4 ms·
The Chinese Room Thinks
- simonh 5y agoThis is a decent refutation of Searle, mainly because Searle's characterisation of the Chinese Room is that it just encodes and processes syntax. Engaging in a conversation requires much more than syntax, it requires analysing and decomposing the meaning of what is being said, which for example could require constructing a model of a situation and reasoning about activities and changes in a situation. It might require critiquing suggested strategy for playing Chess for example, or commenting on economics. Actual conversation takes a lot more than just syntax, so by characterising it in this simplistic way he's stacking the deck in the reader's mind against the Chinese room. This result is incredibly impressive. There's a lot of analysis going on here, far beyond mere processing of syntax. In this sense it exposes the inadequacies of Searle's critique of AI. It won't kill the Chinese Room argument though, it's supporters will simply extend it beyond just syntax. They will still claim there is no thinking mind producing these results and they're probably right, after all as I understand it this system is essentially a hierarchy of nets that filer input to generate output, there's no persistent running cognitive process that does anything like thinking on an ongoing basis. The Chinese Room argument is still bunk though for many other reasons.