In this article, Robert Damper attempts to bring out the benefits of a quasi-modal formalization of thought-experiments, with a direct application on John Searle’s Chinese Room Argument (CRA, henceforth). Since thought experiments are clearly dealing with what can or could be the case, and reasoning in some manner from these “ifs and cans” – to borrow Strawson’s title – the step seems only fitting.
Here’s a concise restatement of Searle’s CRA:
Searle envisages a situation in which he is hidden in a room and is presented questions in Chinese written on an ‘input’ card, posted in to his room by unseen enquirers. Searle knows no Chinese; indeed, he is quite unaware of the enterprise in which he is engaged and is ignorant of the fact that the strange marks on the cards represent questions framed in Chinese. He consults a manual telling him (in English) precisely what equally strange marks to write on an ‘output’ card, which he posts back to the outside world. By virtue of the ‘machine intelligence’ embodied in the manual (which is actually a formalisation of the steps in an AI program), these marks on the output card constitute an answer to any input question. […] Searle concludes that an AI program could give the impression of intelligence to an external observer, but have no understanding. This is contrary to the tenets of strong AI—essentially that computational states are functionally equivalent to mental states—as exemplified in the (then contemporary) work of Schank and Abelson (1977), McCarthy (1979), Newell (1980) and others. (p. 165)Amongst many replies to this thought experiment, the one called “systems reply” is the most debated and to which many rejoinders have been given on each side. Damper summarizes the discussion as follows:
- Systems reply: intelligence resides in the total system not just in Searle himself, who is merely a component. Because a part of the system (i.e. Searle) does not understand Chinese, this does not mean that the complete system does not understand Chinese. 166
- Searle claims to have the decisive rebuttal which he calls the ‘outdoor’ CRA. He argues that he simply (!) ‘internalizes’ everything, committing the entire manual to memory, and then proceeds as before. There is then, he argues, nothing but the human in the system who still does not understand Chinese 166
- The popular rejoinder by proponents of AI to this manoeuvre is to point out that it begs the question by assuming the truth of the CRA
In these terms, we replace the general model with the particularities of CRA:
1. The modal source statement is the ‘theory’ of strong AI, namely that executing an AI program is necessarily constitutive of understanding.
2. The modal source statement S implies the logical necessity that any implementation of the Chinese-understanding program understands Chinese, in accordance with the tenets of strong AI.
3. If Searle were to hand-implement any such program then he would understand Chinese—a ‘weird’ consequence, W.
4. Searle does not understand Chinese (this is an ‘absurdity’).
5. It is possible that Searle can hand-implement the Chinese understanding program.
From the fact that 1-5 are inconsistent (in some way, usually logical), the thought experimenter draws the conclusion that S must be wrong.
The advantage of doing this – and continuing to work with this framework in dealing with the replies, too – is that we can see, in a more transparent way, the manner in which the thought experiment is being put forward. For instance, we can clearly delineate, in the case of a “refuter” (or “destructive TE”), what is the theory against which the experiment functions. Also, we can separate the replies more clearly into no more than five types, according to the step they do not allow. Thirdly, we can identify fallacies and antifallacies (seemingly good steps that are bad, and seemingly bad steps that are good, respectively).
Little is said about these fallacies. Damper borrows from Sorensen the idea that a common fallacy (common to thought experiments only?) is missuposition, which “comes in two flavours”: oversupposing and undersupposing. The first case is that of admitting to much, creating flamboyant worlds and scenarios the possibility of which becomes suspect. The second – a fault which Damper finds in Searle’s CRA – is that of keeping the background assumptions to an ambiguous minimum. “Searle occasionally remarks on the simplicity and ‘conciseness’ of the CRA.13 Yet this very conciseness is in my view no more than a symptom of undersupposing.” (p. 177)
Although it might be true that Searle leaves out parts which might turn out to be important – in some respects – I think this is a very very weak refutation of Searle’s endeavour. It is hard to see what, of the relevant parts, is left out, and it is hard to believe so many people debated to and fro about this thought experiment without requiring an answer to what is in fact an essential question. Damper exemplifies, but unconvincingly: “how it is able to answer context-dependent questions (like ‘what was the question that I asked just before the last one?’)”. Plus, the fact that some important part is missing – and the recognition of any of the fallacies mentioned above, actually – is in no need of the formalization. The link between the (quasi-formal) translation and the replies is clear, but the link between the formal translation and the fallacies is practically absent. One could have accused Searle of not being specific enough, or being too exotic in his assumptions without having formalized his thought experiment beforehand.
PS: It is my impression that Damper assumes – without necessarily making it explicit – two strong propositions as being true: (1) that Searle’s thought experiment is an argument. One argument, that is. (2) that modal logic would tell us anything (new) about the TE. After all, how do we know that this or that formalization is a good one, unless we already have, and use, some criteria that are already there.
0 comentarii:
Post a Comment