Thomas Nagel’s natural teleology

I recently read Thomas Nagel’s mercifully short Mind and Cosmos — mostly just to see what all the fuss was about — and one of the weirder lines of enquiry he pursues concerns what he calls the ‘historical problem of consciousness’. This is the problem of how consciousness came to be, which Nagel wants to distinguish from the ‘constitutive problem of consciousness’, i.e. the problem of what consciousness is. He writes [1]:

The historical account of how conscious organisms arose in the universe can take one of three forms: it will either be causal (appealing only to law-governed efficient causation), or teleological, or intentional.

A causal explanation is the sort of thing a Darwinian account of the evolution of consciousness might offer, while an intentional explanation would appeal to the plans of some agent who made things happen the way they have (guess who). Teleological explanation is murkier, and it is here that Nagel positions himself, attempting to stake out some sort of middle ground between theism and materialism.

Teleological explanations of a sort do appear in science — particularly in biology — but usually with the proviso that they are to be understood as shorthands (or sometimes heuristics) for lower-level causal explanations. So we might explain the web-building behaviour of spiders in terms of its purpose (i.e. spiders build webs in order to catch flies), but on the understanding that the ‘purpose’ of an organism’s behaviour just refers to those of its causal effects which increase the organism’s inclusive fitness.

Even intentions crop up: when a gene has effects which result in its spread through a population (altruistic behaviour in cases of kin selection is a favourite example) it can be helpful to think of the gene as an agent with goals. But this kind of agenthood is underpinned by the fact that genes have causal effects which can influence their own spread, and so can be regarded as beneficiaries of these effects. Genes are agents, then, only insofar as their interactions can be modelled using game theory, and this is a criterion which can be cashed out in purely causal terms.

It can also be tempting to ascribe teleological features to systems which have intrinsic tendencies. So we might say that soap bubbles ‘want’ to become spheres, or (more generally) that dynamic systems ‘want’ to minimise their free energy. But here, as in the above cases, this can be understood in terms of patterns of efficient causality at lower levels which give rise to emergent features at higher ones. (Rather than the spread of traits through populations, in this case the causal explanation might mention singularities in a system’s phase space, for example.)

None of this is quite what Nagel is after. He wants some form of teleology which neither can be paraphrased away using efficient causes, nor which requires a real agent. This jams him into a rather narrow crease, and forces him to say things like “Each of our lives is a part of the lengthy process of the universe gradually waking up and becoming aware of itself” while unable to say anything substantial about what it means for the universe to be the sort of unity which can ‘wake up’. The faint whiff of woo that rises from such statements no doubt accounts for much of the scorn that has been poured on this book since it was published in 2012.

And perhaps not all of it fairly — while Nagel is not quite sure what exactly he is proposing (he admits as much, so props to him) he does have some idea what the presence of ‘natural teleology’ would imply, and that may be enough [2]:

Natural teleology would require […] that the nonteleological and timeless laws of physics […] are not fully deterministic. Given the physical state of the universe at any moment, the laws of physics would have to leave open a range of alternative successor states, presumably with a probability distribution over them.

And then [3]:

The existence of teleology requires that [those] successor states [which lead to the formation of more complex systems, and ultimately life] have a significantly higher probability than is entailed by the laws of physics alone […] Teleological laws would assign higher probability to steps on paths in state space that have a higher “velocity” toward certain outcomes.

These implications offer a way of empirically testing for natural teleology in indeterministic transitions. If, say, indeterministic quantum effects played a critical role in gene mutation, and we suspected that this is somewhere that natural teleology might manifest, we could gather measurements of these events and see whether the spread of our results deviated from the expected probability distribution given by quantum mechanical laws. Being a wise sort of chap Nagel never mentions quantum mechanics, but we can see how this kind of test might work even if natural teleology was hypothesised to operate at some as-yet-undiscovered lower indeterministic level.

If quantum indeterminism is supposed to be a site of teleological influence, then we could be forgiven for wondering why Nagel isn’t out there doing experiments to look for it. But if it isn’t, then Nagel’s claim puts strong constraints on the shape of future physics. All this just serves to highlight something that should already be clear: Nagel’s thesis is highly speculative, and whether or not it should be given serious consideration depends entirely on the hopelessness of alternatives. The subtitle of Mind and Cosmos is the snappy “Why the Materialist Neo-Darwinian Conception of Nature is Almost Certainly False”, so this is not a question he leaves unaddressed. But personally I found it all a bit toothless (I think a more accurate subtitle would have been “Why the Materialist Neo-Darwinian Conception of Nature is Almost Certainly Inconsistent with a Wagonload of Dubious Claims about Cognition and Moral Ontology, Surprise Surprise”). Here, however, is not the place to get into it — if you’ve read the book (or have just been following the fallout) and have some thoughts, please do drop in with a comment.

Notes:

1. Chapter 3, Consciousness.

2. Chapter 4, Cognition.

3. ibid.

A note on free will and semantics

I wish to dispute a claim made occasionally in relation to the ongoing philosophical discussion of free will: that the disagreement between those who I shall call ‘irrealists’ (who I take as believing that free will does not exist) and ‘compatibilists’ (who I take as believing that free will exists, but only in some sense which is compatible with determinism [1]) has no substantial content, and comes down to semantic preference. This seems wrong to me, and in this post I’ll attempt to locate the point of contention between these views. It’s worth mentioning that since it is the compatibilists who are at most risk of seeming obscure, locating this contrast is primarily in their interests. This point in mind, what follows can be read as a modest attempt to motivate compatibilism (though not to present any particular arguments in its favour).

One of the charges against compatibilism is that it redefines free will. The thought is that what most people mean by free will is ‘libertarian’ free will: the idea that choices can be made free from prior determination. This notion contradicts the deterministic idea that any given state of the world necessitates all future states (so all future events are pre-determined by past events, and the possibility of an agent being able to choose between future scenarios is illusory). If this is what free will means then to define it in a way which sidesteps this conflict is to remove the sense in which a willing can be said to be free. In short, the compatibilist stands accused of changing the subject.

This could be a valid charge. If I were to define ‘having free will’ as ‘having a nose’, I could happily say things like “I have free will” without saying anything about free will at all. But not all redefinitions are so empty, and sometimes there are good reasons for them; the issue does not concern redefinition as such, but ill-motivated redefinition. What is required from a definition of free will is some necessary and sufficient conditions (or something approaching them) which will help analyse situations in which free will is said to be exercised. These situations range from matters of inconsequential decision to cases in which uninhibited choice is deeply entwined with moral responsibility. Having a nose has nothing to do with any of that, so the redefinition can be identified as vacuous on account of its abandonment of the subject matter. What the compatibilist needs to do in order to respond to the charge is justify their redefinition.

A helpful example of well-motivated redefinition is provided by the word ‘atom’. As coined by Democritus, an atom is an indivisible unit of matter. In the 19th century when scientists postulated the existence of atoms to explain certain chemical phenomena, their theories succeeded because there really are tiny discrete entities which behave as their hypotheses required. We might say that the term ‘atom’ as used by these theorists began to track a real feature of the world.

Later it emerged that the atoms of atomic theory were not indivisible; the real feature they tracked had properties which did not line up with the Democritean connotation. The semantics of this situation can be thought of in several ways. On the one hand it could be said that ‘atom’ had changed its meaning and that a redefinition (from ‘indivisible base units’ to ‘the things referred to in those successful theories’) was warranted, because ultimately usage is the arbiter of meaning, and a definition should reflect usage (not vice versa). On the other it could be said that the meaning of theoretical terms is stipulated, so definition should here be regarded as fixing meaning. This would amount to saying that two distinct concepts came to be sheltered under the term ‘atom’.

It doesn’t matter which way we see it — the point to take from this is that a term can track a real feature of the world to some extent independently of what people think they mean by it. In the period before subatomic particles were discovered, if asked what an atom was one might have replied either with the Democritean answer or with an answer which made explicit reference to the new atomic theory. Either would have seemed fine, since it was presumed that these definitions picked out the same kind of entity. When it was discovered that this was false, the motivation was in place to do something: either redefine ‘atom’ by jettisoning its Democritean elements, or identify a new kind of atom.

The situation with free will is different to the situation with atoms in the 19th century, but some of the salient features of the latter can help illuminate some of the salient features of the former. One difference between ‘atom’ and ‘free will’ is that ‘atom’ is a theoretical term which was coined in the abstract and later came to be associated with a real feature, while ‘free will’ is a pre-theoretical concept — we deploy it and communicate successfully with it long before we have started to reflect analytically on what we mean by it, or encountered any of the philosophical problems associated with it [2].

This suggests two things. Firstly, it would be unwise to treat ‘free will’ as a concept with a stipulated meaning. Secondly, the real features of the world tracked by ‘free will’ (if they exist) may not be what people sincerely claim they are. So while it is important to ask whether what people think of as free will exists, the deeper question is whether pre-theoretical usage of ‘free will’ tracks any real feature of the world at all, and if so whether this feature has the properties required to properly underwrite such usage.

The concept of free will is typically used to distinguish a scenario in which a person holds moral responsibility for their actions from one in which they do not. Thus we might say that an opportunist thief steals freely while a pathological kleptomaniac does not, so only the former should be held morally accountable for their actions. If there is a real difference between these two cases, and it is that the opportunist exercises their libertarian free will in choosing to steal while the kleptomaniac’s is bypassed by their pathology, this would seem to provide us with the distinction required to justify our attribution of free will to the former but not the latter. The real feature tracked by the pre-theoretical ‘free will’ would be the theoretically captured ‘libertarian free will’. This is in many respects the natural way to think about things.

But if libertarian free will does not exist (as both the irrealist and the compatibilist believe) the question remains whether ‘free will’ tracks any real feature of world which can support its pre-theoretical usage. This is not a semantic question, and it is one which the compatibilist typically answers in the affirmative, using their claim that such a real feature of the world exists to justify their redefinition of free will. (What such a real feature might be is up for debate, and will have to mesh tightly with meta-ethics — see, for the popular example, Daniel Dennett’s spiel about ‘evitability’ [3]) This question splits irrealism two ways: either an irrealist can agree that ‘free will’ does track a real feature of the world which can underwrite its usage (but that this still does not warrant redefining free will to mean that real feature), or they can deny that it tracks any such feature.

If it is the first then the difference between them and the compatibilist really is just semantic: they would say that the meaning of ‘free will’ is fixed by definition as libertarian free will, while the compatibilist takes a more fluid approach. Given that on common assumption both agree that the usage of ‘free will’ can be supported by something real which isn’t libertarian free will, this leaves the irrealist in a rather weak position, akin to the Democritean who complained that the new atomic theorists were not talking about atoms at all. What’s more, this definitional purism seems artificial given the status of free will as a largely pre-theoretical concept. I take it, then, that this is not the position generally defended by irrealists, which brings us to the second option.

An irrealist about free will who denies that there exists some real feature of the world which supports free will (perhaps because they believe free will is compatible with neither determinism nor indeterminism) is saying something far stronger than that libertarian free will does not exist. On this version of the irrealist view there is no distinction at all between the kleptomaniac and the opportunist which could be couched in terms of free will without abandoning the subject matter (which primarily concerns whether moral responsibility can be attributed to the latter but not the former). This is the version that the irrealist should stand by if they wish to contrast their position with the compatibilist’s substantial claim that the concept of free will can be underwritten in a non-libertarian way.

The distinctions made here also help to sharpen the problem: the language of free will and moral responsibility is near impossible to get rid of, and this in itself is a datum to be accounted for. The compatibilist tries to do this by looking for real (if non-obvious) features of actions and actors that the language of free will hooks on to. Broadly speaking, the irrealist has two options: either they can argue that despite appearances the language of free will is empty, and can (perhaps should) be abandoned. Or they can adopt the milder stance: while there are no real features of the actions we describe as freely willed which make sense of those attributions, there may be features of us as attributers which do — to put it another way, they could argue that free will is a useful fiction.

Notes:

1. In it’s technical sense compatibilism need not commit one to any particular stance on the existence of free will, just on whether determinism precludes free will or not. But in this post I’m concerned only with compatibilists who believe in free will, so that’s how I’ve taken it.

2. A paradigm example of a theoretical term would be an abstract mathematical term like ‘differential operator’, which can be precisely defined but whose use is hard to master. At the other extreme we have words which are used constantly and unreflectively but provoke all manner of havoc when someone tries to articulate their meaning analytically — the philosophical favourite ‘to be’ is a good example. ‘Free will’ lies somewhere between these extremes.

3. Here he goes: Daniel Dennett — Free Will, Determinism and Evolution

Considering ontological nihilism

As Quine once noted, the question “what is there?” admits a truthful reply of just one word: “everything”. The glibness of this point illustrates a difficulty encountered when posing ontological questions. When the philosopher asks us if there are trees, we are inclined to say that of course there are. If the philosopher is frustrated by the ease with which this reply is given, it is likely because they sought to place a stricter meaning on the verb ‘to be’ than we have granted it. The question was not really “are there trees?”, but “are there really trees?”

Armed with this general caveat we can divest ontological nihilism — the view that nothing exists — of some of its surface angst by rendering it more faithfully as the view that no thing exists. In their paper on the topic [1], John O’Leary-Hawthorne and Andrew Cortens (who will from hereon in be assimilated into the unity known as ‘HC’) approach the idea (which they defend) by considering the various degrees of commitment we can hold towards objects in general.

1. Straightforward realism. The world comes to us “articulated into distinct objects” [2]. On this view the ontologically honest sentences are those which take the familiar grammatical form in which nouns are used as subjects. Example: “There is a pebble.”

2. The world consists of stuffs, not of things. Many of what we think of as objects are actually just partial regions of scattered stuff. The ontologically honest sentences are those that employ mass terms [3] which refer to these stuffs. Here the pebble example would more honestly be put as: “There’s a bit of pebble.”

3. Token monism. There is just one object, and that object is the world. Everything we perceive as an object is just some local modification of the world-stuff. (This view is associated with Spinoza and Parmenides. [4]) Our example becomes: “The world-stuff is pebblish there.”

The positions on this list become more ontologically innocent [5] as we ascend, in that they become committed to the existence of fewer and fewer objects. Realism treats all or most of the things we refer to with nouns as objects, the stuff view reduces this number dramatically, and token monism reduces it to one. Ontological nihilism is then the view which reduces this number to zero.

To see how this works it is helpful to consider which sentences would be ontologically honest according to nihilism. The paradigmatic example given by HC is “It is raining.” Sentences of this type are referred to as ‘feature-placing’ [6], and are distinguished by the fact that ‘it’ is not being used as a noun-subject, and hence carries no tacit existential commitments. (As we can see by noting that “It is raining” is not semantically equivalent to “There is a thing which is raining”.) Adding it to our list, we have:

4. Ontological Nihilism. There are no objects. Ontologically honest sentences are those, like feature-placing sentences, which are completely ontologically innocent. Our example becomes: “It is pebbling there.”

Arguments against

The primary argument against nihilism appeals to the indispensability of (ontologically guilty) subject-predicate grammar. We are, of course, far more at home with descriptions like “there is a pebble” than “it is pebbling there”. The question is then whether this familiarity is simply a matter of historical contingency, or if subject-predicate grammar provides an intrinsically good way of representing reality which we have, as it were, discovered in the course of our linguistic development. If it is the latter then the case can be made that this utility is best explained if reality really is divided into objects prior to its perception, i.e. if object realism of some sort is true.

HC criticise this argument by challenging the indispensability of subject-predicate grammar directly. They do so by attempting to provide an account of how a descriptively rich language can be constructed based on ontologically innocent sentences. The first step, as we’ve seen, is to replace noun-subjects with feature-placing verbs: “there is a pebble” becomes “it is pebbling there”. Next, predicates of nouns are replaced with adverbs: “there is a white pebble” becomes “it is pebbling whitely there”, then adverb-dropping inferences are dealt with, then counting terms, tense terms, generalising quantifiers, etc. Finally, charges that these constructions smuggle objects back into the picture are met. I will not go into details here — let us just grant that there may be some hope for this approach. For the remainder of this post I will assume that it succeeds.

If the indispensability argument can be undercut in this fashion, then pending further objections to nihilism the playing field is levelled. What’s more, nihilism’s greater parsimony seems to sway things in its favour. Still, a point won on appeal to Occam’s Razor is unlikely to provoke revolution. What ontological nihilism needs is some substantial arguments in its favour.

Arguments for

HC offer two positive arguments, both of which aim to show that nihilism can help to dissolve certain metaphysical problems which are (they allege) in grave need of dissolution. The first concerns the persistence of identity over time, the second concerns composite objects.

In the case of identity we are asked to consider a watch that has been taken to bits then reassembled with a faulty part replaced. Some feel it is the same watch as before, others feel it is a new one. Who is correct? HC note that the temptation is to view this as a pseudo-problem, and to convey the dispute as a semantic one. But it’s hard to see how this can be the case given object-realism. If we are serious about objects, then the problem remains stubbornly metaphysical: is it the same watch or not? It seems as if the realist’s realism leaves them unable to diagnose their own suspicions. For the nihilist the problem just doesn’t come up. There are no objects, only verb-like happenings of the mysterious ‘it’ (“it was watching then and now it’s watching likewise”), so there is no problem of object identity.

There is a sense in which this would get us out of the fix, but I do not think it is a particularly compelling strategy. We can always dissolve a problem by denying the existence of its subject matter, but this will rarely do much to alleviate the tensions that gave rise to it in the first place. It is inconsequential whether it’s the same watch or not, so the example illustrates their point well. But consider a situation in which moral outcomes hang on identity criteria. If we were to develop teleporters in the far-flung future, whether or not the person who emerges from one end is the same person as whoever entered at the other seems roughly equivalent to asking whether teleporting someone is murder. To take a more sober example, the question of whether we have moral obligations to ourselves (a question with obvious implications for the ethics of suicide) seems to hinge on whether we remain the same person from decade to decade as our cells are gradually replaced.

If the ontological nihilist doesn’t want to embrace moral nihilism — and deny the existence of yet another subject matter — how do they make good on such quandaries? If nihilism is to provide a genuine dissolution to the problem of persistent identity, then it will have to do so in the cases that matter.

The situation with composite objects is mildly better, but ultimately amounts to the same thing. In what sense can we say that a table exists, as opposed to just that its parts exist and are “arranged tablewise”? [7] Again the problem seems false, again the realist is powerless to identify it as such, and again HC dissolve it by nuking the concept of objects altogether. While the link to consequential matters is perhaps less obvious in the case of composite objects than it was in the case of persistent identity, HC’s treatment of it is uncompelling for much the same reasons.

The general worry is that the nihilist is going to lose the ability to articulate ontological contrast of any sort. This leaves them with a problem when it comes to making sense of disputes which are couched in terms of ontological commitments, but whose content is not wholly ontological (moral disputes in particular). There are no doubt ontological disputes for which dissolution is the wisest approach (do gaps and crevices exist?), but there are others which are harder to dismiss as contentless (do human rights exist, or are they just useful fictions?). It is the latter that the ontological nihilist will have to account for if their constructive project is to succeed.

Notes:

[1] John O’Leary-Hawthorne and Andrew Cortens – Towards Ontological Nihilism (link doesn’t provide access, I’m afraid)

[2] Here HC quote Michael Dummett.

[3] ‘Mass term’ is a key notion in Quine’s Word and Object. Examples include ‘red’ as in “Add some more red!”, and ‘kirsch’ as in “Good grief Dorothy, how much kirsch did you put in this?”

[4] It’s worth distinguishing token monism — the view that there is only one particular — from type monism — the view that all particulars are of the same metaphysical type. (Examples of type monism include materialism and idealism.)

[5] ‘Ontological innocence’ is a phrase which, once again, comes from Quine, and is appropriated here by HC.

[6] Nope, not Quine this time — ‘feature-placing’ is a term coined, according to HC, by PF Strawson.

[7] HC’s discussion of composite objects pays heavy lip-service to Peter van Inwagen’s Material Beings.

The knowledge argument and phenomenal concepts (or, Mary’s Room redecorated)

A month or so ago I wrote a post [1] about the knowledge argument against physicalism in which I offered my take on a reply known in the literature as the ‘ability hypothesis’. Roughly put, the ability hypothesis says that while the physically omniscient Mary does have a novel experience when she leaves her monochrome room and sees a rose, her acquisition of this new knowledge-of-what-red-is-like is not an acquisition of a new true belief but of a new ability — the ability to imagine or recall the experience of red.

The analogy to consider is with something like juggling: I could know all the physical facts about juggling and still not know how to juggle, for the simple reason that I spend all my time reading books about the neural peculiarities of legendary jugglers when I could be practicing my 4-ball shower. Naturally, this isn’t a problem for physicalism. Thus if the knowledge argument rests on an equivocation between knowing in the sense of knowing how and in the sense of knowing that — as the ability hypothesis says it does — then it doesn’t stick as an objection to physicalism.

I made a real howler in that piece. I cited David Papineau as having fleshed out a “particularly thorough version” of the ability hypothesis in his paper Phenomenal and Perceptual Concepts, when he does nothing of the sort. The ability hypothesis is in fact usually associated with David Lewis [2], while Papineau’s own reply to the knowledge argument lies several turns in the road beyond it (making it a counter-objection to an objection to a reply to an argument against physicalism — but you all still love the philosophy of mind, right?) In the name of karmic adjustment I shall now try to give Papineau’s argument the airing I denied it last time.

In order to respond to the ability hypothesis one would have to show that with her novel experience Mary acquires not just some new know-how, but also some new ‘know-that’. Consider the following modification of the knowledge argument: rather than being let out of her room to behold a rose, or turning on the colour television, Mary remains in her monochrome environment but while she sleeps some shady character slips a piece of red card into the room via a hatch in the wall. When she wakes and sees it, Mary has a novel experience. But unlike when she sees a rose, or the Australian flag on her television, she does not know which colour concept to associate with it. She knows all about ‘red’ of course — which chemical substances are red, which region of the spectrum it occupies, etc. — but since she has never seen it before, and since a piece of card could be any colour, she doesn’t know that what she’s experiencing is red, as opposed to, say, blue.

Intuition suggests that despite this Mary can still form a concept which references the experience — let us call it Q — with which she can have thoughts such as “that piece of card causes Q”, “I had Q earlier”, etc. If she then discovers or is told that the piece of card is red, it seems that she’s suddenly furnished with all sorts of know-that which she didn’t have (and, more importantly, couldn’t have had) previously. For example, she knows that light in the red region of the colour spectrum elicits Q when it hits her retina. Since ex hypothesi she already possessed all the physical knowledge it cannot have been exhaustive, and the knowledge argument holds tight.

Concepts like Q are called phenomenal concepts. When we start taking phenomenal concepts seriously, the ability hypothesis begins to look thin. The ability to recall an experience is precisely what enables the consistency of reference required for a concept to hook onto it [4], thereby opening the possibility of new know-that. There does not seem to be anything analogous going on in the case of juggling.

This is the juncture at which Papineau’s point becomes relevant, because it inverts the above considerations and uses the notion of a phenomenal concept to flesh out a physicalist response to the knowledge argument. It goes like this. If physicalism is true then the experience referenced by Q simply is some physical or physically instantiated property — call it P. So when Mary learns that light in the red region of the colour spectrum elicits Q when it hits her retina, this is the same as learning that light in the red region elicits P when it hits her retina. But she already knew this, because P is a physical property. So it seems that on a physicalist view Mary has not in fact added anything to her description of the world. What she has done is discovered that P and Q refer to the same thing. Papineau illustrates the point like so [5]:

Suppose a researcher into educational history knows of all the 117 children in Bristol Primary School in 1910—including Archie Leach.  Then she learns, on reading Movie Magazine, that Cary Grant was also at the school in 1910.  In a sense, she has learned something new.  But this doesn’t mean that there was an extra child in the school, in addition to the 117 she already knew about.  In truth, Cary Grant is one and the same person as Archie Leach.  Her new knowledge is only new at the level of concepts.  At the level of reference there is nothing new.  The objective fact which validates her new knowledge that Cary Grant was at that school is no different from the objective fact that validated her old knowledge that Archie Leach was at the school.  (Moreover, if she comes to learn that Cary Grant = Archie Leach, the fact which makes this identity true is similarly none other than the fact she always knew, that Archie Leach = Archie Leach.)

This seems to me to be both nifty and simple. If we accept phenomenal concepts, Mary’s knowledge is not about the world, but about the mappings between two conceptions of the world she’s formed from ‘different angles’. If on the other hand we don’t accept phenomenal concepts, then the ability hypothesis still stands.

Notes:

[1] Mary’s room and the Myth of the Given

[2] David Lewis – What Experience Teaches

[3] Phenomenal concepts are not uncontentious — they seem, for example, to be exactly the sort of thing rendered impossible by Wittgenstein’s private language argument. But hey, there’s hardly any consensus on how successful that is as an argument, so maybe the intuitive plausibility of phenomenal concepts is points against it. (See Papineau’s paper below for more on this.)

[4] David Papineau – Phenomenal Concepts and the Private Language Argument

On the indispensability of mathematical objects

The Quine-Putnam (QP) indispensability argument for the existence of mathematical entities cropped up in a discussion I was having recently about the difference between materialism and metaphysical naturalism. I’d suggested that the existence of things like numbers marks a potential fault line between the two: while materialism explicitly precludes the existence of abstract objects it may be a path open to the metaphysical naturalist to argue that abstract objects are indeed real and natural, though immaterial, and the indispensability argument might well provide the motivation for taking it.

The SEP puts the QP argument like so [2]:

1. We ought to have ontological commitment to all and only the entities that are indispensable to our best scientific theories.

2. Mathematical entities are indispensable to our best scientific theories.

3. Hence, we ought to have ontological commitment to mathematical entities.

The force of this argument derives from the fact that we tend to believe in the existence of things like quarks and black holes (SEP’s examples) for similar reasons. Our best models of particle physics and cosmology make heavy reference to quarks and black holes, respectively, and on these grounds we tend to believe that quarks and black holes are real things, despite the fact that they can’t be observed directly.

So how about mathematics? Mathematical theories which refer to things like numbers, vectors, conic sections, sets, and so on, are vital to the formulation of our best scientific theories — this is not under dispute. By the same token shouldn’t we believe in these entities too?

While I think there’s something in this, it seems to me that the first premise of the QP argument as quoted is unjustifiably strong. Much hinges on what exactly is meant by an entity being indispensable, but if it just means that they feature in our best scientific models, where one aspect of ‘best’ is adherence to the principle of parsimony (i.e. they’ve been maximally shaved of extraneous entities by Occam’s Razor), then it seems quite possible that entities could satisfy this without existing. In fact the history of science offers some potential candidates, for example phlogiston.

It’s debatable whether phlogiston theory really was the best theory of combustion around at its time, but the point here is not to argue that particular instances of entities referenced by theories are or were both indispensable and non-existent, just that such instances are perfectly possible. Minimally, then, we should weaken 1 to “we ought to consider the indispensability of entities to our best scientific theories as good evidence for their existence” [2]. Indispensability should not secure our ontological commitment, though it should push us in that direction. The conclusion of the revised argument is then (the weaker one) that we have a good prima facie case for the existence of mathematical entities.

But while this may push the burden of proof onto those who deny the existence of mathematical entities, we have not yet considered their arguments to see if this burden can be met. Indeed this is where the contrast between quarks and numbers lies — a contrast which is masked when the argument is presented as it originally was. For unlike quarks and black holes, which despite their peculiarities are physically and temporally located things with causal powers, giving them an ontological status similar to other physical entities, mathematical entities (if they exist) are abstract — immaterial and causally inert.

With this comes metaphysical and conceptual objections. One is that if mathematical objects exist and are causally inert, it seems impossible that we could know about them. Since it appears that we do know a thing or two about them, those who profess their existence are left in a tight spot [3]. There are other objections, but here I just wish to note that these are objections which do not apply in the case of quarks and black holes, and all of which may motivate us to revise the initial generous assessment of the likelihood of the existence of mathematical entities downward, perhaps catastrophically.

Some might argue that philosophical criticisms of the above sort should not be given any weight, and that science should be the only source of our ontological commitments. Quine’s naturalised epistemology certainly seems to say something along those lines, but since Quine also proclaimed that ‘philosophy is continuous with science’ it’s doubtful whether he would have recognised the contrast required for the wholesale rejection of ‘purely’ philosophical objections. What’s more, there are other situations in which philosophical reasons do seem to decide issues of ontology in fairly non-controversial ways. For example, interaction problems are one of the main reasons substance dualism is not widely adhered to. Even when not considered deciding they are at least considered relevant.

In summary, the QP argument (or some variation of it) does offer a similar case for the existence of mathematical entities to that given by common reason for the existence of unobservables like quarks and black holes. But this case alone is not decisive. What is decisive in the case of quarks and blacks holes is the further absence of philosophical problems caused by them, while the presence of such problems in the case of mathematical objects gives us reason for pause, perhaps reason enough to reject their existence altogether.

Notes:

[1] Stanford Encyclopaedia of Philosophy – Indispensability Arguments in the Philosophy of Mathematics

[2] I’ve dropped the ‘and only’ clause here because we’re interested in indispensability only insofar as it provides justification for the existence of certain entities, not against. (Note that the ‘and only’ clause is not required for the original argument to work.)

[3] This is the objection to mathematical Platonism from epistemological access, usually attributed to Paul Benacerraf. See the relevant section in the SEP’s entry on Platonism in the Philosophy of Mathematics.

Expressivism and the Frege-Geach problem

Being somewhat green when it comes to this topic I could be way off the mark here, but a thought occurred to me about the Frege-Geach objection to expressivism which I felt worth writing a brief post about (if only so that someone better-versed might appear ghostlike from the intertubes and tell me to shut up).

Expressivism is a meta-ethical position which claims that moral judgements express attitudes rather than facts. On this view, when I say that murder is wrong I am not describing a moral aspect of reality (in some way that could be true or false), but expressing an evaluative attitude towards murder (perhaps an emotive attitude akin to disapproval, or an aesthetic attitude akin to distaste). Since it holds that moral judgements do not have truth-values, expressivism is a kind of moral anti-realism.

The problem described by Geach (drawing heavily on Frege) [1] appeals to the difficulty of accounting for certain moral judgements in expressivist terms. Moral judgements involving conditionals cause particular concern, for example:

If tormenting the cat is wrong, then getting your little brother to torment the cat is wrong.

Call this proposition FG. Propositions like this are the bread-and-butter of moral reasoning, and any theory of moral language which doesn’t account for them has surely failed. The expressivist struggles with FG because assenting to it does not require assenting to the first clause (or the second, for that matter). That is, no evaluative attitude need be expressed toward cat torment (or the instigation of cat torment) in order to assent to FG as a whole. This leaves them with the question of what evaluative attitude is being expressed by FG – a question which seems difficult to answer.

As far as I can tell, most if not all discussions of the Frege-Geach problem take it for granted that FG is a moral judgement, presumably because of the appearance in it of the moral predicate ‘is wrong’. But isn’t it possible that despite appearances it isn’t a moral judgement at all?

The way FG is phrased is slightly ambiguous. One could argue that in order to assent to it we would also be assenting to the judgement “getting your little brother to do something wrong is wrong” (I can’t think of any moral theory which would disagree with that, but it does seem non-trivial). So we might glibly contend that there are evaluative attitudes expressed by FG. But then this could easily be countered by drawing out the hidden judgement and folding it back in explicitly. Doing so (for the sake of removing ambiguity at the cost of some clarity) we get FG’:

If tormenting the cat is wrong and getting your little brother to do something wrong is wrong, then it is wrong to get your little brother to torment the cat.

The thing about FG’ is that no matter what your moral views are you’re going to agree with it. Of course, this is exactly what Geach was picking up on. But if there is no moral content to it (i.e. there are no moral disagreements to be had about it) then how can it be a moral judgment, whether it contains moral predicates or not? And if it is not a moral judgement, why should expressivists be expected to provide the evaluative attitude it expresses?

A critic might be concerned that the problem is being dodged here. Even if judgements like FG’ are not strictly speaking moral judgements, the expressivist will still have to account for their role in moral reasoning to be a successful theory of moral language. This is true, but the point is just that if FG’ and its ilk are not moral judgements, then the expressivist may account for them in ways other than by explaining the evaluative attitude they express. Doesn’t this leave them with options?

Notes:

[1] Peter Geach – Ascriptivism (1960).

Lessons from Feyerabend

A standard account of the Copernican revolution might run like this: as the body of celestial observations grew (aided by the invention of the telescope in the early 17th century) anomalies were recorded which revealed inadequacies in the Ptolemaic system, motivating astronomers to consider rival theories. Among these was the heliocentric system of Copernicus, which ultimately turned out to be the best (it explained the data in the simplest, most elegant way, adopted the least auxiliary hypotheses, etc.). Despite the efforts of the Church it gradually came to be accepted. According to this story observation and theory are independent of each other — observation is the neutral party which can topple old theories and adjudicate between new ones.

This presumed independence of theory and observation is one of the main targets in Paul Feyerabend’s Against Method [1]. Feyerabend illustrates his point by considering the tower argument against the motion of the earth, which was for a time one of the main sources of resistance to the heliocentric model of the solar system. He quotes Galileo [2]:

[H]eavy bodies. . . falling down from on high, go by a straight and vertical line to the surface of the earth. This is considered an irrefutable argument for the earth being motionless. For, if it made a diurnal rotation, a tower from whose top a rock was let fall, being carried by the whirling of the earth, would travel many hundreds of yards to the east in the time the rock would consume in its fall, and the rock ought to strike the earth that distance away from the base of the tower.

Needless to say, this argument eventually lost its purchase. Nowadays we understand that the observation that rocks dropped from towers fall perpendicular to the earth is not evidence that the earth is stationary. The force of the reasoning derived from something we have lost [3]:

[Galileo] tells us that the everyday thinking of the time assumes the ‘operative’ character of all motion, or, to use well-known philosophical terms, it assumes a naive realism with respect to motion: except for occasional and unavoidable illusions, apparent motion is identical with real (absolute) motion.

Without absolute motion the tower argument makes little sense, and a crucial component of the Copernican revolution was the adoption by astronomers of Galilean relativity. As the standard story tells it this shift in attitudes toward motion occurred at the level of theory, while the observations — e.g. that falling rocks fall straight to earth — remained stable. Feyerabend disputes this. Naive realism about motion cannot be considered part of the old Ptolemaic theory; one reason is that it is not required, but the more important is that it couldn’t even be stated until Galileo had provided an alternative. Naive realism about motion was not part of the explicit theory, it was an implicit natural interpretation of the observations [4].

Natural interpretations, unlike explicit hypotheses or postulates, cannot be said to be distinct from the data. On the contrary: they permeate the observation language itself. To put it another way, they are part of what the terms used to state and record observations mean to the people using them. According to Feyerabend, it is not correct to say that in championing Copernicus Galileo was offering a new theory that did a better job of accounting for the data. Quite the opposite: it was inconsistent with the data as then understood. As he writes [5]:

[A] theory may clash with the evidence not because it is not correct, but because the evidence is contaminated.

Galileo did not just argue for a new theory, he provided a new observation language. From this Feyerabend concludes that scientists must from time to time proceed by what he calls ‘counterinduction’, that is by actively developing theories which contradict the current body of observation. By doing so they can unearth (and perhaps then discard) the natural interpretations it contains.

While others have made similar points about holism and the theory-ladenness of observation, Feyerabend is more suspicious than most of overly prescriptive views of scientific method. Taking all this a step further, we can note that his idea that new theories provide not just new explanations of the data, but in effect change the very data to be explained is one that applies to rational discourse in general.

Take materialism and idealism. Idealists have on occasion been accused of denying the existence of their own hands [6]. This criticism likely rings hollow for them — while the idealist does deny that the hand exists in any mind-independent sense, they need not deny the existence of the hand as such, because on their view the things they refer to as their hands were just perceptual entities all along. In virtue of their idealism they mean something different by ‘my hand’.

Similarly, materialism often gets criticised on grounds like “it is the nonsensical idea that mindless physical stuff can crash together to create minds”. The materialist will be unmoved by this, because what is materialism if not the thesis that physical systems can indeed be mindful? As with the above comment on idealism it is at best a cynical rephrasing of the position, and at worst a question-begging criticism of it.

The pattern crops up all over the place, particularly in relation to the big questions of morality, truth, and the existence of God. Beliefs on these topics have deep semantic ripples, and the opportunities for those with differing views to talk past one another are manifold. Feyerabend’s philosophy of science can help us see that this needn’t be because of malice or disinterest on either side, but may simply come down to the fact that charitable interpretation is really hard — it requires more than paying attention to someone’s definitions, it requires stepping into their entire paradigm. The general point can be put somewhat crudely: it is not just the case that the beliefs we hold depend on what we mean by the terms that compose them, but also that what we mean by those terms will depend on (and change with) the beliefs we hold that involve them.

Notes:

1. Paul Feyerabend, Against Method (4th edition). First published in 1975. (This really was a pleasure to read.)

2. Galileo Galilei, Trattato della sfera, quoted AM p50.

3. AM p54.

4. ‘natural interpretation’ is a phrase Feyerabend borrows from Francis Bacon. See AM p55.

5. AM p15.

6. This argument sometimes gets accredited to GE Moore. Presumably this comes from his essay A Defence of Common Sense, though in it he seems to be making a subtler point.

Some thoughts on mathematics

Any successful philosophy of mathematics will have to face up to these two facts:

1. Mathematical truths are, or at least seem to be, necessary truths.
2. Mathematics is immensely useful in physics and other sciences.

At first glance these appear to be in conflict. One way we might try to account for necessity is by holding that mathematical truths are analytic truths, i.e. that their truth is derived from the meaning of the terms that occur in them, and nothing else. The problem here is that if mathematics is analytic, its utility becomes mysterious. Analytic truths are by definition not about the world, they are about the relationships between meanings.

In the opposite direction we face a similar dilemma. Because mathematics is so helpful in our theories of the world, we would expect it to be in some sense about the world. But the world is a shifty place, full of contingent structure and beings that might not have been. If mathematics describes a contingent world, how can its truths be necessary?

Platonists try to resolve this tension by positing that there is some ideal, non-contingent portion of reality which mathematics describes. If true this deals with necessity, and then utility is handled by saying that the physical world is somehow derivative from or dependent on this ideal world. This approach raises many questions, arguably more than it answers. 

But perhaps more tellingly, Platonism offers accounts of things it doesn’t need to. In positing that an ideal portion of the world accessible via mathematics is necessary and that the physical world is dependent on it, Platonism implies that the mathematics which is useful in this world will be useful in all other possible worlds. In other words, it provides an account not just of the necessity and utility of mathematics, but also of the necessity of its utility.

But this necessity needn’t be accounted for, because while a truth being necessary and useful does imply that it couldn’t have been false, it does not imply that it couldn’t have been useless. This shows us that Platonism is stronger than required, suggesting that a weaker theory might be able to do the same explanatory work with less of the metaphysical baggage. Here’s a thought about how this might work.

If we stick to our guns and maintain that mathematical truths are indeed analytic truths (contra Platonism), we’re left with troublesome questions of utility. We might contend that theorems of group theory, say, just follow from the definitions of a group, and are necessarily true in this stipulative, almost trivial sense. But then why is group theory so handy for doing quantum mechanics? Perhaps we can say that it’s useful because some bits of the physical world exhibit group-like structure. Ah ha! says the Platonist, but the physical world is contingent, so it might not have exhibited group-like structure. How can you square that with the necessity of group theoretic theorems?

But at this point we might wonder what exactly the problem is. We have two things: a theory which consists of some definitions plus theorems derived from them – i.e. a system of analytically true propositions – and a mapping between the theory and a portion of the physical world. But there’s no reason to suppose the mapping between theory and world is anything but contingent. In another world group theory might be irrelevant to physics (or to any other descriptive aspect of it). But this doesn’t mean that group theory might have been false, it just means that group theory might not have been useful. Mathematics that is true and useful in one world might be true and useless in a world with different physics. 

What I’m advocating here is something like a mix between mathematical formalism and mathematical nominalism. On their own these theories face objections – nominalism can’t cope with necessity, and formalism seems at a loss when it comes to utility. But once we understand these questions as decoupled (that coupling them was a mistake in the first place) we can see that there’s scope for tethering the good ideas in formalism to the good ideas in nominalism. This hybridisation seems to me to provide a framework in which an account of both the necessity and utility of mathematics can be approached, and without taking on ad hoc metaphysical commitments.

A revision (09/06/14):

I’ve realised that I was being unfair to Platonism when I said that it implies that the utility of mathematics is necessary utility. I didn’t offer any justification for this because it seemed obvious to me, but having received a question about it which I didn’t have an answer for it now seems likely to me that, like many things which seem obvious at first, it is false. I had in mind the notion that were Platonism true, the abstract entities referred to by mathematical terms would be the same in any given possible world, and then that since the physical world is derived from them the physics of each would appeal to the same mathematics. Unfortunately that looks like a non sequitur. I suppose the Platonic world could contain all the abstract objects referred to in (any) mathematical statements, with the utility of each object (or class of objects) being contingent on the physical world. (I’m not sure whether a Platonist would want to argue in that manner, but perhaps it’s an option open to them.)

Hopefully this doesn’t obscure the main point of the post, which is that despite appearances questions about the necessity and utility of mathematics don’t cause problems for each other, and that their decoupling can help us gain a sense of what a successful philosophy of mathematics might look like. I still think that a formalism/nominalism combo can do this better than Platonism, though in light of this revision I can’t claim the reasons I’ve given here are especially good ones. Still, should any errant Platonist wander this way I’d welcome any comments about how the utility of mathematics might be worked out in that sort of framework, and whether what is mathematically useful in describing our world might not have been.

Mary’s Room and the Myth of the Given

Frank Jackson originally put his knowledge argument against physicalism (aka Mary’s Room) like so [1]:

Mary is a brilliant scientist who is, for whatever reason, forced to investigate the world from a black and white room via a black and white television monitor. She specializes in the neurophysiology of vision and acquires, let us suppose, all the physical information there is to obtain about what goes on when we see ripe tomatoes, or the sky, and use terms like ‘red’, ‘blue’, and so on. She discovers, for example, just which wavelength combinations from the sky stimulate the retina, and exactly how this produces via the central nervous system the contraction of the vocal chords and expulsion of air from the lungs that results in the uttering of the sentence ‘The sky is blue’.… What will happen when Mary is released from her black and white room or is given a color television monitor? Will she learn anything or not? It seems just obvious that she will learn something about the world and our visual experience of it. But then it is inescapable that her previous knowledge was incomplete. But she had all the physical information. Ergo there is more to have than that, and Physicalism is false.

The argument runs:

1. Mary has all the physical knowledge of colour vision there is prior to turning on the television.
2. When Mary turns on the television, she gains some new knowledge of colour vision. (She learns what it is like to see red, say.)
3. Hence, the physical knowledge of colour vision does not exhaust all there is to know about colour vision.
4. Physicalism implies that the physical knowledge of colour vision is all there is to know about colour vision.
5. Hence, physicalism is false.

Objection 1: begs the question

One objection given by Dennett [2] is simply that premise 2 begs the question against physicalism. To say that no amount of physical knowledge of sensations (physiology, word-usage, etc.) can amount to knowing what the sensation is like is just to say that physicalism is false, according to Dennett, and so can hardly be used as the basis of an argument against it.

We can see from this that Dennett signs up to certain aspects of the argument. He seems to accept that were Mary to have a novel experience when she turns the television on, this would be a problem for physicalism. On physicalism, then, Mary should somehow be able to ‘synthesise’ the sensation of red by examining, absorbing, and understanding a whole heap of propositions about colour vision – and nothing else. So even if we accept Dennett’s unfavourable analysis of the logic, the price seems to be admitting that physicalism has implications which border on the magical.

Dennett’s response to this would be that the magic has already been injected into argument by way of the massively unrealistic conditions stipulated in premise 1. And so the wheel keeps turning. For now though, let us forget this line of reasoning. Instead I want to turn to another objection to the knowledge argument, one which I have realised recently is much, much better.

Objection 2: appeals to the myth of the given

The new objection contends that even if Mary were to have a novel experience, this does not entail that she acquires any relevant new knowledge. If this is so, the fact of her having a new experience does not conflict with the physicalist premise that her complete knowledge of the physical facts is exhaustive of the facts in general. The argument misfires completely, and its refutation need not sign physicalists up to any queasy voodoo.

There are many variations on this objection (a thorough one is given by Papineau [3]), but here I want to link the general idea with Wilfrid Sellars’ notion of the ‘myth of the given’ [4]. So from now on let’s assume (a) that Mary knows everything there is to physically know about colour vision (in the broadest possible sense) and (b) that she really does have a novel experience when she turns on the television (contra Dennett).

What does (a) mean? Roughly speaking, to ‘have all the physical knowledge’ is to know all the observable facts about brains, light, colour-word usage, etc., plus all the theoretical facts about how these relate to each other. In other words, (a) stipulates that Mary has a certain amount of propositional knowledge – a body of beliefs satisfying whichever requirements you feel are necessary to earn the title ‘knowledge’. According to (b) she has a novel experience, thus learning something new. But how does having a new experience entail learning a new fact? Perhaps it doesn’t.

OK, but surely she gains some sort of new knowledge? Isn’t that how we’d say it: “Mary now knows what red is like”? Indeed we would, but it’s unclear why this should be relevant. If I say “I know what spinach tastes like”, for example, all I am saying is that I have tasted spinach. Maybe we should add an extra condition that I should be able to recall that taste, in order to fully capture the notion of ‘knowledge-of-what-something-is-like’. Even with this caveat, my transition from not knowing what spinach is like to knowing what it is like does not imply that I know any new facts about spinach. All it implies is that something about me has changed. So there may be new facts about me – I probably even know these new facts – but I don’t know anything new about spinach or spinach flavour [5].

If we call knowledge in the sense of knowing what something is like ‘know-like’, and propositional knowledge ‘know-that’, we can say that when Mary turns on the television and sees red, she gains some new know-like but no new know-that. At a push we might say she does gain some new know-that, but only in the sense that she now knows facts like “I have experienced red”. But this isn’t new knowledge about colour or colour vision, this is new knowledge about her own past – of an event that has occurred only since pressing the television’s on button.

So the question is: why should any of this be a problem for physicalism? The implication of physicalism is that the physical facts determine the mental facts. There is no implication that people who know all the physical facts will have encountered certain sensations, or be able to relive those sensations. In order for Mary’s novel experience to pose a problem, it would have to amount to her epistemic acquisition of some relevant new fact (where relevant means ‘about colour or colour vision’). But to assume that it does is just to conflate sensing with knowing – know-like with know-that – and this is the myth of the given.

It seems to me, then, that the knowledge argument falls straight into Sellars’ trap. In order to conflict with physicalism, premise 2 has to be taken as claiming that Mary not only has a novel experience, but gains some relevant new propositional knowledge. But there is no reason to think this. That she does indeed have a novel experience is a claim that tugs hard at our intuitions, but without some separate account of how the gap between know-like and know-that is to be bridged, the argument is left looking thin.

Edit: the citation I give of David Papineau in this post is highly dubious — Papineau does not support the objection I outlined. Please see this post for correction and elaboration.

Notes:

1. Frank Jackson – Epiphenomenal Qualia
2. Daniel Dennett – Consciousness Explained
3. David Papineau – Phenomenal and Perceptual Concepts
4. Sellars develops this idea in Empiricism and the Philosophy of Mind
5. I suppose it might seem that some know-that has slipped in under the wing of ‘recalling the taste’. But I don’t think it has – to be able to recall a taste is just to be able to relive it in some sense, i.e. to summon it back to consciousness (or some derivative version of it).

Richard Rorty on the incorrigibility of the mental

Here’s a question: what sort of problem is the mind-body problem? I think it’s fair to say that it is typically regarded as an ontological problem: first there’s a set of questions about what kinds of things minds and bodies are, then a worry arises about how we can square the answers we’re inclined to give. In its modern guise as the hard problem of consciousness, this worry centres in on the contrast between physical things and so-called ‘phenomenal’ things – items of conscious experience like pains and itches. The difference is easy to spell out, and has intuitive force. Hence physicalists tend to be in the business not of solving the hard problem, but dissolving it – of attempting to show that our intuitions about the ontology of phenomenal items are based on a mistake.

One such effort at dissolution is offered by Richard Rorty in his 1979 book Philosophy and the Mirror of Nature [1]. Rorty thought that the peculiarity we attribute to phenomenal things, while contentful, points not to their ontological character but to their epistemic character. In other words, phenomenal things are distinguished not by what they are, but by how they’re known. In his view this distinction only comes to be regarded as ontologically problematic because of unjustified tendencies among philosophers to think of certain states of entities as if they were entities themselves.

What makes phenomenal things different from physical things is that there is no meaningful distinction between their appearance and their reality. For a pain, to be is just to be felt. On the other hand, a tomato’s reality (or at least some portion of it) is independent of its experience; a hallucinated pain is still a pain, but a hallucinated tomato is not a tomato. Another way of putting it is that phenomenal things have a first-person ontology: they exist only insofar as they’re experienced [2].

So:

1. p is phenomenal iff. p’s appearance exhausts its reality. 

No physical things satisfy this criterion. But pains do, and pains exist. So isn’t that all there is to showing pains are non-physical?  

In a sense, yes, but this doesn’t necessarily mean much. There are many non-physical things around, such as a rock’s solidity or a person’s health, but these don’t cause us concern. The reason these are ‘non-physical things’ is not because they mysteriously lack physicality, but because they are not things at all, in any strict sense of ‘thing’. A fortiori, they are not physical things. They are just states of things. What we would need for there to be any genuine ontological friction here is something which is both a thing in a strict sense of ‘thing’, and non-physical to boot. 

If we fix the word ‘entity’ to mean thing in a strict sense of ‘thing’, and ‘thing’ for anything we refer to with a noun (like states or properties or entities), we can now ask: are pains entities or some other kind of thing? Typically we understand pain as a state of a person or organism. Pain is something you can be in. So if the non-physicality of pains is to be more pressing than the non-physicality of states in general (i.e. if the mind-body problem is to be worthy of any more attention than the health-body problem), it seems that the criterion for being phenomenal (1) should either give us independent reason to take pains not just as states but as fully-fledged entities, or it should imply that some other entity is non-physical (some part of the organism in the pain-state, perhaps).

In other words, if there is really an ontological problem here, it should persist even if we adopt nominalism about states of organisms, by resisting the tendency to talk about them as if they were entities. (For example, if we insist that “René is experiencing a pain” is just a way of expressing the more strictly literal “René is in pain”.) Rorty contends that it does not.

Our criterion (1) is ambiguous on the question of what sort of things p can be. If we write it explicitly in terms of states, we end up with something like this:

2. A state p of S is phenomenal iff. whenever it seems to S that S is in p, S is in p. 

If it seems to René that he is in pain, then René is in pain. Contrast this with physical states: it may seem to me that I am in good health, but I could be wildly wrong about that. Rorty now provides us with a handy term: a proposition is said to be known incorrigibly by a person if they can’t be wrong about its truth or falsity. Using this we can rewrite 2 like so:

3. A state p of S is phenomenal iff. whether S is in p or not is known incorrigibly by S

We can see here how what was supposed to be ontological turns out to be epistemic. If we are nominalists about states, what it means for a state to have a first-person ontology is just that it has a first-person epistemology. If we are not nominalists about states, then why are we more worried about the non-physicality of pains than we are about the non-physicality of health and nations?

This line of thought is not intended to eliminate the mysteries of the mind. After all, why should some states be known incorrigibly by those in them and others not? What it does do, however, is raise doubts about whether an ontological thesis like physicalism is really in deep conflict with our intuitions about consciousness. What’s more, recognising phenomenal states as an epistemic rather than ontological curiosity suggests that a whole different set of philosophical resources can be brought to bear on the hard problem than is usually thought, for example theories of language, justification, and truth.

Notes:

1. Richard Rorty – Philosophy and the Mirror of Nature. See Section 1.

2. ‘First-person ontology’ being John Searle’s phrase. See, e.g. Why I Am Not a Property Dualist