Minds and Computers : An Introduction to the Philosophy of Artificial Intelligence
Download 1.05 Mb. Pdf ko'rish
|
document (2)
appearance of understanding is evidenced by the system – su
fficiently well to convince a human – there is something crucial lacking in the operations of the system. They don’t, in and of themselves, mean anything. In other words, the syntactic operations of the Chinese room, although they pass a Turing test, lack semantics. 17.2 SYNTAX AND SEMANTICS On the strength of the Chinese Room thought experiment, we might be tempted to mount this argument against computationalism: P1 Having semantics is a necessary condition for having a mind. P2 The syntactic operations of formal systems are not su fficient for having semantics. _____________________________________________________ The operations of formal systems are not sufficient for having a mind. _____________________________________________________ Computationalism is false. Premise 1 is not in dispute – it is clear that mental states have seman- tic content. Premise 2, however, is certainly arguable. There are two ways we might interpret the second premise, and, consequently, the interim conclusion that follows from it. The weaker interpretation is the claim that is licensed by the thought experiment; 177 however, a much stronger interpretation is appealed to in deriving the claimed falsity of computationalism. The weak interpretation of the second premise is that there is a formal system such that its operations are not su fficient for having semantics. The stronger interpretation is that there is no formal system whose operations are su fficient for having semantics. The Chinese room thought experiment does not show that there can be no formal system whose operations are su fficient for generat- ing semantics. Consequently, the argument above fails to show the falsity of computationalism. What the thought experiment does show, however, is something a little stronger than the weak interpret- ation of the second premise I’ve given above. The weak interpretation of premise 2 is essentially trivial. It’s a given that there are plenty of formal systems whose operations don’t meet conditions for having a mind. What is interesting about the Chinese room thought experiment, however, is that it shows that there is a formal system whose operations alone are su fficient for passing a Turing test, yet, intuitively, the system lacks understanding entirely. We might, then, interpret the Chinese room thought experiment as an indictment on the e fficacy of the Turing test. After all, if some- thing can pass the test despite a complete lack of understanding, it doesn’t seem the test is at all a reliable indicator of the presence of a mind. Before we draw this conclusion, however, we should reflect on the system described in the thought experiment in light of what we know about formal systems and natural language processing. The thought experiment describes a system which, while logically possible, is not physically possible. To implement this system, we would need to draw up the generation tree for all possible Chinese conversations that can be had in the course of several hours. Given that a generation tree for possible conversation states would be considerably more complex than the generation tree for possible chess states, it should be clear that constructing a complete genera- tion tree for even the first twenty possible conversational exchanges is simply not computationally tenable. It is safe to say that, regardless of future advances in practical computational power, no computer will ever be able to pass a Turing test by following the method which the Chinese room implements. Ordinarily, arguing against the physical possibility of a thought experimental situation obtaining does no philosophical work since, generally, we are using thought experiments to test claims of logical relations. 178 For instance, physicalism holds that a complete physical description is su fficient as a complete description of the mind. This sufficiency is a logical claim. Consequently, while it is physically impossible that there could be a scientist such as Mary, the thought experiment described in Chapter 5 still speaks against this su fficiency – to the extent that your intuitions are primed by the thought experiment – since the situation described is logically possible. If the claim concerning the Turing test was similarly a logical claim, then the Chinese room thought experiment would indeed speak against it. Recall, however, that the claim is not that passing a Turing test is su fficient for having a mind, but rather that, were something to pass a Turing test, we should be prepared to attribute mentality to it. The Turing test is an empirical test. Consequently, when consider- ing possible counter-examples to the e fficacy of the test as a reliable indicator of the presence of a mind, we should restrict our consider- ation to empirically possible systems. So the Chinese room thought experiment fails to straightforwardly show the falsity of computationalism and o ffers no indictment on the e fficacy of the Turing test. It does, however, still show something rather important. The thought experiment does, I think, show that no amount of syntactic operation in isolation from the external world is su fficient for generating semantics. I could be in the Chinese room performing this procedure for years – given enough books – and it seems, intui- tively, that there is no way to begin to understand the meaning of the symbols I am processing. The reason is that my operations lack an appropriate connection to the external world. To understand Chinese just is to understand how elements of the language – written or spoken – relate to things outside the system of language. Languages are systems which encode and communicate meanings. These meanings, however, are not generated by mecha- nisms and inputs entirely internal to the linguistic facility. Certainly linguistic mechanisms are implicated in the conferral of meaning to linguistic entities but necessarily implicated is an appropriate connec- tion to the external world. The lesson to draw from the Chinese room thought experiment is that embodied experience is necessary for the development of semantics. In order for our mental states to have meaning, we must have antecedent experience in the world, mediated by our sensory apparatus. In other words, semantics do not develop in isolation but, rather, this development is conditional on experience in relation to the empirical world. 179 This necessity of embodied experience for the development of semantics does not, in and of itself, speak against computationalism. It merely shapes the explanatory burden on the computationalist, requiring them to provide a computational account of the meaning conferring mechanisms. This will involve, inter alia, an account of the computational conversion of sense data to various kinds of Download 1.05 Mb. Do'stlaringiz bilan baham: |
Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling
ma'muriyatiga murojaat qiling