:: Article

Impossibility and Non-Existence

Interview by Richard Marshall.

[Additional pics: David Lynch]

Impossible worlds can help making distinctions between necessary equivalents: add a world compatible with all one can know a priori, but where Ali is different from Clay, and you will distinguish those meanings. Such a world is metaphysically impossible: a way things could not have been.

It’s the so-called ‘imagery debate’, one of the most intractable problems in the philosophy of cognition. There are four main candidate answers on the market, I think: mental representations represent (1) linguistically (in whatever way language represents, without the intervention of sensory modalities); (2) pictorially (in whatever way pictures represent, via quasi-spatial features and the involvement of sensory modalities); (3) a mixture of (1) and (2); (4) magically (they just represent, with no further questions asked on how they do it).

Franz Berto’s research focuses on metaphysics, in particular Meinongianism (the theory of objects put forward by the Austrian philosopher Alexius Meinong), ontological commitment and modal metaphysics (the study of the ontological status of possible worlds). He also studies what’s known as continental philosophy, with a particular focus on the dialectic and rationalist philosophers, and the philosophy of logic and computational philosophy. Here he discusses possible and impossible worlds, paraconsistency and dialetheism, the logic of the imagination, whether conceivability entails possibility, fictionalism, noneism and Meinongianism, Quine vs Meinong, whether reality is digital or analogue, and whether the concept of identity can fail to apply to somethings even when we can count them.

3:AM: What made you become a philosopher?

Franz Berto: I wanted to stay home and think about fundamental problems. I ended up working in eight universities across six different countries. I thought I had signed up for armchair metaphysics: it turned out to be a window seat in the economy class. 🙂

3:AM: One of the things you’re interested in is possible and impossible worlds. Can you introduce these two items to us and say what philosophical issues are raised by these things? I guess this broadens out into a question about why you think such seemingly esoteric things are important?

FB: Possible worlds are ways things could have been (the Nazis could have won WWII). Impossible worlds are ways things could not have been (two times two couldn’t have been five).
Some fifty years ago, under the shadow of Quineanism, many philosophers considered possible worlds esoteric. Then came the ‘intensional revolution’: possible worlds were used to analyse knowledge, belief, information, meaning, content, essence, supervenience, conditionality, and more. They entered the toolbox of logicians, semanticists, computer scientists, game theorists: a success story of philosophy. Already in 1986, in On the Plurality of Worlds, Lewis called possible worlds ‘a philosophers’ paradise’. (Well… He meant, possible worlds as he understood them. I’ll get back to this.)

Impossible worlds are considered esoteric these days, but I suspect they will earn their keep by helping with some troubles of the possible worlds analyses of those concepts. Arguably, such concepts are not merely intensional but, as people say, ‘hyperintensional’, which means that they require distinctions more fine-grained than the ones the standard apparatus of possible worlds can represent. For instance, it has been proposed that sentences mean the same thing when they express the same proposition, understood as a set of possible worlds: those where the sentence is true. Assuming that logical necessity is unrestricted and Kripke is right (usually a safe bet), ‘Muhammad Ali is Cassius Clay’ and ‘If Obama is funny, Obama is funny’ express the same proposition: the total set of worlds — for they are true across all of them. But how can they mean the same? Only one is about the guy who won the Rumble in the Jungle. Only one, arguably, is knowable a priori.

Impossible worlds can help making distinctions between necessary equivalents: add a world compatible with all one can know a priori, but where Ali is different from Clay, and you will distinguish those meanings. Such a world is metaphysically impossible: a way things could not have been.

There are many other approaches to hyperintensionality: truthmaker semantics, situation semantics, theories of aboutness, 2D semantics, transparent intensional logic, justification logic… These (families of) theories may be variously translatable into each other, e.g., some things that have been called ‘impossible worlds’ in the literature look very much like truthmakers or situations of a certain kind. It’s a burgeoning field with lots of open questions.

Here’s a philosophical issue raised by worlds: what are these things? In the aforementioned book, Lewis recommended that we stick to his version of the paradise and take possible worlds as a sort of disconnected spacetimes, which represent the possibility that x is F by having a real x which is really F as a part. That (assume) there could be talking donkeys is represented by some disconnected spacetime’s hosting real chatty donkeys. I think this won’t work for impossible worlds, for a number of reasons (though Ira Kiourti has defended, very nicely, the idea that one can have Lewisian impossible worlds). But one can take worlds, possible and impossible, as things which model possibilities and impossibilities in some other way, e.g., as language-like entities which represent in whatever way language represents. Suggestions of this kind have a number of troubles in their turn (Lewis called them ‘paradise on the cheap’): I say more in a forthcoming book I’ve been working on with Mark Jago, Impossible Worlds. I suspect, anyway, that whatever metaphysical issues are raised by worlds, they won’t be very different from those that will be faced by many other approaches to hyperintensionality: what are truthmakers, situations, etc.

3:AM: Are impossible worlds ruled by paraconsistency or dialetheism of some sort or could you have a notion of contradiction in such a world – an absolute notion – that even a dialethist couldn’t handle? (You might quickly sketch what paraconsistency and dialethism are here too!)

FB: Paraconsistent (e.g., relevant) logics, are logics where contradictions don’t have arbitrary consequences (‘A and not-A’ doesn’t ‘explode’, delivering arbitrary B’s), as it happens in classical logic and also in some non-classical logics, such as intuitionistic logic. Dialetheism is the view that some contradictions are true.

One who thinks that ‘A and not-A’ shouldn’t entail arbitrary B’s (perhaps also, that A shouldn’t entail ‘If B then B’, or ‘Either B or not-B’, for arbitrary B’s), given that B may have little to do with A, can endorse some paraconsistent logic without believing that anything of the form ‘A and not-A’ could ever be true. On the other hand, believers in true contradictions (I know several of them: Graham Priest, JC Beall, Dave Ripley, Koji Tanaka, Zach Weber, and more) should endorse some paraconsistent logic if they don’t want to be committed to everything being true. (I know only two people who think that: Paul Kabay, and the guy I did my PhD with, Vero Tarca — nomen omen: ‘Vero’ means ‘True’ in my tribal tongue.)

One of the characterisations of impossible worlds from the literature takes them as logic violators: given some logic L, they are worlds where some L-theorem or L-validity fails. That doesn’t rule out their complying with some other logic. Paraconsistent logics are weaker than classical logic, but they usually have, for instance, the entailment from ‘A and not-A’ to A. One may have worlds that are impossible relative to classical logic, but compatible with (closed under) some paraconsistent logic.
I did believe there to be contradictions even dialetheists couldn’t handle, and that one could motivate them via complicated metaphysical considerations. I didn’t take this to be a big issue for dialehteists, for I thought they’d reject commitment to the truth of those specific contradictions anyway. That’s a project I’ve given up on, some time ago. But I may get back to it.

3:AM: How do you see the relationship between impossible worlds and the logic of the imagination? Are you understanding imagination in terms of something like Chalmers’ ‘positive conceivability’?

FB: One good thing about Chalmers’ notion of positive conceivability, which seems to me close enough to a notion Stephen Yablo also wrote about, is that it is cashed in terms of representing a situation. Conceiving that P, so understood, is not just the tokening of a sentence in one’s mind. It is, rather, the act of mentally representing a situation that makes P true.

One may hope to provide some semantics for ‘Conceiving that P’, which resorts to worlds, for a widespread strategy in epistemic and doxastic logic has it that intentional notions like believing, knowing, being informed that, can be understood as restricted quantifiers over possible worlds (how the restrictions should work is a tricky issue). And if, against what Hume once said, we can conceive the absolutely impossible, impossible worlds can help us to model the situations making true the impossible P’s we can conceive.

Chalmers’ view of positive conceivability, it seems to me, doesn’t quite settle the issue of what kind of mental representations are involved here. Both Yablo and Chalmers speak about imagination, but ‘imagining’ (just like ‘conceiving’) is highly ambiguous: people use the term to refer to free mental wandering, daydreaming, hallucinating, thinking counterfactually, supposing. The borders between these mental activities seem to often be fuzzy in their turn, and what mental representations they involve is a tangled issue. I think that some input from cognitive psychology may help philosophers with it. I’m not an expert myself, but I’m studying — and now I have good researchers helping me.

3:AM: Does it follow that you think we can imagine logical impossibilities like, for example, an empty box with something in it? Is Hume wrong to think that conceivability entails possibility and if so how can we conceive of absolute impossibilities?

FB: I believe Hume is essentially wrong, but the issue is tricky due to the aforementioned ambiguity of ‘imagining’ and ‘conceiving’, and the connected issue of how mental representations represent. It’s the so-called ‘imagery debate’, one of the most intractable problems in the philosophy of cognition. There are four main candidate answers on the market, I think: mental representations represent (1) linguistically (in whatever way language represents, without the intervention of sensory modalities); (2) pictorially (in whatever way pictures represent, via quasi-spatial features and the involvement of sensory modalities); (3) a mixture of (1) and (2); (4) magically (they just represent, with no further questions asked on how they do it).

Some cognitive psychologists, like Allan Paivio, have defended a ‘dual coding’ theory of mental representations whereby some represent linguistically, some pictorially. People like Kosslyn and others attach great importance to representations of kind (2); others, like Fodor or Pylyshyn, claim that it all reduces to (1): they need not deny the phenomenological evidence of our entertaining pictorial imagery in our mind, but they claim that what does the representing work has little to do with imagery.

I’m neutral on (1)-(3). But I have an argument by cases. Roughly: if the conceivability of P at issue for Humeans involves representation of kind (1), we can conceive the impossible, for (I argue) impossibilities can be represented by meaningful bits of language. If the conceivability of P amounts to the pictorial imaginability of a situation verifying P, then the issue is whether this works purely qualitatively, i.e., only via phenomenological resemblance with the imagined scenario. If we are in a pure case (2), the range of situations imaginable in this way is too limited to have a significant role in modal epistemology. If not, and we have a case (3) of mixture of the linguistic and the pictorial, conceiving will involve some arbitrary labeling-linguistic component, which turns out to be sufficient to peek at the impossible. And if conceiving is neither pictorial nor linguistic, it is not clear what it may be. I just don’t buy the invocation of representational magic, as per (4).

3:AM: How do your views link with fictionalism?

FB: Nearly in no way, unfortunately. That’s due to my ignorance. For a while, I’ve only had very superficial knowledge of fictionalist views. So, while I didn’t find them prima facie convincing, that was a very big prima. At some point I stumbled upon Yablo’s ideas while studying his wonderful book, Aboutness, and I found his story on ‘noncatastrophic presupposition failure’ intriguing. But I didn’t get into the details in that case either. I am aware that fictionalist strategies in metaphysics are burgeoning, but I haven’t worked on metaphysics, ontological commitment, and the like, for a while. I’m more into logic these days. I wish I had more time to read stuff.

3:AM: Another strange area of philosophy you’re deeply involved with is noneism or Meinongianism. Before looking at the philosophical issues this raises could you sketch for us what noneism is?

FB: People often call ‘Meinongians’ (from the Austrian philosopher Alexius Meinong) those philosophers who claim that some things do not exist (in the metaphysics of modality, people call ‘possibilists’ those who claim that some things only exist at worlds different from the actual one, and thus don’t actually exist. I won’t get into the links between Meinongianism and possibilism here).
Given this characterisation, there’s a number of Meinongians, or former Meinongians, in contemporary philosophy: Terence Parsons, Ed Zalta, Richard Routley, Dale Jacquette, Filippo Casati, Naoya Fujikawa, Graham Priest, Tim Crane (I believe), Suki Finn (I think), Maria Reicher (I think), Jacek Pasniczek (I think), and more I can’t remember off the top of my head.

Meinongians are opposed to Quineans, who, on the contrary, claim that everything exists. (Are they really opposed? Do they actually make contrary claims, or are they taking past each other? I’d like to get back to this.)

Noneism is a kind of Meinongianism, defended mainly by Routley and Priest. One may be a Meinongian pluralist about existence, and think that those things that do exist, can exist in different ways (probably this was Meinong’s own view). Noneists are not like that: they believe there’s only one way of existing. They gloss it as the having of causal features, or the potential for causal interactions. Anna Karenina, my merely possible sister, and the exponential function, then, are all nonexistents. Routley thought the only existing things to be those that exist now — he was sort of a Meinongian presentist. In this view, also Socrates is, currently, a nonexistent.

3:AM: Now the Quineans amongst us think that there are knock-down arguments against this position. Could you sketch for us what the thrust of their objections are before you explain why you disagree with them?

FB: Quineans have invoked the motto ‘No Entity Without Identity’, and claimed that nonexistents lack identity criteria specifying the conditions for such objects to count as the same thing, or rather different ones. Meinongians have answered by deflating identity criteria, or by providing identity criteria for their nonexistents. Such criteria have been criticised in their turn. It’s an ongoing discussion (thus, no knock-down: when someone knocks you down, that stops discussion — at least for a while).

Quineans have also invoked selection problems (these are epistemic issues, unlike metaphysical ones concerning identity criteria): how can we pick out a nonexistent and say true things about it, since we can’t single it out via causal interactions, and we often seem to lack uniquely identifying descriptive conditions? Meinongians have replied by deflating issues of selection, or by providing uniquely identifying descriptive conditions, or by pointing at intentional phenomena or theories of arbitrary reference that may do the trick. Such attempts have been criticised in their turn. It’s another ongoing discussion (no knock-down here either).

Quineans have also claimed that Meinongians misunderstand the concept of existence. This is the topic I’ve been most interested in.

3:AM: So why aren’t you convinced that nonism is a dead issue – at least in the terms given by the Quineans?

FB: I think nowadays few are convinced that noneism, or Meinongianism, is a dead issue. (Of course, those who ignore the literature have just heard stories about Russell killing Meinongianism and Quine throwing salt on it.) I do think that Meinongianism is full of open problems, though — just like Quineanism.

Meinongianism comes in a variety of theories, often very different from, and incompatible with, each other — just like Quineanism again. Those who believe that existence is what Quine said it is, are divided between presentists and eternalists on the nature of time and its occupants, realists and nominalists on abstract objects, friends and foes of identity criteria, supporters and enemies of colocated objects, etc.

So are Meinongians divided: some claim that some singular terms of ordinary language can lack denotation (others denote existents, others yet, nonexistents), some want all terms to denote. And their nonexistents may come in as rich a variety as Quinean existents: each Meinongian I know of accepts that fictional characters, like Sherlock Holmes, qualify. Some add possibilia, like my merely possible sister. Some add ideal objects and objects postulated by false scientific theories, like frictionless planes or Leverrier’s Vulcan (the planet), or mythical objects like Vulcan (the god). Some, like Routley, also add past existents that currently are nonexistent, like Socrates, and perhaps future existents, like the first newborn of the XXII Century. Towards all of them, Meinongians may take a realist stance by claiming that these things are what they are, and have the intrinsic features they have, largely independently from us. Or, Meinongians may claim that some of them, like fictional characters, metaphysically depend on the intentional activities of the authors of the relevant fictions. Meinongians may opt for this or that criterion of identity for objects, or have different criteria for things of different kinds, or – again, like some of their Quinean mates – just despise the philosophical plea for identity criteria. Each of these Meinongian options will have its pros and cons when it’s about intuitiveness, explanatory power, simplicity, or bullets to bite – just as happens for the many, different, and reciprocally conflicting Quinean ontologies.

3:AM: So what is the ontology of Meinongianism? Does it imply that existence is a real property in some way?

FB: The core issue — the one I have been postponing so far — is indeed whether existence is a real property or not. One calls ‘real’ a property that makes a difference: some things have it, others lack it.

It is here that the original split occurs.

According to Quineans, the concept of existence is captured by the quantifier. It is sometimes claimed that Quineans (should) deny that existence is a property, or that it is a property of individuals. That’s a mistake, as Quineans like van Inwagen pointed out. Quineans can have their existence property — a property of individuals. That’s the property of being something. It just isn’t a real property: anything is something.

For Meinongians, existence is a real property. That can’t be right for Quineans: to claim that some things do not exist is to claim that some things are such that there are no such things, which is preposterous.

Thus, the core question is: why should ‘exists’ mean this rather than that?
(As a side remark, notice that one may endorse a certain pluralism on existence, and claim that some senses of ‘exist’ indeed express quantificational concepts, while others don’t. But Quineans have traditionally claimed that there is no way ‘exists’ can ever express a real, non-quantificational feature of individuals.)

3:AM: Van Inwagen would argue that Meinongians are guilty of the fallacy of attributing to the being of a thing what properly belongs to its nature. So what would you say to that kind of objection? Are you making that move, or would that be to misunderstand what is going on in your thinking?

FB: I would say that Quineans are guilty of the fallacy of attributing to the being of a thing what belongs to its being a thing. I find being something a nice feature (at least to the extent that being self-identical, or being either abstract or not, are). But I object that it either cannot be, or cannot always be, what ‘exists’ means.

I don’t mean that ‘exists’ (or, its counterparts in other natural languages), in the Quineans’ mouth, means something else from what it means in the Meinongians’ mouth. Van Inwagen is right on this, too: ‘exists’ means whatever it means in English. Meanings are largely determined in connection to use (this is a very vague claim — making it more precise would require too much space). Quineans and Meinongians can, in general, use English competently. ‘Exists’ won’t switch meanings between the Meinongians’ and Quineans’ mouths. And if it means what the Quineans say it means, then it means that also in the Meinongians’ mouth although, according to their mistaken view about the meaning of ‘exists’, ‘exists’ doesn’t mean that.

But if the vague claim that meaning is largely determined in connection to use (or rather, some precisification of this) is right, then, to assess what ‘exists’ may and may not mean, it makes sense to look also at how that word works in English. And once one looks, one will see that Meinongianism is in good shape. For instance, in The Semantics of Existence the philosopher and linguist Friederike Moltmann has explored in detail the topic of the title, and one of her conclusions is that ‘exists’ often acts as a first-order extensional predicate in sentences whose semantics is very different from the one of quantificational or there-sentences. That’s bad news for the view that the concept of existence is just captured by the quantifier.

3:AM: Is reality digital or analogue or, as Floridi argues, are these not features of reality at all?

FB: It’s either digital or analogue. One who thinks that this is too trivial a claim to be worth making should read Floridi’s nice works, according to which discreteness and continuity are features of modes of presentation of reality, not of reality itself. The view resembles Kant’s stance on a cosmological antinomy in the transcendental dialectic of the Critique of Pure Reason. In ‘ingenuous realist’, pre-Kantian metaphysics, some defended the view that reality can be decomposed into ultimate, discrete units. Others claimed that reality is continuous and divisible in infinitum. Kant has it that both parties are mistaken: they want to speak of the world as a whole, as it is in itself, thereby going beyond the limits of scientific knowledge established in the transcendental analytic.

I’m on the side of ingenuous realists on this. One may say that this is also the side of science: some people working in fundamental physics wonder whether the structure of spacetime is discrete once one gets to the Planck scale. While our best physical theories do not settle the issue either way, practitioners make perfect sense of the question.

I say that if one claims that continuity or discreteness cannot be features of reality, reality ends up in a bad shape, e.g., it has no parts and is unextended, or identity must work in peculiar ways: there must be things out there whose identity is vague, or identity must be a relative notion. Of course, one’s modus tollens can be another one’s modus ponens.

3:AM: And can the concept of identity fail to apply to somethings even when we can count them?

FB: If identity is objectual, non-vague, Leibnitian (entailing congruence with respect to all properties) — it can’t (a number of people have argued this before me, including Otávio Bueno and Matteo Morganti; a great book on the topic is French & Krause’s Identity in Physics).
Some say there’s an ambiguity in counting. One may take it to involve arranging the stuff in a list with a first item, a second, a third… Or, one may take it to involve assigning a cardinal to a bunch: ‘Three things here’. And some believe that the concepts of identity and difference may fail to apply to things we can’t count in the first sense. I think facts of identity and difference can supervene on facts of cardinality. Those who believe otherwise have surrendered to empiricism: they focus on permutations of those indiscernible particles of quantum mechanics, this overwhelmingly successful bit of natural science, and lose their nerve. I won’t be shaken from my metaphysical armchair.

3:AM: And finally, for the readers here at 3:AM, are there five books you can recommend to take us further into your philosophical world?

FB: I’ll do it the Italian way: I’ll recommend friends.

Crispin Wright, Truth and Objectivity

Diego Marconi, Lexical Competence

Friderike Moltmann, Abstract Objects and the Semantics of Natural Language

Graham Priest, Beyond the Limits of Thought

Greg Restall, An Introduction to Substructural Logics

Richard Marshall is still biding his time.

Buy his new book here or his first book here to keep him biding!

End Times Series: the first 302

First published in 3:AM Magazine: Saturday, February 10th, 2018.