Greg Detre
Saturday, 11 May, 2002
written after reading Sarah W's PI essay
the private language arg seems to contain more than one line of argument
we cannot point to individual sensations (private ostensive definitions), because we cannot delineate them, we cannot cleanly parcel up our perceptual world into individual sensations
e.g. just as we can't point to the division between two partially-mixed liquids in the same container...???
maybe it only makes sense to talk of our entire sensation picture (i.e. the brain as a fully-distributed system), that it can't be divided into independent components as is necessary in order for us to use syntax
along the same lines: we have no criteria of identity for sensations, that is, no means of saying for sure that THIS is the same sensation that i remember from last week
i.e.(???) we have no rules for using our self-devised symbol 's'
this is kind of like the problem of induction (and the Goodman paradox) - when a new sensation comes along, how will we know whether it is an 'S' or not, given that we can't fully specify a sensation in such a way that we can know whether to include or exclude examples - i think this is what he means when he says that in calling something 'S' we don't have rules for using 'S' - see the 'it takes two to make a language' point below
hang on - what about if you can fully specify a sensation so that you have rules for saying whether or not some new sensation should be deemed an 'S' or not???
e.g. a number - bad example, cos Child says it's not really a sensation - so what IS a sensation then??? or more specifically, what's wrong with calling a concept of number a sensation (since after all, it's a linguistic concept by which we aggregate experience for which we have public (behavioural???) criteria)??? what did Child say to my question about this in the tutorial???
this in turn relates to the problem of memory - that we don't know for sure that our memories are reliable
(see Dennett on Chase and Sanborn??? is Dennett's Chase/Sanborn (i.e. Stalinesque/Orwellian) argument the same as this, or is it saying something more subtle about the problem of private, incorrigible qualia, which fits with Witt???)
Witt is also attacking our notion of qualia in general as private, real and incorrigible ('objectively subjective') - again, see Dennett
a language is a distributed system - you cannot have a language with one speaker, because you would have no means of verifying whether the way you're using that language is drifting/reliable
without a pair of speakers to anchor each other, i.e. if the private linguist doesn't have someone whose sensations he can directly compare with his own (i think this would have to be simultaneously somehow, to avoid memory problems) ... then what(???)
I think this maybe is related to Quine�s idea of �observational sentences�, and because we have no means of exhibiting private samples, we have no means of anchoring our language to a set of exhibited sensations that we can term �S� � but if we ignore the memory objection, we can
(this is kind of like the problem of memory - in fact, is it at all different??? - it cannot be simply saying that language needs two people because communication is what is adaptive about it, or anything dumb/evolutionary like that)
i don't know whether Witt advances this argument, but there seems to be an inherent problem with the idea of comparing sensations full stop - no, that's not true, cos i can conjure up an image of (say) both red and yellow at the same time and compare them
is there any problem with our attention mechanism in general, with the idea of an inner sideways glance??? that doesn't seem to be Witt's point
strong/logical behaviourism - a sensation *is* its public criteria
perhaps we can only learn about sensations through other people, i.e. learning about *their* sensations by watching their reactions and imputing those behaviours onto ourselves (or vice versa) (kind of like an inverted simulation theory of mind) - this seems wrong, and nowhere in Witt
verifiability and behaviourism??? logical positivism??? is he making an epistemological or metaphysical claim???
beetle example - seems to imply that we *do* all have qualia/sensations (no, he doesn't � there might be nothing, see bit on Wittgenstein in Dennett on quining qualia), but simply that we have no means whatsoever of comparing *our* qualia to anyone else's - and so that any linguistic description of them must be referring to the public manifestation of sensations, rather than the private experience of them - in order to reject a PL though, this arg seems to need to be supplemented by either:
the arg that it takes two people to form a language cos they need to anchor each other (which i don't think is really Witt's point, and i'm not too sure i can really make convincing at all)
the arg that i don't *know* i am in pain, i just *am* in pain etc. ...???
or the args that i have no means of saying whether this 'S' is the same as last week's 'S' cos i don't have criteria of identifying it - but then what work does the beetle example do in this case??? ahh, it seems to be saying that the *only* criteria are *public* criteria
does a modern/scientific understanding of mental mechanisms/processes help at all???
e.g. NNs' associative/classifying properties as subconscious means of anchoring a symbol representation with a brain-state/sensation- representation???
or using an MRI (permanently projected onto our retina (a la Greg Egan's patch) to identify a given sensation in terms of its NCC???
know vs have pain - the grammar of our pain-concept
what does it mean to say that someone else has a pain...??? see sarah essay on the two+ components of this idea
verifiability conditions � do we have the same problems about the intelligibility of talking about �someone else�s pain� as talking about �did it rain yesterday� stuff...???
concept as skill???