Appendix

Longer Footnotes

0.1 A Concession to Relativism: Human Centred Metaphysics

The ‘argument’ that Davidson offers against conceptual relativism[1] crucially involved rejecting the pre-conceptual given.  Since conceptual idealism is committed to rejecting this too, it would also seem committed to Davidsonian Absolutism.  But is there not room for some concession to conceptual relativism that is compatible with conceptual idealism?  By investigating such a possibility, we may discover a middle-way between conceptual realism and conceptual idealism.  This possibility presents an attractive alternative to the irresolvable debate discussed in section 3.4.  But there may be a high price to pay: the very possibility of metaphysical explanation that conceptual idealism and conceptual realism attempt to justify.

The realist insists that Reality as it is ‘in itself’ (which we will indicate with capitalisation) is quite independent of our conception of it, and the insight of the conceptual idealist can be captured by pointing out that this Reality must remain quite meaningless to us.  To talk about reality at all, we must talk about reality as it is conceived in some way.  It is essential to reality that it can be conceived, described and made sense of.  However, the conceptual idealist may avoid the criticisms aimed at absolutism by arguing the following way: it is possible that we might come across aliens that ‘conceptualise the world’ in such a different way to us that little or no mutual understanding is possible.  That is to say, we can imagine the situation where the detailed study of creatures convinces us that whatever they are doing, they are not talking about the medium sized dry goods that we normally talk about, nor employing any other part of our ontology.  What they are doing will remain, for all practical purposes, a matter of speculation.  But we can say that if they are employing concepts that are fundamentally distinct from our own, then they must be talking about quite different things to the ones we talk about.  No doubt our concepts are determined in part by our abilities and interests, and other creature may not share those abilities and interests.  If we were so bold as to say that they ‘had a different conceptual scheme to ours’ (and we may well want to remain more cautious, but if we did so say), then we would also be forced to admit that they inhabited (i.e. thought about) a ‘different world’ to us.  Since we can say little or nothing about this ‘different world’, it will (quite literally) make no difference to us.  The job of the metaphysician is to analyse the concepts we use in order to say some about the world we inhabit.  This line of thinking is more inline with Kant’s transcendental idealism, if it is perhaps more cautious than he was (at times) about the notion of a noumenal Reality underlying our reality.

The caution is well advised, for down that route lie the bogies of anti-realism and anti-metaphysics.  If we are only in cognitive commerce with the world ‘as it appears to us’, we will be less inclined to celebrate any metaphysical results of conceptual analysis.  Metaphysics is supposed to be fundamental in that it reveals the true nature of things, not just the way they are to us.  To retain credibility and hold on to claim (i), the conceptual idealist must hold that the world really does contain the things that we tend to pick out: the everyday substances that we talk about (such as human beings, animals, trees, rocks and so on).  One may say that these things exist relative to our conceptual scheme, but then in order to be true to conceptual idealism, one will have to insist that there is no more fundamental ‘absolute’ existence that we can talk about.  Conceptual idealism is committed to rejecting a ‘view from nowhere’, a view uncontaminated by a particular conceptual framework[2].

Such a cautious balance can be seen in the work of David Wiggins, who argues for a moderate essentialism[3].  He defines conceptualism[4] (or ‘conceptualist-realism’) as the conjunction of the following claims:

(1)     “that the possibility of singling out of an object in experience depends upon the possibility of singling it out as a this such;”

(2)     “that there is no surrogate or reductive level (for instance, the level of description of retinal stimulation or whatever),”

(3)     “that our cognitive access to reality is always through conceptions that are conceptions of what it is to be this or that sort of thing, these conceptions being a posteriori and at every point corrigible by experience, yet present in advance of the recognition of any particular object as a this such.”

Claim (2) serves to exclude the possibility that objects are reducible to mere stimuli of a kind that we do not normally take them to be (and phenomenalism and other kinds of empirical idealism are thereby excluded).  Claim (3) is a way of expressing the moderate conceptual idealism under discussion, given its restricted applicability to ‘our cognitive access to reality’.  Note the explicit restriction to a posteriori concepts that are modifiable in the light of experience.  This makes Wiggins’ position compatible with certain realist and externalist conceptions of natural kind terms[5].  Claim (1) can be seen as an application of conceptual idealism for de re thoughts: the possibility of singling out an object depends on bringing that object under a concept – a ‘this such’.

The balance between absolutism and anti-realism that Wiggins is seeking is not easy to spell out.  He talks about a ‘two-way flow’ conception of singling out, acknowledging both the corrigibility of our concepts to experience, and the role these concepts have in deciding what is singled out.  (One metaphor that he has used on more than one occasion in that of a fishing net, the size and mesh of which determine not which fish are in the sea, but which may be caught).  At one point he puts it this way:

The object is what it is, whether or not it is singled out: but it does not simply individuate itself or, in and of itself, differentiate itself from other things.  There are no lines in nature (even though, after the imposition of lines, there are edges for us to find there).[6]

But he admits of the italicised sentence that he has yielded ‘to the temptation to try and convey something by issuing the denial of something that is really nonsense.’  (And surely the denial of nonsense is just more nonsense).  He also urges a cognate criticism of the sentence that follows, which one might suspect of being anti-realist.

The result of this precarious balancing act, if Wiggins does indeed avoid the criticisms that can be levelled at anti-realism on the one hand and absolutism on the other, is a position that is supposed to yield a ‘modest metaphysics’ of essentialism.  Whether or not one admits that the sortal concepts we use are relative to our (human) interests, one can still insist that the objects picked out by them are nevertheless real objects.  Perhaps they are real ‘relative to the human conceptual scheme’, but since we cannot know of any other reality, this qualifier is of little consequence.  If one maintains that it is nonsense (for us) to talk about Reality independent of our concepts (the denial of conceptual realism), then the analysis of our sortal concepts will still have the appeal of telling us about the essential nature of our world.  The question as to whether there are ‘other worlds’ that correspond to conceptual schemes incommensurable to ours can be disregarded as idle speculation.

It is worth comparing this moderate conceptual idealism with Kant’s defence of a modest metaphysics.  At least in the Critique of Pure Reason Kant argued that nothing could be known about the thing-in-itself, considered in abstraction from the faculty of judgement.  This leaves open the possibility of conceptual relativism, since we cannot know that minds very different from our own might conceive of things very differently.  Nevertheless, the analysis of that faculty, and the faculty of intuition, still offered objective knowledge of the nature of reality.  While it was in a certain sense relative knowledge, it could be rendered absolute and universal by qualifying it carefully.  For example, in the Transcendental Aesthetic we find the following defence of knowledge of the phenomenal world:

As to the intuitions of other thinking beings, we cannot judge whether they are or are not bound by the same conditions which limit our own intuition, and which for us are universally valid.  If we join the limitation of a judgement to the concept of a subject, then the judgement will posses unconditioned validity.[7]

This suggests a general formulation of a ‘modest’ conceptual idealism that avoids the problems of the absolutist kind: that our reality is limited by our concepts.  If something is to be considered possible for us, then it must correspond to some concept that we could have.  In retrospect, this human centred modesty is compatible with the formulation of conceptualism given by Morris (that is, with claim (B)).  But it is worth noting that permitting relativism to a conceptual scheme may endanger the commitment to substantial metaphysical claims.  Putnam, for example, draws anti-metaphysical conclusions from just such a combination of positions[8].

Putnam argues against Davidson’s absolutism as follows.  Imagine a world containing three atomic objects, x1, x2 and x3.  How many objects does it contain?  The most straightforward answer is ‘three’, but if you are of mereological bent, you will count 7: x1, x2, x3, x1 + x2, x1 + x3, x2 + x3 and x1 + x2 + x3.  Call a language that only admits 3 objects CL (after Carnap’s preferences) and a language that admits of mereological sums PL (after Polish logicians preferences).  Imagine that at least one of x1, x2 and x3 is red and at least one is black.  If one does not maintain (as the conceptual realist does) that these differing conceptual schemes can be assessed relative to a ‘view from nowhere’, then the question as to which is correct becomes problematic.  It is also not clear how these two languages (with their corresponding conceptual schemes) relate to each other.  Putnam argues (against Davidson) that the question as to whether statements in PL (such as there is an object that is partly red and partly black) correspond in both logical form and meaning to their ‘translations’ into CL (‘There is an object that is red and an object that is black’) is itself relative to a conceptual scheme.  He may have a point, since such a question certainly cannot be answered in either PL or CL, since they cannot even be framed.  The ontological commitments of one language can thus not be said to have any more metaphysical weight than the other.  The moderate conceptual idealist is thus caught between the opposing alternatives of quietism and absolutism.

I will not argue conclusively for the view that moderate conceptual idealism collapses into quietism here.  It suffices for our present purposes to see that it a possibility.  Moreover, this possibility shows how someone who held a moderate form of (ii) and (iii) might consequently come to reject (i).  The relevance of this point will become clearer when we discuss Wittgenstein’s views on philosophy.

0.2 The “Private Language Argument” and the Transcendental Deduction

Some commentators have found similarities between the private language argument and the transcendental deduction.  For example, Scruton: “The transcendental deduction has been revived in recent years, most notably by Ludwig Wittgenstein (1889 – 1951).  In the famous ‘private language’ argument in his Philosophical Investigations Wittgenstein argues that there can be no knowledge of experience which does not presuppose reference to a public world.” (Scruton, Kant, p. 35).  But if some of the implications of their conclusions are similar, the premises they start from are most certainly not (see Philosophical Investigations,  §24).  Furthermore, if McDowell’s discussion of Wittgenstein’s remarks on private language can be seen an attempt to interpret it as Kantian – allowing sensations as “a limiting case of the model of object and designation” – then the discussion in chapter 5 shows some salient differences between the two arguments.

0.3 Quietism and Disjunctivism

I want to pause here to briefly compare the above line of reasoning with another different but related argument.  This is a line of reasoning that leads one to a disjunctivist position with respect to sensory experience.  The form of argument has something in common with the quietist approach under discussion, but there are also important differences.

It is a common philosophical thought that what we experience is a representation of the world, rather than the world itself.  We can see the appeal of this thought by considering the cases of illusion and hallucination.  There are certain situations, created by abnormalities of either the environment (in the case of illusion) or one’s mental state (in the case of hallucination), where the experience one has is qualitatively indistinguishable from veridical perception, but is nevertheless caused by different conditions.  Compare the situation where I am looking at a pink elephant to the case where, unknown to me, some chemical has changed my mental state such that it seems to me as if there were a pink elephant in front of me, when in fact there is not.  In the one case my experience is caused by the presence of a pink elephant, in the other there is no such direct causal connection with a real animal. Assume that the two experiences are otherwise identical[9].  They both have the same vivid qualitative feel.  I am apt to behave in identical ways.  If the elephant is charging in my direction, I will take evasive measures.  I am also apt to form the same beliefs.  For example, that I am in immediate danger.

Now, in the case of hallucination, it seems clear that I represent the world to be a certain way, when, in fact, it is not.  It is natural to conclude that in this case what we have immediate perceptual access to is a representation of the world, rather than the world itself.  What exactly this means, is, of course, a substantial philosophical question.  For the sake of the present comparison, let us assume that it means we have immediate acquaintance with something like sense data: mental entities that can misrepresent the world.  Now, given that the phenomenology is identical to the case of veridical perception, it seems parsimonious if we postulate that in this case, too, what I have immediate access to is not the world itself, but a representation of the world.  The difference is that in the former case, but not the latter, the representation is correct.

This line of reasoning has convinced many of the need to postulate sense data or qualia.  Whatever other problems there are with it, there is one point that makes the conclusion seem radically misguided.  It just seems false that what I have immediate access to is only a representation of the world.  When I veridically perceive an elephant, what I experience is an elephant, not some mental copy of one.  To assume otherwise is to assume a disassociation from reality that has struck some philosophers as intolerable[10].  It means, for example, that when we refer to objects in our experience, we are mistaken about the nature of those objects: we are really referring to collections of sense data.  But that is not what we find in experience.  We find the world.

In some ways this conflict of intuitions is a mirror of the considerations that brought us into tension over ineffable qualia.  On the one hand, we have an irresistible line of reasoning to the existence of representational mental entities.  On the other, the brute force of a basic intuition.  I do not think we need place too much emphasis on the difference between basic intuitions and the result of argument here.  In both cases the results of the arguments can be replaced by (similar) brute intuitions, and arguments can be found for the basic intuitions. This is at least hinted at from the mirrored structure of the two cases[11].  The issues involved in the two cases are somewhat different, however.  I have said nothing in the argument for sense data that would indicate that they are ineffable.  Of course, as discussed above, the temptation to assume they are is almost irresistible.  But that is a separate issue, and it does not obviously follow from the reasoning presently under consideration.

One way to resolve the conflict that the argument for representational perception gives rise to is to live with a quite different explanation of hallucinations from that of veridical perception.  That is, we give up on our desideratum of a unified account.  In the one case we have direct perception of the world, in the other we have a misrepresentation of the world.  Such a disjunctivist position is quietist in the following respect: it accepts that while we would like a philosophical explanation here, none can be given.  It accepts that no philosophical explanation can be given of the fact that hallucinations can have the same qualitative phenomenology as veridical perception.  That does not necessarily rule out a scientific explanation.  It maybe that there is an explanation in terms of common brain states[12].  But it denies that postulating sense data or other representational entities in the case of normal perception can gain us any philosophical insight.

A quietist position might go even further than this, though.  It is also possible to deny that the postulation of any representational entities will provide insight into the nature of experience – veridical or otherwise.  I don’t wish to argue that here.  I have taken this brief digression only in order to indicate some of the similarities and differences between the quietist position on qualia and the disjunctive approach to the philosophy of perception.