Thanks to Jones for his post to the Grice Club.
I tried the Commentary section, but I may find it easier to provide the commentary in a different post.
Jones writes:
"Turing's "imitation game" is sometimes thought of as providing a definition or analysis of the concept of thinking or of intelligence and as such it may seem crude by the side of the subtle analyses of language we find in Grice."
Indeed. One should revise, especially, I was thinking, the analysis of 'mean', and 'think', in one of the least read of Grice's "Logic and Conversation Lectures": the last one, number seven.
While written in 1967, it may be seen as a response to Ryle, and why not Turing, since it's all about the psychological relevance of 'mean' as a 'thinking' process.
Jones continues:
"In twentieth-century analytic philosophy the most profound cleavage between schools of thought might be seen as that between philosophers such as Russell who saw failures of reasoning as arising (at least in part) from pathologies in natural languages, and advocated the use of idealised formal languages as a remedy, and those whose interest was primarily in the analysis of those natural languages, warts and all (and would not recognise the features of which Russell complained as pathological)."
Indeed. But we should recall that Grice speaks in the MOST FAMOUS of the "Logic and Conversation" lectures -- number II -- of a 'common mistake' between formalists and informalists, or what he would later call neo-traditionalists (like Strawson -- I think the 'neo-' is ironic, since a traditionalist, like a conservative, hardly wants anything 'new') and the modernists like the 'heirs to PM' (Principia Mathematica). We should also be reminded that amongst Grice's unpublications there's a lovely,
"Definite descriptions in Russell and in the vernacular"
that Beale makes much use of in his "Content" book -- and that some member of the Bertrand Russell Society should pay serious attention for. For one reason, it mentions Russell explicitly, when Grice, as an Oxford man, had no such obligation even to mention the Cantab. competition!
Jones goes on:
"To read Turing as if he belonged to the latter school may serve to illustrate some nice philosophical insights, but his opening paragraph in the classic paper surely places him firmly in the former camp."
Good to be reminded of this, and thanks for the quote.
What does the quote read. Let us consider it in some detail. I always read it as a compliment to RYLE, whose 'variety' of 'ordinary language philosophy' does not quite fit Grice's and Austin's "Play Group" model (recall Owens's obituary of Ryle for the Aristotelian Society, where he notes that it was Austin's, and later, after Austin's death, Grice's group, that attracted more of a fan basis, than the more pro-Establishment group that gathered around Ryle -- a previous generation, too --: Ryle, Wood, Owens, and a few others.
Ryle did practice 'ordinary language philosophy', but not alla Austin or Grice.
In generational terms, Turing belongs with Grice. (Grice born 1913, Turing born 1912 -- Ryle born 1900).
Turing writes:
"I propose to consider the question, "Can machines think?"".
In the film "The Imitation Game", the detective makes the question, and Turing then dubs it 'stupid'. He expands on this stupidity. He thinks that if we can say that Jones, say, thinks different from Smith, we "might just as well say" that a machine thinks different from a human.
I should revise the screenplay, which is genial!
Turing goes on:
"This should begin with definitions of the meaning of the terms "machine" and "think.""
I always love to qualify this to say: "and can, of course". Because 'can' is perhaps the most important verb, philosophically. It introduces a MODAL claim, as Peacocke (professor of metaphysics at Oxford, sometime) and Kant would have it.
Consider:
i. Can machines think?
ii. Should machines think?
iii. Do machines think?
In terms of implicatures, as per the screenplay to "The Imitation Game", we should also consider the implicatures it triggers. In this case, the implicature that someone or other (a detective in Manchester wonderfully acted by Rory Kinnear in "The Imitation Game") DOUBTING that machines can think.
Turing goes on:
"The definitions might be framed so as to reflect so far as possible the normal use of the words, but this attitude is dangerous."
By normal, we should assume Turing means what Austin and Grice (and perhaps Ryle) would call 'ordinary'. I never actually did like 'ordinary', since I tend to use the hyperbolic variant of it all too often ("Extraordinary!"). But 'normal' has its own problems too.
"If the meaning of the words "machine" and "think" are to be found by examining how they are commonly used it is difficult to escape the conclusion that the meaning and the answer to the question, "Can machines think?" is to be sought in a statistical survey such as a Gallup poll."
Here Turing turns from 'normal' to 'common', which I dislike too. As when we say that Smith is a 'common' name. (In fact perhaps Smith IS a common name, but that is neither here nor there -- and in any case, if Smith thinks that Smith is a common name he can always qualify it, as Nowell-Smith did!).
Turing goes on:
"But this is absurd. Instead of attempting such a definition I shall replace the question by another, which is closely related to it and is expressed in relatively unambiguous words."
Jones comments:
"Turing side-steps questions about thought and intelligence, substituting his imitation game."
Indeed, he substitutes what Grice would have as a conceptual analysis of 'think' with a 'thought-experiment'.
It is always or often said that Grice created such thought-experiments (more specifically Martian-type thought-experiments) with his "Some remarks about the senses", in WoW) so there shouldn't be a problem there.
A conceptual analysis of 'think' should provide sufficient and necessary conditions for claims such as:
A thinks.
Or
Rodin's Penseur thinks.
(I think this is what Ryle attempts in his book on THINKING -- NOT "The concept of mind" -- and we should double check how close to Turing he gets, since it is acknowledged that Turing (not a philosopher as Hodges hastes to repeat in his Stanford entry for the man) got his essay published in 'a journal of philosophy' as Mind was because it was liked by its editor, Ryle.
Jones goes on:
"Instead of claiming that computers can think, he claims that they will at some point succeed in some more sharply defined accomplishment."
This seems to imply (or implicate) that they 'can'.
"Can" can trigger some implicatures. Sinnott-Armstrong made part of his career on this:
Does "can" imply "ought"?
was a title of one of his essays. Sinnott-Armstrong concludes that "ought" conversationally implicates "can".
Turing and Jones are here playing with 'can' and 'will'.
Pigs can't fly, but they will.
Implicature watch, there! :)
But a modal-claim should perhaps (some think) NOT be temporally qualified like that. Opponents to AI (Artificial Intelligence), some sort of Griceian-oriented, like Searle, or Haugeland, or Dreyfuss, would protest that the 'can' in "Can machines think?" should NOT be temporarily qualified. Philosophers should be concerned if, as a matter of principle, machines do think. Perhaps.
Jones goes on:
"This kind of side-stepping is common among those of a positivistic temper, and hence this is more like what Carnap might have done than the work of Grice."
Point taken.
"There is still much room for debate about the merits of Turing's choice and the influence which that choice has had upon the development of the discipline of AI."
Indeed.
"Gibert Ryle is an interesting philosopher in this regard. He may easily be thought of as belonging to the latter camp with Grice his student, but this placement is easily contested."
Yes, Grice may be regarded as a student of Ryle. Grice has big words of appreciation for Ryle in "Prejudices and predilections, which become the life and opinions of Paul Grice" (aka "Reply to Richards"). But he is concerned on how Ryle's attitude sort of improved the atmostphere of Oxford at a time when Ayer was 'L'Enfant terrible', as Grice calls him.
Grice is seeing Ryle as a good Oxford administrator man in the Philosophy Sub-Faculty.
(Grice's actual tutor was Hardie -- but it is true that Grice's first published essay was ALSO published in "Mind" when Ryle edited it -- "Personal Identity", 1941).
The irony is that Ayer's positivism, as brought to Oxford, ultimately was caused by Ryle, when he advised Ayer "to go to Vienna", which was perhaps merely an implicature that Ayer took literally.
Jones goes on:
"One of [Ryle's] earliest papers places him close to Russell. "Systematically misleading expressions" engages in analysis of just that kind of "pathology" in natural language which provoked Russell's advocacy of idealised languages."
Indeed. And a beautiful piece of analysis it is too. Of course Grice was ALL into 'mislead'. Grice's motto may have been,
"Misleading, but TRUE".
"Some nightingales sing beautifully" (I'm recently fascinated by this song, "The long long trail, so pardon the ornithological reference, rather otiose).
If it turns out that ALL nightingales sing beautifully,
"Some nightingales sing beautifully"
may be thought of as misleading, and perhaps even 'systematically misleading'. BUT TRUE!
But one should revise what Ryle was actually onto!
I particularly love his analysis of the "Fido"-Fido theory of meaning.
(All of this has fortunately been properly reprinted -- and there's even a Ryle Memorial Lecture in Canada -- Ayer was one of the lecturers -- but of course they don't have to deal with Ryle. Ayer dedicated his to Hume, out of which "Hume" in the Past Masters Series edited by Oxford University Press arose).
Jones goes on:
"Later [Ryle] might be blamed for sending off the infant terrible"
Lovely we chose Grice's expression for Freddie (as his friends call him)! I wonder if this French idiom originated as having to do with SOME real 'infant terrible', somewhere. Or is the implicature that, on the whole, French infants are more terrible than infants of other nationality. I should be surprised if the French, when referring to a terrible infant use the English, 'terrible infant', as a sort of vengeance. (Note incidentally that 'in-fant', means 'he cannot speak', which does not apply to Ayer AT THAT TIME).
Jones:
"... to Vienna facilitating Twentieth Century Positivisms most persuasive polemic "Language Truth and Logic"."
Indeed. And this was early enough. Ryle's first thing was apparently his review in "Mind" of Heidegger's "Sein und Zeit".
Ayer was for a time a member of the Thursday play group, again led by Austin, which met at the prestigious All Souls -- Hampshire and Hart -- another Anglo-Jewish, like Ayer --, and quite a few other classy undergraduates were part of it. Vide Berlin (another Anglo-Jewish), "Austin and the early beginnings of ordinary language philosophy" in Berlin et al, volume on Austin. Meanwhile, Grice, as he later confessed, felt an outsider, 'having come from the wrong side of the tracks', a mere 'scholarship boy' who had ended in Corpus Christi (a Midlands-oriented college), only because his forte at school (Clifton) had been "Classics".
Why Ayer changed the lovely atmosphere of those quiet Thursday evenings at All Souls just because his tutor said "Go to Vienna!" is perhaps a proof that Vienna has appeals other than dreaming spires...
(Vide Toulmin, "The Vienna of Wittgenstein").
Jones goes on:
"Even [Ryle's] mature magnum opus seems to have as its principal objective one which a positivist might approve, the extirpation of the "Ghost in the Machine"."
Indeed. It's a lovely phrase, since 'ghost' is essentially cognate with 'Geist', and of course the "Holy Ghost" of the Anglicans and Catholics. (But the Anglicans and Catholics don't speak of machines so freely). Not to be confounded with the 'holy goat' (In "Four Weddings and a Funeral, "Mr. Bean" makes this malaprop as "Father Gerald":
In the name of the father, the son, and the holy spigot. Spirit!
In the name of the father, the son and the holy goat. Eh... *ghost*.
---- But I wouldn't be surprised if Ryle is translating from Descartes.
On the other hand, let us not forget the SECOND noun in that lovely phrase, 'the ghost in the machine': 'machine', which Turing worshipped.
Indeed, loads have been written, I think, on Turing's use of 'mechanical', as the adjectival correlation for machine. And right he is.
--- INTERLUDE ON "MECHANICAL" AND "MACHINE"
Main modern sense of "device made of moving parts for applying mechanical power" (1670s) probably grew out of mid-17c. senses of "apparatus, appliance" and "military siege-tower." In late 19c. slang the word was used for both "penis" and "vagina," one of the few so honored. Political sense is U.S. slang, first recorded 1876. Machine age is attested by 1851:
The idea of remodelling society at public meetings is one of the least reasonable which ever entered the mind of an agitator: and the notion that the relations of the sexes can be re-arranged and finally disposed of by preamble and resolution, is one of the latest, as it should have been the last, vagary of a machine age. ["The Literary World," Nov. 1, 1851]Machine for living (in) "house" translates Le Corbusier's machine à habiter (1923).
---- END OF INTERLUDE.
Jones concludes his interesting post:
"In this one might see Ryle as more positivistic than Carnap, whose ontological pragmatism might be more sympathetic to mental language."
Indeed. The keyword here seems to be "ANALYTIC BEHAVIOURISM", what Ryle is defending. Grice considers it in his "Method in philosophical psychology: from the banal to the bizarre", and in earlier publications, "Intentions" (circa 1946), where he favours a dispositional account to 'intention', having Ryle in mind.
Oddly, in "Method", in Block's reading of it, Grice proposes as an alternate to Ryle's analytic behaviourism something like an application of a Turing-machine.
We have a squirrel (or squarrel, as Grice prefers) eating nuts.
We see the input: the squarrel seeing the nuts because the squarrel is hungry.
And then we see the output: the squarrel eating the nuts.
Grice is interested in the 'black box' -- the psychological or
ψ
predicate as Grice prefers, now seen as a 'theoretical term'. Thisψ
is seen as a mere 'connector' between input (perceptual recognition) and output (behavioural manifestation. Grice sees this obviously NOT as 'behaviouristic' in Ryle's way, since he is not trying to ELIMINATE
ψ
Rather, Grice is EXPLAINING
ψ
in terms of its role in accounting for the passage between the input and the output.
For all the banal-cum-bizarre typical Griceian parlance here, it is said that Grice is drawing mainly from D. K. Lewis, who would be the first to have applied Turing's ideas to the philosophy of mind.
But Grice does it in his very own Kantotelian way!
Cheers
Posts
No comments:
Post a Comment