One response to Daniel Dennett's confident statements about the mind being instantiatable by computers is to propose that a Rube Goldberg device could produce the same sort of results as a computer, given enough time and a little luck (i.e., no breakdowns in the components). Actually, I think something of this sort has been proposed already, and DD himself replied that these devices operate too slowly. Unfortunately, I didn't read DD himself saying this but heard someone else referring to him. In any case, DD's reply would be apt, at least initially, for such a device would be so sluggish that it would fail to convince a human observer that it was driven by thought: a slow motion computer will fail the Turing Test.
Suppose, however, that we give free reign to thought experiments: the following scenario might problematize DD's reply. Suppose, that is, that a device with the same structure as the Rube Goldberg device were somehow miniaturized so that it worked as quickly as a computer and was just as capable (or incapable) of passing the Turing Test as a computer made of normal components: if (per impossibile) the latter sort of computer could pass the Turing Test and in so doing show that it is capable of thought, then the former could as well. In such a case, would we really have a good reason to object to the attribution of consciousness to the a Rube Goldberg computer that has the same structre but which works at a normal speed?
It doesn't seem so; hence DD's reply about such devices being too slow doesn't seem plausible.
Another insight or intuition, however, might lie behind the response that the device is too slow. That intuition would be that cognition consists of a series of judgments, and that each of these judgments involves component acts through which an aspect or part of the whole thing to be judged is apprehended. Judgment can only be about what one already apprehends. The two must be simultaneous. And simultaneity cannot occur in a Rube Goldberg device or in galaxy-wide contraptions.
Suppose, however, that we give free reign to thought experiments: the following scenario might problematize DD's reply. Suppose, that is, that a device with the same structure as the Rube Goldberg device were somehow miniaturized so that it worked as quickly as a computer and was just as capable (or incapable) of passing the Turing Test as a computer made of normal components: if (per impossibile) the latter sort of computer could pass the Turing Test and in so doing show that it is capable of thought, then the former could as well. In such a case, would we really have a good reason to object to the attribution of consciousness to the a Rube Goldberg computer that has the same structre but which works at a normal speed?
It doesn't seem so; hence DD's reply about such devices being too slow doesn't seem plausible.
Another insight or intuition, however, might lie behind the response that the device is too slow. That intuition would be that cognition consists of a series of judgments, and that each of these judgments involves component acts through which an aspect or part of the whole thing to be judged is apprehended. Judgment can only be about what one already apprehends. The two must be simultaneous. And simultaneity cannot occur in a Rube Goldberg device or in galaxy-wide contraptions.
Comments