Friday 11 June 2021

First thoughts on starting to read Annalis Coliva’s Moore and Wittgenstein

I’ve been reading Annalis Coliva’s excellent book on Moore and Wittgenstein: Scepticism, Certainty and Common Sense (2010). The discussion of Moore in chapter 1 seems very helpful. In chapter 2, however, I’m torn between admiring the clear exposition of Wittgenstein’s On Certainty and a sneaking feeling that she subscribes to one aspect of the sort of thing I used to say about On Certainty and now feel uncomfortable with. My hunch is that she has a relaxed construal of knowledge which resembles something like a true belief based on a good enough reason. Further, I think this is motivated by focusing too much on the claim ‘I know’. Spotting that that is fallible – one can claim to know when one does not know – she then locates the fallibility too closely, I fear, with the state of knowledge itself rather than with its first person self-ascription. Still I'm early in the book and may be quite wrong or, equally, later persuaded.

For example, is the following merely a slip or is it significant? First she quotes On Certainty and then picks up a line of thought. 

According to Wittgenstein, here are some examples of the correct use of ‘I know’:...

Someone with bad sight asks me: ‘do you believe that the thing we can see there is a tree?’ I reply ‘I know it is; I can see it clearly and am familiar with it’. – A: ‘Is N. N. at home?’ – I: ‘I believe he is.’ – A: ‘Was he at home yesterday?’ – I: ‘Yesterday he was – I know he was; I spoke to him.’ – A: ‘Do you know or only believe that this part of the house was built on later than the rest?’ – I: ‘I know it is; I got it from so and so’. (OC 483)

In these cases, then, one says ‘I know’ and mentions how one knows [das Grund], or at least one can do so. (OC 484) (Coliva 2010: 62)

A paragraph later she is able to offer the description and conclusion she wants which runs:

[T]he fact that in general our use of ‘to know’ in the first person present be made on the basis of grounds – whatever they contextually happen to be – is a criterion for saying that that use of ‘I know’ was meaningful though, perhaps, mistaken. Thus, if I had talked to NN’s twin brother this would (probably) falsify the claim that NN was at home and would (certainly) show that I did not know, but only believed to have talked to him. My claim ‘I know that NN was at home yesterday’ would thus be false, but my use of ‘I know’ would not be nonsensical. By contrast, if I claimed to know that NN was at home yesterday but if I were unable to offer reasons in favour of such a claim, that would show that I misused the expression ‘I know’. (ibid: 62-3)

A condition on the ordinary use of ‘I know’ – she says – is that one is able to offer grounds. But one may be in error, or merely not know, if, for example, the grounds offered fail, for example if one has mistaken his twin brother for NN. But the passage just before that gets her to this conclusion runs:

Let us consider the second of Wittgenstein’s examples from the above mentioned quotation. The fact that I talked to NN yesterday and offer it as my ground for saying that I know he was at home is a criterion for the correct assertion ‘I know that…’. A criterion is neither a necessary nor a sufficient condition for knowledge that p. For, clearly, I could have talked to NN’s twin brother, and NN could have been at home while I had no conversation with him. (ibid: 62 both italics added)

Here Coliva seems to think that the fact that I talked to NN fails to act as sufficient grounds for the claim to knowledge because – for all I know, let’s assume - I could have talked instead to NN’s twin brother. That possibility of error implies, she says, that actually talking to NN is merely a criterion on the model of criteria where they are neither necessary nor sufficient for knowledge. But with no other undermining circumstances described, why is (the fact of) talking to NN at his home not sufficient for knowing him to have been at home? 

This seems to hint at a highest common factor of experience in which one could never take in the fact of talking to NN – because it is possible to confuse NN and his twin – and hence the invocation of that fact as grounds for a knowledge claim should be interpreted merely as a highest common factor analogue. Not the full fact as she actually states it but that merely under an interpretation of what it shares with talking to his twin brother: seeming to talk to NN (or his twin). (Of course, it does change things if his twin lives with NN and I know that I cannot tell them apart. In that case the fact of seeing NN rather than, as it happens and by luck, his twin cannot impress itself on me and hence my capacity for knowledge is undermined. But none of this has been specified.)

This looks like the difficulty I’ve seen elsewhere of drawing conclusions for knowledge from the fallibility of claims expressed via ‘I know…’.

Here’s a longer passage a little later from pp74-6. Coliva here follows Wittgenstein so my qualms may be with him rather than her. But she does endorse the view that emerges. I will intersperse lengthy quotation with my comments.

Wittgenstein, as we saw in §2.2, thinks that ‘the difference between the concept “knowing” and the concept of “being certain” isn’t of any great importance at all’ except, however, ‘where “I know” is meant to mean: I can’t be wrong’ (OC 8). When we say that we are certain ‘we express complete conviction, the total absence of doubt, and thereby we seek to convince other people. That is subjective certainty’ (OC 194, cf. 245, 415). But – asks Wittgenstein – ‘when is something objectively certain?’ (ibid. my italics); and answers, ‘When a mistake is not possible’ (ibid.). That is to say, when the possibility of being mistaken is ‘logically excluded’ (cf. ibid. my italics). (Coliva 2010: 74-5)

So knowing and being certain are not substantially different except in this context though note that it is still explored via an ‘I know’ claim (rather than, eg., looking at when we might properly ascribe knowledge in the third person). When ‘I know’ is meant to mean: I cannot be wrong, then there is a difference. It would be good to know what this is.

Well when we say ‘I am certain’ that expresses, she says, complete conviction, the total absence of doubt. So that sounds to be equivalent of saying ‘I cannot be wrong’. So that sounds just the same as the use of ‘I know’ we’ve just had to mean the same thing: I cannot be wrong. Hence I’m confused.

The passage from Wittgenstein continues with the distinction of subjective and objective certainty. My hunch is that subjective certainty is a status that applies to the claims ‘I know’ and ‘I am certain’ - both expressive of the total absence of doubt - whether or not they are true. That is, one is subjectively certain if one claims knowledge falsely. We might say: although Jones was certain, in his own mind, he turned out not to know and, in fact, to be quite wrong. (There is no equivalent: subjectively know, ie. falsely.) Objective certainty applies to a combination of subjective certainty and the idea that mistake is actually not possible.

Given that subjective certainty did not deliver the promised distinction between knowing and being certain, is the idea that a distinction enters only with objective certainty? This would falsify the initial too wide claim but understandably so. But if so it implies that Wittgenstein and Coliva hold that knowledge is compatible with some possibility of being wrong: being right though one might have been wrong; being right though only by luck. But that seems to me to undermine the very idea of knowledge. 

Note that it is perfectly possible, as above, that the claim ‘I know’ is false but here we seem to be talking about knowledge itself not merely false claims to have it. She continues:

At another place, however, he writes,

‘Knowledge’ and ‘certainty’ belong to different categories. They are not two ‘mental states’ like, say ‘surmising’ and ‘being sure’ (…). What interests us now is not being sure but knowledge. That is, we are interested in the fact that about certain empirical propositions no doubt can exist if making judgments is to be possible at all. Or again: I am inclined to believe that not everything that has the form of an empirical proposition is one. (OC 308)

But, as we saw in §2.2, sometimes ‘I am certain’ can occur in place of ‘I know’, and yet, Wittgenstein now says, ‘I know’ doesn’t express a subjective conviction. (ibid: 75)

This is puzzling assuming - possibly utterly wrongly - my approach above. I’d like to say that ‘I know’ expresses (it does not ensure) a (hence fallible) claim to objective certainty though we might – on the suggestion above – ascribe to the person ‘subjective certainty’ should they prove to be in error about the facts. Perhaps the latter could be flagged as indicated. The sincere claim expresses an objective claim about the world that may prove false and indicates that the speaker holds it with great conviction, which still holds even if they are wrong about the world. I am yet to see why this is helpful, however.

This makes it puzzling to see how for Wittgenstein ‘I am certain’ can express a subjective conviction and thus a mental state. In the contexts in which ‘I know’ can be substituted by ‘I am certain’, I suggest to interpret both these claims as in fact equivalent to the following one: ‘for all I know, things are thus- and-so’. For this last assertion makes clearer how my certainty is here a function of the information, evidence and grounds I have in favour of a given hypothesis. This, however, doesn’t guarantee that things are as I – with good right – claim them to be. Thus, one can say that my certainty is subjective, as it is a function of a subject’s state of information. (ibid: 75)

There seems to be something strange and very wrong about glossing the sincere claims ‘I know’ or ‘I am certain’ as for all I know, things are thus-and-so. If someone says that, then they ought to withdraw the claim ‘I know’ or ‘I am certain. Unless it is an empty philosophical musing, ‘for all I know’ undermines the claim of either. Lots of sentences about cultural practice in distant countries can follow my own use of ‘For all I know’ and none of them are then knowledge (they would be mostly false, even).

This doesn’t mean that the claim ‘I am certain’ would somehow be arbitrary and whimsical. Simply, my certainty doesn’t exclude the possibility of error: I could have overlooked or misinterpreted some evidence and wrongly taken it to support ‘p’; or else, I could fail to have some decisive piece of evidence against it. (ibid: 75)

To the contrary, I am inclined to say that the previous suggestion about ‘for all I know’ really does make the follow up claim ‘I am certain’ arbitrary and whimsical! It ought to be withdrawn immediately by a competent speaker. The rest of this seems, again, to ignore the difference between what is claimed and what might prove to be the case. ‘I know’ is fallible. Again, I suspect the problem comes from looking at first person rather than third person ascriptions.

So, we can make sense of Wittgenstein’s claim that, in this case, one’s certainty would be subjective. Yet, can we also make sense of his claim that it would then be a mental state? After all, it won’t differ from knowledge and Wittgenstein, as we saw in §2.1, stressed the fact that knowledge isn’t a mental state. He writes,

One may for example call ‘mental state’ what is expressed by tone of voice in speaking, by gestures etc. It would thus be possible to speak of a mental state of conviction, and that may be the same whether it is knowledge or false belief. (OC 42)

I think that the kind of certainty Wittgenstein dubs ‘subjective’ isn’t itself a mental state, just as knowledge isn’t, for him, but that it can be accompanied by a particular ‘mental state of conviction’, that can be manifested in the tone of voice and in one’s manners. That there be such aspects of ‘tone’ and that there be also a particular mental state of conviction – a special feeling, as it were – is entirely plausible, from a phenomenological point of view. This, however, doesn’t exclude the possibility of error. I may feel as certain as ever of having taken my keys this morning when I left home, and yet find out later that I didn’t. A particular feeling is not a sufficient condition (nor a necessary one) for the truth of the hypothesis one claims to be certain about, as it could occur both when one really knows that p and ‘p’ is thus the case and when one doesn’t know it because in fact ‘p’ isn’t the case. So one’s claim to be (subjectively) certain and its accompanying feelings of total conviction don’t guarantee the impossibility of being mistaken. (ibid: 75-6)

This seems to be making unnecessarily heavy weather of the fact that claims to know are fallible. But the suggestion somehow lurking here that one might self ascribe knowledge – in an ‘I know’ claim – via introspection of feelings of certainty seems bizarre. She says: A particular feeling is not a sufficient condition (nor a necessary one) for the truth of the hypothesis one claims to be certain about. Who knew? But without that bizarre idea, I am not sure what this paragraph claims. Note that one would not normally claim to be subjectively certain unless, perhaps, one realises that one has taken a drug and has become unreliable. One might, perhaps, say: I am weirdly subjectively certain about facts I couldn't possibly know so please ignore me.

Wittgenstein, therefore, isn’t interested in this kind of certainty [TT: a good thing too since it seems silly], nor, in this case, in the use of ‘I know’ within our usual language games, where it can be substituted by ‘I am certain/sure’. Rather, he wishes to clarify that peculiar use of ‘I know’ which is a synonym of ‘I can’t be wrong’ and which expresses objective certainty (OC 569). (ibid: 76)

But didn’t we have this assimilation earlier applied to subjective certainty? Perhaps not if the conjunction is played up. That is, he is interested in the use of ‘I know’ not just where it can be substituted by ‘I am certain/sure’ but where, in addition to that substitution, it actually also possesses objective certainty. If so, though, the word ‘express’ is in the wrong place. ‘I know’ expresses, fallibly, a claim to objective certainty. But also: surely this just means cases where one’s knowledge or certainty claim successfully excludes the possibility of mistake. And that still does not distinguish knowledge and certainty.

Still, I've two chapters to go and it may be that the butler really does it. I post this mainly in the hope of getting quick corrections to my email address.

Coliva, A. (2010) Moore and Wittgenstein: Scepticism, Certainty and Common Sense, Basingstoke Palgrave Macmillan