Most of us are familiar with a sceptical thesis attributed to Kant (Reference Kant, Guyer and Wood1781/2007): we lack knowledge of things ‘as they are in themselves’, or Ding an sich. What is perhaps less known is Langton’s (Reference Langton1998) de-idealized interpretation of it, which attributes to Kant a more palatable thesis of ‘epistemic humility’, viz. that we lack knowledge of the intrinsic properties of things. In a paper published posthumously, Lewis (Reference Lewis, Braddon-Mitchell and Nola2009) agrees; not with the Kantian scholarship, but with the humility thesis itself, or at least ‘something very like it’. Moreover, he does so by providing an argument that relies on Ramseyan, as opposed to Kantian, premises. The subject of this paper is Lewis’s argument for the humility thesis, or what he calls ‘Ramseyan humility’.
The thesis that Lewis ends up arguing for, as Langton (Reference Langton2004) explains, is different from the one she attributes to Kant in several key respects. Subsequently, this paper will attempt to tackle Lewis’s humility thesis on its own terms, without getting bogged down in detail over the extent to which it differs from Langton’s Kantian thesis, and the extent to which this thesis, in turn, differs from more orthodox interpretations of Kant. Lewis’s thesis warrants our interest because, as we shall see, not only does Lewis provide an argument for it that is rigorous and plausible, but the thesis itself, though not as challenging as Kantian idealism, also proves to be an obstacle to philosophical and scientific investigations into the fundamental nature of our world.
Here is how. Ramseyan humility, as argued for by Lewis, is the thesis that while we can know which nomological roles are occupied at our world, we can’t know the identities of the fundamental properties that actually occupy them. The argument he gives, roughly, is that since the nomological roles occupied at our world might have been realized by distinct fundamental properties, and in ways that make no observational difference, we can’t know the identities of the fundamental properties that actually occupy these roles. Were this argument successful, even if we forgo Kantian idealism and thereby grant that we have knowledge of the nature of things in the external world, we would still be ignorant, and in a seemingly irremediable way, of the identities of the fundamentalia on which these things depend. Thus, there would be an important sense in which our knowledge of the actual world would still be incomplete.
What to make of all this? Lewis, in the same paper, notes that there is a way around the humility thesis, which involves assuming two theses concerning qualia: panpsychism and the identification thesis. Panpsychism of the relevant form, i.e. a type of panphenomenalism, is the view that (i) all instantiated fundamental properties at our world are qualia – where these include our own qualia.Footnote 1 Moreover, the identification thesis, or revelation, is the view that (ii) by being acquainted with our qualia, we can thereby know their identities.Footnote 2 The conjunction of (i) and (ii) provides us with an exception to the humility rule: if we could know the identities of our qualia, and if it turns out that these qualia are in the set of the fundamental role-occupying phenomenal properties at our world, we could thereby know the identities of some of the fundamental role-occupying properties at our world, namely the ones with which we are acquainted.
It would appear, then, that we have a potential way of refuting the humility thesis. Nonetheless, as it stands, a detailed explanation of exactly how the above conjunction could refute this thesis, as well as a thorough investigation into the plausibility of doing so, remains absent. Lewis’s own explanation is terse, and his assessment of the plausibility even more so: he denies the identification thesis, and ergo this route to refuting the humility thesis, on grounds that it is incompatible with materialism. Moreover, much of the subsequent literature concerning Ramseyan humility has focused on issues orthogonal to those to do with humility vis-à-vis qualia.Footnote 3
This paper aims to make up for this neglect by explaining and evaluating the response to humility that concerns buying into the two aforementioned theses concerning qualia. The structure of the paper is as follows: I will begin by explaining Lewis’s argument for humility – while drawing out the implicit assumptions and secondary arguments that are salient to our project (Section 1). I shall then explain why revelation by itself doesn’t suffice to undermine the argument for humility (Section 2), but that it combined with a type of panphenomenalism does so, and in a way that also provides us with an exception to the humility thesis (Section 3). Finally, I shall assess the plausibility of trying to get rid of our humility in this manner (Section 4). I will argue that the humility thesis remains intact, although for reasons different to those alluded to by Lewis. In brief, I argue the kind of panpsychism needed to defeat Ramseyan humility is both implausible and unmotivated. Moreover, the historical contextualizing of the debate shows that using it to undermine Ramseyan humility is self-defeating, as it involves assuming a form of idealism to undermine a form of epistemic scepticism. The paper ends with a postscript that briefly discusses the prospects of refuting humility by combining revelation with views other than panpsychism (Section 5).
1. Ramseyan humility
Let us begin with some background. Take a postulate of our final theory: T(t 1 … t n, o 1 … o n). The t-terms are the theoretical terms of the theory, and the o-terms are the terms of our ordinary language or O-language. Some of the t-terms will refer to fundamental properties, and when they do, it is assumed that the O-language is capable of saying so. Now replace the t-terms in our theory with variables, and prefix an existential quantifier to our formula. What we now have is a Ramsey sentence: ∃x 1 … ∃x n T(x 1 … x n, o 1 … o n).
There are two reasons why the Ramsey sentence is significant for our purposes. First, it says that theory T has at least one actual realization. This is simply entailed by our definition of the Ramsey sentence. Second, it implicitly defines the t-terms in terms of the variables and the o-terms. Roughly, for any given t-term, even one that refers to a fundamental property, we define ‘t’ as the entity x which occupies such-and-such nomological roles.Footnote 4
We are now ready to spell out the argument for humility, which can be formulated as follows:
-
(1) For the unique actual realization of T, there is a possible and distinct realization of T.
-
(2) If there is a possible and distinct realization of T, no possible observation tells us which realization is actual.
-
(3) Therefore, no possible observation tells us which possible realization of T is actual.
-
(4) If no possible evidence tells us which possible realization of T is actual, we don’t know which fundamental properties obtain at our world.
-
(5) Therefore, we don’t know which fundamental properties obtain at our world.Footnote 5
Premises (1), (2) and (4) are in need of explaining. Let us consider (1) and (4) first, since the premise we can undermine concerning qualia, and so the one we need to look at in detail, is (2).
We get (1) by assuming combinatorialism in conjunction with quidditism. Combinatorialism is a principle of recombination, which takes the laws of nature to be contingent, and assumes that for any given combination of items of the same category, say the monadic property combination –A–B–C–, there are possible re-combinations of them, e.g. –C–A–B–, –C–A–, –A–B–. Quidditism, by contrast, is a thesis concerning property individuation, which states that properties are individuated by something other than the roles they occupy, e.g. by certain quiddities like intrinsic natures or suchnesses. For present purposes, combinatorialism tells us that for the actual realization of T, there are possible realizations of T, i.e. realizations where the nomological roles required for T are realized by different sets of fundamental properties to the actual sets. Moreover, quidditism allows for these possible realizations of T to be distinct realizations, e.g. they will be distinct if the fundamental properties that possibly occupy the relevant nomological roles are distinct from their actual occupants vis-à-vis their quiddities. The conjunction of combinatorialism and quidditism, then, gives us (1).Footnote 6
As for (4), Lewis (Reference Lewis1979) regards all knowledge as a matter of self-location in modal and spatiotemporal space. So to say that we know which realization of T is actual, is to be able to locate ourselves in the set of possible worlds, and at spatiotemporal regions within these worlds, where a specific set of fundamental properties, which happens to be the set that actually satisfies T, satisfies T. Conversely, to say that we don’t know which realization of T is actual to say that we lack the means to locate ourselves in the set of possible worlds (let alone in the relevant regions within these worlds) where the specific set of fundamental properties, which happens to be the set that actually satisfies T, satisfies T. Therefore, if we lack this knowledge, we lack the means to identify which set of fundamental properties obtain at our world. In other words, conditional (4) holds true.
No one disputes premise (4), whereas many would find (1) objectionable. In particular, property structuralists who take properties to be individuated by their nomic roles will deny (1) on grounds of quidditism being false.Footnote 7 For our purposes, however, we can remain neutral on (1) and (4), as what concerns us is (2). There are two assumptions that motivate (2), which are as follows:
I assume that no fundamental properties are named in O-language, except as occupants of roles; in which case T will name them over again, and will say that the property named by so-and-so T-term is the occupant of such-and-such role.
I’ll assume, however, that O-language does suffice to express all possible observations. (Lewis Reference Lewis, Braddon-Mitchell and Nola2009, 206)
Let us dub the first assumption the naming principle: the principle that no fundamental properties are named in the O-language except as occupants of roles. And let us dub the second the sufficiency principle: the principle that O-language suffices to express all possible observations.
The naming principle is significant because if it were true, then the Ramsey sentence would be true whether the realization of the roles is the actual realization or a possible realization. That is, if the roles are occupied, and the o-terms only name fundamental properties as occupants, then the Ramsey sentence (which defines the T-terms using O-language) would be true whether or not the actual occupants or some distinct possible occupants realize the roles.
The sufficiency principle is significant because if it and the naming principle were true, then it would turn out that the Ramsey sentence would be true whether the realization of the roles is the actual realization or a possible realization, and there would be no further observation to tell these possibilities apart, which is to say that if the two principles were true together, there would be no possible observation that would give us evidence to go beyond the Ramsey sentence. We could observe that the nomological roles are occupied, and consequently we would observe that T is satisfied, but that is all we would observe.
Once we note these implications, we see that (2) is supported by an argument of the following form:
-
(i) If the naming principle and the sufficiency principle are true, then no possible observation gives us evidence that goes beyond the Ramsey sentence.
-
(ii) The naming principle and the sufficiency principle are true.
-
(iii) Therefore, no possible observation gives us evidence that goes beyond the Ramsey sentence.
-
(iv) If there is no possible observation that goes beyond the Ramsey sentence, T will be confirmed whether the realization is a possible realization or the actual realization.
-
(v) Therefore, T will be confirmed whether the realization is a possible realization or the actual realization.
-
(vi) If T will be confirmed whether the realization is a possible realization or the actual realization, if there is a possible and distinct realization of T, no possible observation tells us which realization is actual.
-
(vii) Therefore, if there is a possible and distinct realization of T, no possible observation tells us which realization is actual.
We have just seen an explanation for (i). Lewis appears to take (ii) as a given. In support, we can note that the naming principle is true according to a plausible and widely held belief about how the sciences (implicitly) define their theoretical terms. In brief, such theoretical terms, including the ones that pick out fundamental properties, are defined causally in terms of the causal-cum-nomological roles the referents of these terms play, in contrast to anything to do with the intrinsic natures of these referents.Footnote 8
The belief behind the sufficiency principle is likewise commonly held and plausible; at least prima facie. The idea, roughly, is that we can only observe entities via the effects they have on us (and our measuring devices), as opposed to via their intrinsic natures.Footnote 9 (Which would explain why the sciences define their theoretical terms causally). Accordingly, insofar as the O-terms refer to these entities in terms of their causal roles, O-language then suffices to express all possible observations.
Premise (iii) follows straightforwardly from (i) and (ii). As for (iv), this has to do with the nature of the relationship between the postulate T and its Ramsey sentence. The Ramsey sentence of T logically implies all the O-language sentences, which are theorems of T.Footnote 10 Consequently, if the Ramsey sentence is satisfied, then so is T. Similarly, if there is no observation that goes beyond the Ramsey sentence, there is no observation that tells us anything beyond that T is realized either. So (iv) would appear to hold true.
Premise (v) follows straightforwardly from (iii) and (iv). By contrast, how we get premise (vi) isn’t obvious. It follows from premises (iii) and (v). Since no possible observation gives us evidence that goes beyond the Ramsey sentence, and T will be confirmed whether the realization is a possible realization or the actual realization, we get the following: if T will be confirmed whether the realization is a possible realization or the actual realization, if there is a possible and distinct realization of T, then no possible observation tells us which realization is actual.
The conclusion (vii) follows from (v) and (vi). Since T will be confirmed whether the realization is the actual realization or a possible realization, and no possible observation tells us which realization of T is actual, it turns out that if there is a possible and distinct realization of T, no possible observation tells us which realization is actual. Hence, we get (vii), which is identical to premise (2) of the previous argument.
The overall lesson here is that premise (2) of Lewis’s argument for humility appears to be supported by an implicit argument, which in turn rests on an uncontroversial assumption about the relationship between our final theory T and its Ramsey sentence, and on two plausible, but not entirely uncontroversial, principles to do with what is named by our O-language and what we can observe. In the next section, we see how the identification thesis bears on these principles.
2. The identification thesis
The identification thesis is the thesis that ‘anyone acquainted with a quale knows just which property it is’ (Lewis Reference Lewis, Braddon-Mitchell and Nola2009, 217). This notion of knowing our qualia is loaded and in need of unpacking.
We can begin by noting that the identification thesis is a thesis that Lewis (Reference Lewis1995) thinks is built into our folk conception of qualia. Moreover, according to him, this conception not only supposes that we know our qualia, but that we do so ‘in an uncommonly demanding sense of “knowing what”’ (p. 141). Whether our folk conception actually makes these assumptions is a point of controversy, which we can presently leave aside. The point relevant for us is what these assumptions entail.
For Lewis, knowing our qualia in the demanding sense of ‘knowing what’ is to possess ‘identifying knowledge’ of our qualia. He elaborates:
I say that according to the Identification Thesis, the knowledge I gain by having an experience with quale Q enables me to know what Q is – identifies Q – in this sense: any possibility not ruled out by the content of my knowledge is one in which it is Q, and not any other property instead, that is the quale of my experience. Equivalently, when I have an experience with quale Q, the knowledge I thereby gain reveals the essence of Q: a property of Q such that, necessarily, Q has it and nothing else does. (Lewis Reference Lewis1995, 142)
The part of this passage that begins after ‘equivalently’, concerns what Johnston (Reference Johnston1992) calls the doctrine of revelation. We can formulate the doctrine as follows:
[Rev] For any given quale Q, by having an experience with Q, we know or are in a position to know the essence of Q.
The claim being made here is that if we have an experience with a quale, we can know the essence of that quale. For example, if I have a feeling of nausea, I can thereby know the essence of that feeling-cum-quale. Similarly, if I have an experience with a certain pain quale, I can know the essence of that pain quale. According to this view, what I can’t learn experientially but can by other means, say by the sciences, are the accidental features of these qualia – not their essences.Footnote 11
Revelation is an altogether more controversial thesis than simply knowing what our qualia are, as it makes two additional claims. First, that our qualia have essences. Second, that being acquainted with our qualia puts us in a position to know these essences. Yet it is clear from the passage that Lewis equates the identification thesis with revelation.
Once we recognize that identifying our qualia involves knowing their essences, it is tempting to suppose that revelation, by itself, suffices to undermine the argument for humility, as it appears to contradict the sufficiency principle. That is, one might reason as follows: if we know the essences of our qualia, it isn’t the case that our O-language, which defines the entities in our world by their roles, as opposed to their intrinsic essences, suffices to express all possible observations. There will be observations, namely ones we make by being acquainted with our qualia, that aren’t captured by our O-language. Therefore, premise (ii), and thereby (2) of the argument for humility that implicitly assumes (ii), is undermined.
This reasoning is persuasive, but there are two problems with it. The first has to do with the notion of essences at play in the formulation of revelation. Essences are usually regarded as mysterious. They are taken to be certain sui generis natures, such as intrinsic aspects, suchnesses, etc. However, the Lewisian notion of essences at play in the formulation of revelation is rather different. The essence, in general for Lewis, is a property of an object such that it necessarily has it and nothing else does. On this account, the essences could be sui generis natures – if these natures are the only properties had by objects necessarily and uniquely. But they needn’t be. Certain structures, causal properties, physical properties, etc. can also be essences so long as they satisfy the definition. For example, Lewis (Reference Lewis1995, 143, fn 7) notes that the atomic numbers of chemical elements, like 19 for potassium, are clear candidates for being their essences. Similarly, he notes the C-fibre firing might be the essence of our pain qualia.
The point for us, then, is that the essences of our qualia might just be some set of physical properties, which are defined causally by the O-language. For example, they might be certain patterns of neural firing. To clarify, this is to say nothing of whether revelation, interpreted thus, is plausible. The claim, rather, is that simply knowing the essences of our qualia, insofar as we assume the Lewisian notion of essences, doesn’t guarantee that we know any intrinsic natures that aren’t mentioned in the O-language. Therefore, revelation being true, by itself, won’t show that the sufficiency principle is false. Hence, it won’t suffice to undermine (ii).
The second problem with the aforementioned way of trying to undermine the argument for humility is that, even if we could use revelation to show that the sufficiency principle is false, there is another similar principle that can act as its substitute, which makes the argument go through. To elaborate, let us grant that the essences of our qualia we happen to know are certain intrinsic natures. Once we grant this, we thereby grant that the sufficiency principle, which states that O-language suffices to express all possible observations, is false.
Now let us consider another principle, the sufficiency principle*: O-language suffices to express all possible observations about the fundamental properties at our world. This principle is compatible with revelation being true – even if revelation turns out to be true because we know the intrinsic essences of our qualia. That is, even if we know the intrinsic essences of our qualia, insofar as our qualia aren’t fundamental, it will turn out that O-language suffices for expressing all possible observations about the fundamentalia.
Moreover, and rather crucially, this principle can replace the original sufficiency principle to make the argument go through. The role the sufficiency principle played in the argument is that it, in conjunction with the naming principle, showed that not only would the Ramsey sentence be true whether the realization of the roles is the actual realization or a possible realization, but that there would be no further observation to tell these possibilities apart. It so happens that we get the same result if we replace the sufficiency principle with the new principle. If no fundamental properties are named in the O-language except as occupants of roles, and if O-language suffices to express all possible observations about the fundamental properties at our world, then all we could observe is that the roles are occupied; not which fundamental properties actually occupy these roles.
So, as before, there would be no possible observation that would give us evidence to go beyond the Ramsey sentence. And since this is all we need to get the argument for (2) off the ground, the lesson is that revelation, once again, won’t suffice to undermine the argument for humility.
3. Panpsychism
Enter panpsychism. Revelation by itself mightn’t suffice to undermine the argument for humility, but it does when we combine it with a version of panpsychism. In fact, the combination doesn’t just undermine the argument. As we shall see, it provides us with an exception to the humility thesis as well.
To qualify, panpsychism, in general, is the view that every (concrete) thing that exists has mental properties. Not any form of panpsychism, however, can be combined with revelation to give us an exception to humility. According to Lewis, the relevant form is the view that ‘all the actually instantiated fundamental properties are qualia’ (p. 217). So the relevant form is a contemporary version of panpsychism, which typically goes by the name ‘panphenomenalism’ or ‘panexperientialism’.
Some motivation will help clarify this position. Proponents of this view, like the proponents of the humility argument, take it that scientific terms refer to entities (properties, etc.), but only via the causal-cum-nomological roles they occupy. Moreover, while we can observe these entities, we can also only observe them via these roles. That is, our observations, by sense perception or using our scientific tools, only give us access to the causal roles occupied by the entities in our world, as opposed to the intrinsic natures of these entities. Subsequently, it appears that an epistemic humility thesis holds true concerning the limits of the empirical sciences: such sciences can’t tell us about the intrinsic nature of the entities in our world.
Now, a lot of us tend to suppose that we do, however, have access to the intrinsic natures of something, viz. our experiences. We know, via introspection, that these natures are phenomenal. Consequently, it appears that all the knowledge we have of the intrinsic natures of things tells us that such natures are phenomenal. Once we put these two theses together, it makes sense to suppose, or so argue the proponents, that the intrinsic natures of other things might also be phenomenal. In fact, on grounds of parsimony, it makes sense to suppose that the intrinsic natures of all entities, including the fundamentalia, are phenomenal. Thus, we arrive at panphenomenalism: all the instantiated fundamental properties at our world, including the ones that occupy the causal-cum-nomological roles at the fundamental level, are phenomenal.Footnote 12
Here, we would be remiss not to mention a way of arguing for a related view, which is in turn, sometimes, also used to argue for panphenomenalism. The view is Russellian monism: roughly, that matter has intrinsic properties that simultaneously constitute consciousness and occupy the causal-cum-nomological roles described by the sciences. The central motivation for this view is what Alter and Nagasawa (Reference Alter, Nagasawa, Alter and Nagasawa2012) dub ‘the happy-compromise argument’, for it provides a compromise between dualism and physicalism while simultaneously retaining certain merits of both views. In short, like dualism, Russellian monism accommodates the intuition that consciousness isn’t reducible to physical qua scientific entities, as it forgoes such reduction. Moreover, like physicalism, it is able to simultaneously integrate consciousness into the world described by the sciences, as it has phenomenal properties occupying the causal-cum-nomological roles.
Russellian monism doesn’t entail panphenomenalism, e.g. we may take the consciousness constituting relation to be not of one of identity, but one of joint instantiation. This is the view Chalmers calls panprotopsychism: fundamentalia ‘have certain specific properties that are precursors to consciousness and that can collectively constitute consciousness in larger systems’ (Reference Chalmers, Alter and Nagasaki2015, 248). But this constituting relation is, arguably, as or even more mysterious than the relation Russellian monism is supposed to avoid, viz. the one where certain physical properties instantiate the phenomenal ones. If so, we ought to forgo panprotopsychism and once again arrive at panphenomenalism.Footnote 13
It is crucial to note that panphenomenalism itself can also be unpacked in two ways, only one of which can be employed in argument against Ramseyan humility. The difference rests on whether the fundamentalia, whose intrinsic natures are phenomenal, are amongst the phenomenal properties directly available to us via introspection. According to contemporary proponents of panphenomenalism, like Rosenberg (Reference Rosenberg2004) and Strawson (Reference Strawson2006), the phenomenal properties we can introspect aren’t fundamental, but supervene on a set of fundamental properties at a more micro-level. But what’s more, these fundamental properties are still phenomenal in some sense, despite being distinct in nature from our own phenomenology. For example, according to Chalmers, these ‘micro-phenomenal properties’ characterize what-it-is-like to be micro-physical entities such as leptons and quarks, whereas Strawson regards these ‘phenomenal ultimates’ as being intrinsically experiential.
This version of panphenomenalism – call it ‘M-panphenomenal’ as it is sometimes referred to as ‘micropsychism’ – is relevant to Ramseyan humility without threatening the thesis itself. As noted, to claim that the fundamentalia are microphenomenal is still to grant that they are phenomenal. Thus, one may argue that we know the general natures of the fundamentalia, i.e. that they are phenomenal, on grounds that we know the intrinsic natures of our own phenomenal experiences. This, prima facie, appears true. Nevertheless, it won’t really undermine Ramseyan humility for the humility thesis concerns not knowing the specific identities of the fundamental properties that actually occupy the nomological roles.
We can also make this point in another way. The sufficiency principle*, recall, states that O-language suffices to express all possible observations about the fundamental properties at our world. It is this principle which, as we shall see shortly, is undermined by the conjunction of revelation and panphenomenalism of the relevant form. Nonetheless, if the properties that occupy the causal-cum-nomological roles at the fundamental level at our world are micro-phenomenal, then revelation, understood as a thesis about knowing the essences of our own phenomenal properties, won’t show this principle to be false.Footnote 14
There is, however, another version of panphenomenalism that does the trick. Suppose as before that all the instantiated fundamental properties are phenomenal, but with an added proviso: our own qualia, i.e. the phenomenal properties directly available to us via introspection, are amongst the set of role-occupying fundamental-cum-phenomenal properties. To clarify, the view isn’t that all fundamental properties are directly available to us via introspection. For instance, the fundamental properties that occupy the role of being a quark aren’t things that we can directly introspect per se. Rather, the view is that we can introspect some fundamental properties, viz. those that occupy the causal roles realized in our brain (which might include some quark roles). Moreover, these properties are of the same type as, or somehow similar to, those that occupy the causal roles at the fundamental level elsewhere (including the role-occupiers for quarks outside the brain).
This version of panpsychism would probably strike one as being implausible even if one were sympathetic to contemporary forms of panpsychism. It seems to attribute to things like quarks not only conscious experiences, but the kinds of conscious experiences had by us. I will say more about the plausibility of this view in the next section, as it really depends on how we understand the typehood-cum-similarity relation between the relevant properties. For now, let us grant that it is this version of panphenomenalism, call it L-panphenomenalism, which Lewis (rightly) supposes is a threat to the humility thesis.
Let us first turn to see how the conjunction helps undermine the argument for humility. If the set of all actually instantiated fundamental properties are phenomenal, where this set includes our own phenomenal properties, and if we in addition know the essences of our phenomenal properties, we then know the essences of some of the fundamental properties at our world. So, it turns out that the sufficiency principle and the sufficiency principle* are both false. O-language suffices to express neither all possible observations nor all possible observations about the fundamental properties at our world. There will be observations, namely introspective ‘observations’ of our phenomenal properties, which also happen to be fundamental, that O-language simply won’t express. Hence, (ii) and, consequently, (2) are both shown to be false; and with that falls the argument for humility.
Now, just because an argument is knocked down doesn’t entail that its conclusion is false. It just entails that we no longer have good grounds to suppose that the conclusion is true. In the present case, the reasons we have for rejecting the argument for humility, however, also prove to be reasons for believing that its conclusion is false. If L-panphenomenalism and revelation are both true, it no longer remains the case that we are entirely ignorant of the identities of the fundamental properties that occupy the nomological roles at our world. Rather, we do know the identities-cum-essences of some of these fundamental role-occupying properties. Consequently, the conclusion of the argument turns out to be false.
4. Plausibility
4.1. Contra revelation
From what we have seen so far, revelation and a certain type of panphenomenalism are only jointly sufficient to undermine the argument for Ramseyan humility and to provide an exception to the humility rule. Therefore, if either one of these views is false, so falls the case against humility.
Lewis (Reference Lewis, Braddon-Mitchell and Nola2009) himself denies this response to humility on grounds that he denies revelation. Moreover, he (Reference Lewis1995) denies revelation on grounds that it is incompatible with materialism. This is where an interlocutor gets some room to manoeuvre. First, the version of materialism that revelation proves to be incompatible with is what Lewis calls Australian materialism. This is a version of a priori physicalism (or Type-A materialism) that, roughly, states that all phenomenal truths are identical to, or are a priori deducible from, physical truths – where we understand ‘physical truths’ as the kinds of truths that are discoverable empirically by the natural sciences. Lewis (Reference Lewis1995) explains why revelation – with some auxiliary premises – is incompatible with this view. Very roughly, when we introspect, we don’t discover any physical qua scientific truths, e.g. by having a pain quale, I don’t know that my C-fibres are firing. As Lewis says, ‘Making discoveries in neurophysiology is not so easy!’ (Reference Lewis1995, 142). So insofar as we know essential truths about our qualia on the basis of introspection, and we don’t thereby also know any physical qua scientific truths, these truths about qualia can’t essentially be physical-cum-scientific truths. Hence, Australian materialism must be false.Footnote 15
The trouble is the above reasoning doesn’t make revelation incompatible with other versions of physicalism. In fact, there are versions of physicalism that might turn out to be compatible with revelation. For instance, drawing on the work of Damjanovic (Reference Damjanovic2012), there is a way in which revelation might be compatible with a posteriori physicalism (or Type-B materialism): the view that phenomenal truths aren’t deducible a priori from physical truths even though the physical truths necessitate the phenomenal ones.Footnote 16 According to this version of physicalism, our qualia are physical because they supervene on standard physical properties. However, this supervenience relation isn’t one that holds due to a priori necessities such that we can deduce the truths about our qualia from the truths about standard physical qua scientific truths. Rather, the supervenience relation holds due to a posteriori ‘psychophysical’ necessities, which make it possible for us to know all the relevant physical truths and still remain ignorant of the phenomenal ones.
This possibility in turn might allow revelation to be compatible with physicalism. If we could know the relevant physical truths and remain ignorant of the phenomenal ones, perhaps we could, in a similar way, know the phenomenal truths, and remain ignorant of the physical truths that entail them. If this were so, it would leave open the possibility that these physical truths themselves are standard physical truths, as the sciences describe them. On this picture, then, making discoveries in neurophysiology might not be so easy as introspecting. Nonetheless, this isn’t because our qualia are nonphysical, but because qualia supervene on neurophysiological properties due to a posteriori necessities.
I am sceptical as to the prospects of revelation being compatible with any version of physicalism. But the point is that not everyone who is committed to a materialist ontology will deny revelation, at least on Lewisian grounds. To that extent, they won’t find the present justification for humility convincing. The other point to note is this way of retaining humility would be entirely unpersuasive to anyone who doesn’t have such materialist commitments to begin with. Any qualophile who holds that qualia are nonphysical (i.e. that qualia are neither identical to nor supervenient on any set of physical properties) can readily accept revelation. That is, they can accept that by having experiences, we know the essences of our qualia – where these essences turn out to be nonphysical. Consequently, they needn’t deny the response to humility either.
That said, there might be independent grounds to reject revelation. Revelation makes a very bold claim about what we can know on the basis of introspection alone, and so there is reason to view it with caution. However, as far as I can tell, there aren’t any knockdown arguments against it. The only argument to the conclusion that revelation is obviously false rests on the assumption that when we introspect the nature of our phenomenal experiences, it is neither obvious, nor does it seem likely, that we know the essences of our qualia.Footnote 17 Now, this may well be the case, but it has no direct bearing on revelation, as revelation is a claim about knowing the essences of our qualia – not a claim about knowing of certain properties that they are the essences of our qualia. The two claims, as Stoljar (Reference Stoljar, Braddon-Mitchell and Nola2009) notes, are distinct, as the latter claim requires that we possess the concept of essences, whereas the former doesn’t.
The upshot, then, is that revelation, despite being a controversial doctrine, could very well be true. Consequently, for all that Lewis claims, the present response to humility remains viable. But if Lewis wanted to uphold humility, he could have also taken another route: he could have rejected panpsychism.
4.2. Contra panpsychism
I suspect for many, it is accepting panpsychism, not revelation, which is the main bar to adopting the present response to humility. For some, panpsychism isn’t just counterintuitive, it is significantly so. For others, it is nothing short of sheer madness to suppose that all matter, including socks, toothbrushes, quarks, leptons, etc., possess elements of phenomenology.
If panpsychism proved to be so obviously problematic, we needn’t have seriously considered the present case against humility. A quick dismissal of panpsychism on such grounds, however, would be grossly unfair for there are a plethora of distinct kinds of panpsychism, which differ to the degree to which they are counterintuitive. Moreover, contemporary advocates of these views have not only provided thorough defences of their positions, but new, sober and extremely detailed motivations for them as well (Some of these positions, including the motivation for them, were briefly mentioned in Section 3). Accordingly, even if not all of us are convinced, the recent literature on panpsychism should remove some of our incredulity.
Still, I think we should retain the humility thesis, and on grounds to do with panpsychism. This again has to do with the fact that not all versions of panpsychism are created equal. A closer examination of the case against humility reveals that the required form of panpsychism isn’t just the popular, and well defended, version of it that goes by the name panphenomenalism. What we require, rather, is a particular, and it should be said neglected, sub-branch of it, which I dubbed L-panphenomenalism. Elsewhere, I provide an argument against panphenomenalism more generally, which should apply to L-panphenomenalism as well.Footnote 18 I shan’t repeat that argument here, as one needn’t be convinced by it to reject the present proposal against humility. Undermining L-panphenomenalism suffices.
The trouble with L-panphenomenalism concerns how we are supposed to understand the claim that the fundamental phenomenal properties we have access to via introspection, i.e. the ones that occupy the causal roles in our brain, are ‘of the same type’, or ‘similar’ to, fundamental properties that occupy causal roles elsewhere in our world. There appears to be two basic ways of understanding this claim. The first is to suppose that both sets of fundamentalia are of the same type to the extent that they are both phenomenal. This, however, simply won’t do as it threatens to collapse our view back into M-panphenomenalism. That is, even the micro-phenomenal properties, on grounds of being phenomenal, need to be given the same typehood as the role-occupying properties inside our brain.
The second is to confer the same typehood on grounds that both are phenomenal in a qualified sense, viz. they both fall under what Goff (Reference Goff2009, 290) calls ‘o-experience’: ‘the conscious experience that corresponds to organisms, the kind of conscious experience that in our own case we are immediately acquainted with, that we want a theory of consciousness to explain’. This marks a clear distinction between M-panphenomenalism and L-panphenomenalism, for the latter is wedded to all fundamentalia not just being micro-phenomenal, but phenomenal in a fully fledged way that ‘corresponds’ to phenomenally consciousness experiences had by living creatures, including members of our own species.
If you find panpsychism counterintuitive, you should find this particular rendering of it doubly so. While we may grant the plausibility of the fundamental properties being protophenomenal, or even micro-phenomenal, it is much harder to give any real credence to them exhibiting phenomenally conscious experiences in a way that corresponds to our own experiences. It is, for instance, hard to grant that, or even fathom how, the fundamental properties, which occupy the causal-cum-nomological roles of say being a quark, can have the kinds of phenomenal experiences had by you or I. That is, what-it-is-like to taste the tannins in fine wines, hear Miles’s Davis’s trumpet, or feel the intense pain of a dentist’s drill, strikes one as being too phenomenally rich to be felt by a quark.
L-panphenomenalism, thus interpreted, then, makes an extraordinary, counterintuitive claim: that all fundamentalia are akin to our own rich and varied phenomenally conscious experiences. Now, there is nothing wrong with making an extraordinary claim per se. But, extraordinary claims, as the saying goes, require extraordinary evidence. It is precisely such evidence, however, which is lacking in this particular instance. So far as I can tell, no contemporary panpsychist defends, let alone holds, the view that fundamental properties, which occupy the causal roles associated with micro-level properties, like quarks, count as o-experiences.Footnote 19 What’s more, the typical reasons cited for accepting panphenomenalism can be accommodated by the relatively less incredible M-panphenomenalism. The real trouble with this reading of L-panphenomenalism, subsequently, isn’t that it makes an extraordinary, counter-intuitive claim, but that it makes one that is unmotivated to boot.
Like the case against revelation, the case against L-panphenomenalism isn’t decisive. Nevertheless, unlike revelation, this particular version of panpsychism is found wanting on several accounts, one of which is a lack of motivation. This is damaging to the view, but crucially, this also, arguably, provides sufficient grounds to problematize its use in the argument against Ramseyan humility. For one may argue as follows: given the option between two controversial though tenable positions, one motivated, and the other not, viz. humility vs. our conjunction, more credence ought to be given to the motivated one. What prevails, then, is humility.
4.3. Contra panpsychist idealism
The present endorsement of Ramseyan humility rests not directly on the problems that plague L-panphenomenalism, but rather on the unjustifiability of employing such a doctrine in a strategy against humility given its aforementioned problems. Still, if the final nail on the coffin is to result from questioning the role L-panphenomenalism is supposed to play in such a strategy, we can do better. There is a sense in which employing panphenomenalism, of either variety, is self-defeating once we take the broader picture of Ramseyan humility, and its interrelated sceptical theses, into account.
The basic idea is this. We began this paper by noting that Langton’s interpretation of Kant moves him away from the more ominous, idealist reading of scepticism towards a more palatable epistemic humility thesis, one that in Lewis’s hands takes the shape of Ramseyan humility. It turns out, however, that the way to remedy our ignorance involves buying into a form of panpsychism, which is symptomatic of precisely the kind of idealism that Langton’s Kant was trying to avoid in the first place.
Let me elaborate. Kantian scepticism, as traditionally construed, is still an epistemic thesis: it concerns our ignorance of things-as-they-are-in-themselves. While this is still worrying, it is a milder form of scepticism than that which he explicitly tries to avoid, i.e. Cartesian scepticism, which doubts the existence of objects outside the mind.Footnote 20 (In other words, Kant was worried about not knowing the natures of mind-independent entities, whereas Descartes (Reference Descartes and Moriarty1641/2008) was worried about not knowing whether they even exist in the first place). Descartes’s scepticism is still epistemic, and, subsequently, milder than the metaphysical thesis that denies the very existence of mind-independent matter. This is full-blown idealism, as espoused by those such as Berkeley and Leibniz, which contends that all that really exists are minds and their ideas.
Three clarifications are in order. First, what are ideas in the relevant sense? Though early modern philosophers didn’t agree on a definition, ‘ideas’ was employed as a technical term, which was, roughly, supposed to be inclusive of all the kinds of mental entities accessible to the mind. If this description proves too opaque, we can think of them as the mental contents of our conscious experiences. These may include, though aren’t exhausted by, thoughts, mental representations, sensations and sense-data. Though the early moderns didn’t explicitly talk about qualia, contemporary conceptions of them, as features of experiences available to us by introspection, would have them classified as ideas as well.
Second, was Kant an idealist in the commonly used metaphysical sense of the term? This is a point of contention amongst Kant scholars. Although he took great pains to distinguish his ‘transcendental idealism’ from that of Berkeley’s (1948–1957, Section 1) in the second edition of Critique of Pure Reason,Footnote 21 the extent to which the former, which is connected with his epistemic scepticism, ends up denying the existence of mind-independent matter remains unclear. What is clear, however, is that regardless of whether the orthodox Kant was an idealist, Langton’s Kant certainly isn’t. Matters of scholarship notwithstanding, Langton’s interpretation of Kantian scepticism is attractive precisely because it de-idealizes the scepticism, and replaces it with the more palatable thesis of epistemic humility. In fact, we need to reject idealism if we are to even consider epistemic humility, for we can’t fail to have knowledge of the intrinsic natures of things, let alone their individual identities, if there are no such things to begin with.
Third, what is the relationship, if any, between idealism and panphenomenalism? Panphenomenalists contend that all instantiated fundamental properties are qualia. Now, qualia, as we noted, are precisely the kinds of things, which enter into the contents of conscious experiences. Thus, panphenomenalism, translated into the old jargon, amounts to the claim that it is ideas, which are the fundamental natures of reality. This is idealism, plain and simple.Footnote 22
Panphenomenalists, of course, may object to this untrendy title on grounds that their view isn’t one where our world lacks physical properties, but rather one where they are grounded in the phenomenal. But so was, arguably, Berkeley’s. Berkeley didn’t deny that there were physical objects or properties. He too just assumed that they were grounded in ideas.Footnote 23 Plus, the problem with panphenomenalism isn’t that it has now been found out to be a form of idealism. Contemporary panpsychists often cite Leibniz as a fellow panpsychist. (Leibniz’s notion of fundamentalia, viz. the soul-like ‘monads’, makes his view sound remarkably like an early day version of M-panphenomenalism). Moreover, they do so despite it being well known that Leibniz is lumped with the idealists.
The problem with panphenomenalism being a form of idealism, rather, is contextual: by assuming panphenomenalism as a means to undermine epistemic humility, we end up with an idealism, which is either as, or perhaps more, ominous than the idealism Langton’s interpretation of Kantian scepticism tries to avoid.Footnote 24 Therefore, assuming panphenomenalism, be it in conjunction with revelation, as a means to undermine the de-idealized humility thesis is self-defeating. Better to stick with Ramseyan humility and be done with idealism.
5. Postscript
In this paper, I have confined myself to looking at exceptions to Ramseyan humility that involve buying into both revelation and panpsychism. I argued that while revelation by itself doesn’t suffice to provide an exception to the humility thesis, it does so when coupled with a panphenomenalist version of panpsychism. I then went on to raise doubt about remedying our ignorance in this way on grounds that the relevant version of panphenomenalism is problematic, especially in the present context. Nonetheless, for all that is said here, it might be the case that we can couple revelation with doctrines concerning qualia other than versions of panpsychism to get an exception to humility. Let me end by briefly considering the prospects of such responses.
One such doctrine is property dualism, as espoused by Chalmers (Reference Chalmers1996). According to this view, there are two distinct types of properties that are fundamental: (1) the properties, whatever they may be, that occupy the causal roles at the fundamental level at our world and (2) the phenomenal properties, including our own, which are nomologically necessitated by the causal roles at our world. To clarify, as in the case of a posteriori physicalism, on this picture, phenomenal properties are necessitated by the standard physical qua causal properties in conjunction with the psychophysical laws. However, there are two crucial differences. First, according to property dualists, the necessity relation is only nomological, as opposed to metaphysical or logical. Second, property dualists concede that the psychophysical laws required to get the necessitation aren’t discoverable empirically by any of the natural sciences. This second difference is the basis for supposing that the phenomenal properties (as well as these laws) are fundamental. Now, if these properties are indeed fundamental, and we know their essences qua identities in virtue of revelation being true, it would then turn out that we do, once again, know the identities of some of the fundamental properties at our world.
There might be other views in the vicinity as well. All we need for an exception to humility – qua knowing the identities of some fundamentalia – is revelation in conjunction with a view that states that whatever properties whose essences are revealed to us experientially are also fundamental. Since revelation is a thesis about qualia, property dualism, which takes qualia to be fundamental, in conjunction with revelation, gives us a straightforward exception to humility in this sense. Nevertheless, other views might get us there as well, albeit in a more roundabout way. For instance, suppose that there are two types of properties, i.e. primary and secondary qualities. How to actually characterise these properties is a point of controversy, but for our purposes, assume the former are intrinsic properties of physical objects, e.g. extension and solidity, and the latter the effects these objects have on us, e.g. sensory qualities like taste and smell. Suppose further that we get fundamental properties of both types, and that our qualia are the type of secondary qualities that are fundamental.Footnote 25 If these two suppositions hold true, and revelation is also true, we would once again have an exception to the humility rule.
What to make of such responses to humility? Obviously, the proponent of humility can resist these cases against humility by attacking whatever views we need to presuppose alongside revelation. Since these views will all have to state that qualia are fundamental, all the stock physicalist responses to qualia, i.e. the ones that motivate eliminativism, as well as reductive and non-reductive versions of physicalism, will prove reasons to resist such views.Footnote 26 These responses are well known, and so I won’t discuss them here except just mention that insofar as you are convinced by any of them, you then have good grounds to uphold the humility thesis.
Furthermore, even if you remain unconvinced, there is a sense in which this way doesn’t really provide an exception to Ramseyan humility. The original humility thesis was the claim that we can’t know the identities of the fundamental properties that occupy the causal roles at our world. Positing fundamental qualia, which are epiphenomenal, says nothing contrary to this thesis. To refute this, we need the conjunction of revelation and L-panphenomenalism. So assume dualism, etc. if you like. But this leaves Ramseyan humility, in its original form, intact.
Funding
This paper was written as part of the project New Directions in the Study of Mind, funded by the John Templeton Foundation.
Acknowledgements
Many thanks to Ben Blumson, Alastair Wilson and Henry Taylor for comments on the draft.
