazine: Saturday, December 3rd, 2016.