Not and Other Metalinguistic Stuff
Interview by Richard Marshall.
Mahrad Almotahari is an Assistant Professor and a member of UIC’s Laboratory of Integrative Neuroscience. He earned his PhD from MIT in 2011 and specializes in the philosophy of language and metaphysics. He was the winner of Philosophical Quarterly’s 2013 Essay Prize for his essay entitled “The Identity of a Material Thing and its Matter.” Here he thinks on why direct reference isn’t incompatible with physicalism, on the relationship between language and metaphysics, on his defence of monism, on his disagreement with Kit Fine over this, on transparency, on material objects and whether metaphysical matters boil down to linguistic ones, on why even if they do this doesn’t trivialise the issue and finally on whether anything in ethics is just plain good. Here comes another young turk on the metalinguistic block… zoom zoom…
3:AM: What made you become a philosopher?
Mahrad Almotahari: I think the first time I was really exposed to philosophy was when I was about 14 or 15. I had just joined the speech and debate team at my high school and an older student recommended that I read Sophie’s World. I was told that familiarity with important philosophical ideas would likely be useful for debate purposes. So I read the book. I think I enjoyed the chapter on Bishop Berkeley the most. I doubt that I understood it, but it made an impression. In any event, I had some sense of what philosophy is when I went to college, but I started off as a biology student. I switched to philosophy at the start of my third year after having taken several philosophy classes the year before. My undergraduate philosophy teachers were awesome. I aspire to be like them.
3:AM: For you direct reference isn’t incompatible with physicalism, but could it have been? If it could have been then could a seemingly semantic question have become a metaphysical one?
MA: Right. I think direct reference theory and physicalism are compatible. The former says (roughly) that the sole semantic contribution of certain kinds of expressions is their referent. The latter says (again, roughly) that all of the facts “supervene” on the physical facts. In other words, once God fixed the physical facts about our universe, she didn’t have to do anything more to fix all of the facts. She could call it a day and punch out. I wrote a paper in graduate school with my friend, Damien Rochford, defending the compatibility of these two doctrines from an interesting and fun-to-think-about argument due to Thomas Hofweber. I find it hard to understand how doctrines that are compatible could have been incompatible. I tend to think of propositions as having their truth-conditions essentially. But, in any event, the semantic question about the theory of direct reference is related in at least one interesting way to the metaphysical question about physicalism. Some philosophers have thought that directly referential terms have a certain property (“rigidity”) that strengthens the case for dualism, which is incompatible with physicalism. I’m not convinced by this case, but it does illustrate the potential for semantics to interact with metaphysics in unexpected and surprising ways. I’m inclined to think that this potential is sometimes realized. But there are philosophers who strongly disagree. They believe metaphysics is far less constrained by the study of natural language than I do, and treat my position as an instance of “Picture Thinking” or “the representational fallacy”.
3:AM: This interaction between natural language studies and metaphysics is illustrated in your defence of monism. So can you first sketch out for us the metaphysical doctrine of monism that you’re interested in?
MA: Sure. I’ve argued that material objects are identical with their constituent hunks of matter. They’re literally one and the same. Hence “monism”.
3:AM: Kit Fine attacks this view—could you say what his main objections are?
MA: He has. And so have others. The central arguments pretty much always take the same shape: material objects have properties that their constituent hunks of matter don’t have. For example, a door might be shut, though the piece of wood out of which it’s made isn’t shut. If the door and the piece of wood were literally one and the same, they should have all of the same properties. Since they don’t, they’re distinct. This way of arguing is pretty common in many areas of philosophy. And there are familiar ways of responding. What’s interesting about Fine’s discussion is that he evaluates these familiar responses and offers powerful considerations in favor of thinking that they won’t work against his particular objections. He’s also developed a pretty detailed rival theory of the nature of material objects.
3:AM: So you think Fine is assuming a linguistic environment of a certain kind that does much of the metaphysical lifting so to speak. You say the environment is not ‘transparent’. So could you show what you mean by this and why, if this is accepted, Fine is taking a metalinguistic view to dismantle the metaphysical view of monism?
MA: Transparency, in the relevant sense of the word, is a property of certain linguistic contexts. It’s a property they have if they permit the substitution of co-referring terms. For example, before he encountered Darth Vader in Cloud City, Luke Skywalker knew that Anakin Skywalker was his father. He didn’t know that Darth Vader was his father. It came as quite a shock to him, after all, when he learned the sad truth. But ‘Darth Vader’ and ‘Anakin Skywalker’ refer to the same man. So the context ‘Luke Skywalker knew that _ was his father’ fails to be transparent. Only transparent contexts license the sort of inference needed to challenge monism. One might question whether all of Fine’s arguments involve transparent linguistic contexts. In an early response to Fine, Jeff King pursues this strategy much more seriously than I do. I raise the possibility that Fine’s arguments fail because they equivocate between two different interpretations of the word ‘not’. On the interpretation needed to secure the truth of the premise, ‘The piece of wood is not shut’, ‘not’ is being used “metalinguistically”. It’s being used to say that there’s something objectionable about the linguistic trappings of the thought one is trying to express, not to assert its falsity. So it’s like the use of ‘not’ in the sentence, ‘Louie C.K. isn’t funny; he’s hilarious’. But, on the metalinguistic interpretation of ‘not’, the conclusion that the door is distinct from the piece of wood doesn’t follow.
To compare, suppose certain linguistic conventions require that when speaking about papal decrees issued ex cathedra we use the Pope’s official name, ‘Francis’. Suppose these conventions dictate that it would be objectionable to use ‘Jorge Bergoglio’ to express the same thought. Now imagine that some ignorant Catholic who takes these conventions seriously reasons as follows: Pope Francis issued a decree ex cathedra, but Jorge Bergoglia didn’t; so they mustn’t be the same person. The thing to say in this case is that ‘not’ is being used metalinguistically in the premise, ‘Jorge Bergoglia did not issue a decree ex cathedra’. It merely says that there’s something objectionable about how a certain thought is dressed up. It isn’t being used to challenge the truth of that thought. I came across an example like this one in an interesting article by Benjamin Schnieder, which piqued my interest in metalinguistic negation and got me thinking about how to properly detect it. I then realized that some of the more reliable tests for determining when ‘not’ is being used metalinguistically indicate that Fine’s argument is similar to our imaginary Catholic’s. One quick consideration in favor of my position is that metalinguistic uses of ‘not’ can’t be replaced with negative prefixes. Infelicity results. So, for example, consider how bad it sounds to say, ‘Louie C.K. is unfunny; he’s hilarious’. Now, to my ear, ‘The piece of wood is unshut’ is just as bad as ‘The piece of wood is shut’. We should expect this infelicity if the use of ‘not’ in the relevant premise is metalinguistic. Finally, I’ve argued that if you combine a proper understanding of how metalinguistic ‘not’ works with two or three other observations, you obtain a pretty direct argument for monism. A solid defense often results in a winning offense.
3:AM: You’ve discussed the metaphysics of material objects and there you’ve shown how the role of ‘not’ works. Could you sketch your theory and say whether debates about material objects boil down to disputes about language—and is this a reason for thinking less of them as some might argue—“It’s just a verbal dispute”—or are such disputes substantial?
MA: Well, I’m a little hesitant to call it my theory, since the core observation I rely on was originally articulated by Laurence Horn, but I’ve defended it from recent criticisms and drawn attention to some independent considerations in its favor. In its roughest form—and ignoring certain important complications—the observation is basically that an objectionable sentence can be metalinguistically negated only if a straightforward denial of the truth of what the sentence can be used to assert is incompatible with what its unobjectionable counterpart can be used to assert. So, to return to my earlier example, one might find ‘Louie C.K. is funny’ objectionable because it isn’t as emphatic as the occasion demands. Its unobjectionable counterpart, then, is something like ‘Louie C.K. is hilarious’. Saying ‘Louie C.K. isn’t funny’ and thus denying the truth of the thought that Louie C.K. is funny is certainly incompatible with asserting that Louie C.K. is hilarious’. So Horn’s observation says that one can metalinguistically negate ‘Louie C.K. is funny’. What’s nice about the observation is that it explains why a metalinguistic interpretation of ‘not’ is unavailable in contexts where one might initially expect it to be. I think Horn’s observation, plus some plausible additional claims, can get us to monism.
But there was another part to your question that I haven’t addressed. You asked whether the debate about material objects is a verbal dispute. I’m inclined to think that this particular debate—whether material objects are hunks of matter—is a verbal dispute. That would explain why linguistic considerations allow us to resolve the matter. My point is methodological. In his introduction to The Linguistic Turn, Richard Rorty wrote that “it would indeed be hard to know what methods a philosopher ought to follow without knowing something about the nature of the philosopher’s subject matter”. But sometimes we’re more confident in the reliability of certain methods—in their having produced knowledge—than we are in the nature of the subject matter under consideration. Maybe the best example of this sort of situation is mathematics. I’m much more confident that the methods of mathematicians produce knowledge than I am about the nature of their subject matter. In this sort of situation, it seems to me that we should adapt our conception of the relevant subject matter to explain the reliability of our method. That’s why I’m inclined to treat the relevant debate about material objects as a verbal dispute. I’m adapting my conception of the subject matter of the dispute to fit the method that seems to move us forward. But this isn’t to downgrade the significance of the debate, since I think it’s related to a number of issues which, taken together, render the question substantive.
3:AM: Nietzsche and others have said that all metaphysics is messed up if the language it’s couched in is taken to underwrite metaphysical reality. For example, where language has a subject/predicate distinction the metaphysician shouldn’t assume that this mirrors metaphysical actuality. So how is metaphysics possible? Isn’t every theory, even something as seemingly obvious as Leibniz’s law of identity, worryingly embedded in a metalinguistic perspective?
MA: Every theory—or every theory we’ve bothered to articulate—is spelled out with the help of some language or other. If that’s all you mean by every theory being “embedded in a metalinguistic perspective”, then I don’t see any reason to worry. Taking a theory to be true needn’t involve any assumption about linguistic structure resembling metaphysical structure. Maybe there’s some additional epistemic virtue a theory can have if, on top of being true, it’s couched in a language whose vocabulary “mirrors” the structure inherent in the world. This is an old idea, and some recent philosophers find it very attractive. If the idea entails that analytically equivalent theories can differ with respect to their epistemic virtues, since one might be formulated with the help of a vocabulary whose basic expressions better mirror the structure of the world than the basic expressions of the other, then I find this old idea implausible. It puts too much emphasis on the format of a theory. I’m inclined to think that epistemic merit is determined by what theories say, not how they say it. To that extent, then, I would agree with Nietzsche, as you describe him.
Of course, there may be practical reasons for preferring certain theories on the basis of their format. Personally, I would never formulate any theory of mine with Helvetica. But I haven’t answered your most difficult question: how is metaphysics possible? I’m not ashamed to admit that I don’t really know, but I’m skeptical of answers that try to solve this problem by representing all successful metaphysical inquiry as fitting into a certain mold. Sometimes there’s disagreement about the shape of the mold. One philosopher might say that metaphysical knowledge is knowledge of the conditions of possible experience; another might say that it’s knowledge of the arbitrary semantic decisions we’ve made in setting up our language; yet another might claim that its knowledge about a logically perfect language. My colleague, Dave Hilbert, believes that colors are surface reflectance properties. His work falls squarely within metaphysics, and I think he’s probably right about the nature of color. But Dave isn’t making a claim about the conditions of possible experience, nor is he analyzing the meaning of ‘color’, and he’s certainly not making a claim about how colors would be represented in a logically perfect language (whatever that might be). His work is very closely allied to vision science and physics. My hunch is that we make progress in metaphysics by observing connections with closely related fields where the “rules of the game” are better understood, teasing out consequences and seeing where they lead. The process is messy and resists generalization across the whole of metaphysics. The more connections of this sort we can discern, the more confident we can be that the enterprise is meaningful, substantive, and important.
3:AM: Phenomenal consciousness is an area where some of this kind of worry rears its head. Take qualia—some like Richard Brown think it’s just crazy to deny qualia, whilst others like Dennett say we don’t have qualia. This seems to be a debate about whether investigating consciousness can ever be done properly because if the best scientific neural theories deny the existence of data that many would think are the data needing to be explained (e.g. qualia) then won’t we always be at a kind of impasse?
MA: I hope the science of consciousness hasn’t reached an impasse. For what it’s worth, I’m definitely conscious. If only you knew what it’s like being me!
3:AM: You’ve entered the battle over the notion in ethics of whether anything is just plain good. Is the problem here one where arguments get bewitched by linguistic forms that may distort actuality—so the problem boils down to language? And is anything just plain good?
MA: This was another project in collaboration with a close friend, Adam Hosein. We’re attracted to the view that nothing is just plain good. It’s not a new view. A number of philosophers have argued for it, including Peter Geach, Philippa Foot, and Judith Jarvis Thomson. It can even be traced back to Aristotle. To clarify the view, it might help to consider an analogy. Suppose someone says, “There are fake dollar bills, fake Rolexes, fake barns, and fake democracies, but in addition to these familiar fake items there are also things which are just plain fake—not fake as an instance of some more general kind of thing, like a dollar bill or a Rolex, but fake period.” It’s hard to understand what this person is getting at. The appropriate way to respond to this person, it would seem, is to point out that ‘fake’ isn’t the sort of word that, in isolation, expresses a property which a thing can just plain have or lack. This isn’t to deny that we sometimes use the word ‘fake’ in the way that we use adjectives like ‘red’ and ‘round’. We might say ‘Ted Cruz is fake’, just as we might say ‘Roses are red’. But this is an idiomatic way of saying that Ted Cruz is insincere. The truth of ‘Ted Cruz is fake’ doesn’t require that there be a property of being simply fake. There isn’t any such property. Geach, Foot, and Thomson think that ‘good’ is like ‘fake’, and just as there are particular ways of being fake, though there isn’t a way of being just plain fake, there are particular ways of being good—for example, being a good liar, or being good to eat, or being good for our democracy—though there isn’t a way of being just plain good. Lots of philosophers have criticized the rationale for this view, which I haven’t gone into here. Adam and I believe that there’s a way of arguing for the view which bypasses the familiar problems that people have identified. Our argument does rely on certain linguistic claims, but sadly we haven’t bewitched anyone. (As far as we know.)
3:AM: And for readers here at 3:AM, are there five books you could recommend to take us further into your philosophical world?
MA: A Natural History of Negation, by Laurence Horn; Normativity, by Judith Jarvis Thomson; Semantic Relationism, by Kit Fine; and both volumes of Steve Yablo’s philosophical papers: Thoughts and Things.
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, March 12th, 2016.