3. Own Reflective Musings

In conclusion either side fails to convince me consistently. While I can hardly disagree with most of Searle’s responses to the critiques against his Chinese Room, I do not agree with the premises of his argument. Neither do I believe that his Chinese Room is a valid analogy to what is actually going on to begin with. For instance the assumption that a computer programmed with any scale of decision-tree could ever pass The Turing Test with syntax alone seems to me a bit far-fetched at best. For while I grant that there is a finite number of Chinese Symbols (approximately 5,000) the combinations of these symbols are infinite, at least as long as Searle does not define an upper limit to the length of sequence allowed. And for further consideration each of these symbols – even each of almost any given symbol sequence – might carry multiple meanings, and therefore also multiple responses, in accordance to the entire conversational context they are put in. There is not even any reason to exclude the context of the current cultural paradigm and recent events. For that matter each question would have almost limitless “correct” responses, and individual human beings seldom give the same answer to the same question twice. On these grounds I reject the notion that anything could pass The Turing Test without inherent semantics.

I will readily agree that syntax does not necessarily entail semantics. However when it comes to a sensory perception of the world based on ones and zeroes, or even the presence or absence of neural firings (these are analogous to a degree Searle would scarcely admit), I fail to see how semantics can be anything but a specialised form of syntax. Perhaps in order to make myself clearer it is in order to distinguish between purely grammatical syntax and the kind of synergetic or associative syntax I am referring to, which I consider applicable to this argument. After all the meaning of the word “apple” comes only from the association between the neural sequence the word produces and the sequence the image, the taste, the smell etc. an actual apple induces and our ability to recognise these separate experiences as “the same.” Searle’s insistence that The Demon is a separate entity from its “host” and that we should presume what we are trying to prove, because computers do not have these unexplained causal powers of intentionality, is not sound argumentation.

On the other hand while Searle might aim too low, Dennett and Hofstadter, aim that much more too high (although granted I am inclined to lean more in that direction). In particular their reasoning as to why just any “hunk of junk” could not be a mind seems a bit void to me. Is the distinction between a static describable mapping and a sophisticated representational system truly so clear-cut? Are there no gradual intermediary states? Is there such a thing as a semi-mind? And even though it might not be considered relevant to the problem of strong AI (although I could argue that it is) Dennett and Hofstadter’s writings left me wondering where exactly these two stand on the issue of determinism. Furthermore I also wonder whether sociological and cultural networks play no role at all in their scrutiny of neural networks.

So since I find both theories to be somewhat lacking, I have composed questions to both sides that I think need to be answered in order for them to be consistent to a satisfactory degree.

Questions for Searle:

• What exactly is the cause of the causal powers of intentionality to which you keep referring, and how much so must a neural network be sufficiently like our own in order to obtain these powers?
• When the word “symbols” can clearly be representative of anything, which is capable of carrying information (zero/one – blue/red quarks?) how can semantics be anything but a specialised form of associative syntax?
• Since by own admission ‘the [human] brain is a digital computer’[1] and ‘it follows from a theorem proven in 1936 by Alan Turing that any general-purpose digital computer can take on the guise of any other general-purpose digital computer.’[2] why should the same not apply to human beings – consciousness, causal powers, intentionality, and all?
• How would a perfect simulation of an abstract conceptual function (i.e. consciousness) differ from the “real” thing?[3]

Questions for Dennett & Hofstadter:

• How could a consciousness exist in complete isolation?[4]
• If the instantiation, which is the foundation of what we call consciousness is a deterministic process - as it seems to be in your theory - can this “biproduct” of a completely logical function be said to be a mind at all, rather than just a mirage projected when a process set in motion takes its consequential course to a set destination?[5]
• If so could we attribute consciousness to anything at all - human beings included - and is this not the natural repercussions of your theory?
• Where do you draw the line? What differentiates a static mapping from a self-updating system? Just how static does a mind have to be before it ceases being a mind?
• Would there ever be a point where a computational sequence would simply be too slow to be considered the instantiation of a mind?
• Have you considered that Searle might be right on another level than you?