John R. Searle (2002) 'Twenty-One Years in the Chinese Room', in John  Preston and Mark Bishop, eds. (2002) _Views into the Chinese Room: New Essays on Searle and Artificial Intelligence_, Oxford : Clarendon Press, pp. 51-69.

John R. Searle (1980) 'Minds, Brains, and Programs', in _Behavioral and Brain Sciences_, vol. 3, pp. 417-424. 

Diane Proudfoot (2002) 'Wittgenstein's Anticipation of the Chinese Room', in John  Preston and Mark Bishop, eds. (2002) _Views into the Chinese Room: New Essays on Searle and Artificial Intelligence_, Oxford : Clarendon Press, pp. 167-180.

Ludwig Wittgenstein (1953) _Philosophical Investigations_, trans. G.E.M. Anscombe, Oxford: Blackwell. 


That it is believed possible to make a machine pass the 'Turing Test' (in Turing's 1936 paper) - to respond to an input with an output similar to a human - does not mean that the computer and human are doing the same thing. John Searle's notorious 'Chinese Room argument' (in an essay 'Minds, Brains, and Programs', of 1980) refutes the 'test', by claiming that machines fall short in understanding the symbols they process. Expressed differently, the observation is that the syntactical, abstract or formal program of a computer program are not the same as semantic or mental content associated with the human mind. The cognitive processes of the human mind may be simulated but not duplicated as such. The thought-experiment proceeds, as follows:
'Suppose that I'm locked in a room and given a large batch of Chinese writing. Suppose furthermore (as is indeed the case) that I know no Chinese, either written or spoken [...] To me, Chinese is just so many meaningless squiggles.' (1980: 417) 

Given linguistic instruction and rules, Searle imagines that he is able to answer questions that are indistinguishable from those of native Chinese speakers. But:
'... I produce the answers by manipulating uninterpreted formal symbols. As far as the Chinese is concerned, I simply behave like a computer; I perform computational operations on formally specified elements. For the purposes of the Chinese, I am simply an instantiation of the computer program.' (1980: 418)  

Searle's position is based on the linguistic distinction between syntax and semantics as applied to the digital computer (Turing machine), as a 'symbol-manipulating device' where the units have no meaning in themselves. Even if it is argued that there is some sense of intentionality (or agency) in the program, it is not the same as that exhibited by humans (what Searle refers to as 'as-if intentionality'). Searle in the Chinese Room is an 'instantiation of a computer program' drawing on a database of symbols and arranging them according to program rules. But it can also be argued that the experiment is simply a description of the workings of the CPU (central processing unit) and not the computer system as a whole. This 'Systems Reply' to the Chinese Room argument claims 'that even though there is no semantic content in me alone, there s semantic content somewhere else - in the whole system of which I am part, or in some subsystem within me.' (2002: 53) All the same, the room itself or any other part of the wider system does not understand Chinese either. To think otherwise, is what Searle refers to as 'verificationist reductionist urge' like the Turing Test itself and 'Strong AI' in general (2002: 52). It would seem that no amount of simulation will duplicate the Chinese speaker. 

To Searle, the distinction between humans and machines is 'obsolete': humans are already biological machines. An artificial machine can in principle think, but only in simulation; for it to duplicate it would also need to replicate causal powers - produce consciousness itself (2002: 56). This has proved impossible thus far. The problem for Searle lies 'not that computational processes are too machine-like to be conscious, it is rather that they are too little machine-like' (2002: 57). By this, he means that computation relies on an abstract mathematical process and not in terms of energy transfer like some other machines; energy transfer is not sufficiently part of computation. The Chinese Room argument is useful not least in disputing of the truth claims of computer science. The appearance of understanding Chinese is an illusion. Indeed it is a fantasy that is produced through a number of misreadings of scientific progress and dead metaphors. For instance, increased computational capacity ('Moore's Law' in action) does not necessarily mean that producing a conscious computer somehow is closer (2002: 68).   

It remains a truism that a set of formal principles alone is not sufficient for the production of meaning. Computation cannot in itself be said to have access to or know or understand content. This is what Wittgenstein refers to as talking 'without thinking' (from 'Wittgenstein's Lectures on Philosophical Psychology 1946-47', 1989: 7). To Wittgenstein, the Chinese spoken by the 'living reading-machine' could not even be described as a genuine speech act, for 'a proposition isn't just a series of sounds, it is something more. [...] The sentence, as it were, plays a melody (the thought) on the instrument of the soul.' (in Proudfoot 2002: 168) 

The reading-machine is simply manipulating symbols, that Wittgenstein likens to the workings of a pianola - that translates marks into sounds, merely following patterns. The lack of understanding is apparent:
'The living reading machine produces as output solutions to arithmetical problems, texts spoken aloud, proofs of logical theorems, notes played on a piano, and suchlike. These 'machines' may be born, like an idiot savant, or trained.' (from _Philosophical Investigations_, 1953: 157; in Proudfoot 2002: 168-9)

In the Chinese Room, this is symbol-manipulation that is insufficient for understanding. Although there are similarities with Searle's position, Wittgenstein's arguments are rather different - for 'meaning-blindness is not a barrier to the use of language (Proudfoot 2002: 173); and furthermore, denying that understanding, thinking, intentionality, and meaning, are simply evident in the process of symbol manipulation (2002: 176; hence, to Diane Proudfoot, Wittgenstein prefigures a 'connectionist view of mind). To Wittgenstein, when we say a word we refer 'to the whole environment of the event of saying it. And this also applies to our saying that someone speaks like an automaton or parrot.' Understanding comes about through a particular history, and through participation in a particular social environment (Proudfoot 2002: 177-8).