... a robot need only perform onboard perception and cognition where speed of response requires it – roughly the equivalent of instinctive functions in a biological organism. (FG)
Dr C is, unlike me, a "real" doctor - a medical one. Not only does this make him more useful to the world (forced to keep only one, which wold you keep: and MD or a mathematician?), it also makes him more knowledgeable than I about the human animal. Where I happily made sloppy use of the word "instinctive" in the above quotation, he quite rightly applies it with more care:
It is my opinion, and I plan to spend a lot more time on this, that all functions of a biological organism are "instinctive." Actually, I would put the word "reflexive" there. (DrC)
After an excursion over the next few paragraphs into the differential placement of information and storage and processing (which you can skip by clicking here), I'll come back to that.
The article is one which I wrote last year (one of a linked pair, this one being in Scientific Computing World, the other in Imaging and Machine Vision Europe) on pattern recognition and robotics. It makes brief reference to archaeological use but the relevant part is one in which I discuss, in Dr C's words, "separating out the physical locale of knowledge". I wouldn't have used the word "knowledge" myself, but put that aside for now - the principle is easily illustrated is human terms.
Incoming sensory information is processed in different ways, and at different levels. Response to that processing also occurs at different levels according to urgency.
For instance, if I see a thorn bush ahead of me as I wander along a rural footpath I will think about it; I will not immediately act on the knowledge of its existence. Is any action required of me, I will ask myself? Does the path the path bypass the bush? If not, does the bush significantly block my way? What do my past experience and accumulated knowledge of thorn bushes have say about this particular situation? And so on. I keep walking, assimilating further information from my eyes as my viewpoint changes, making leisurely analyses and, eventually, taking some sort of action such as stopping or altering direction or just walking on. All of this fairly sophisticated analysis is going on primarily within my brain.
If, on the other hand, a thorn from a previously unnoticed bush punctures the skin of my leg, the sequence of events is much shorter, quicker and cruder. A signal about the pain is sent to my brain, but there is no waiting about for a mentational response before acting. Somewhere in my spine, a local executive decision is made to withdraw from the source of pain and, secondarily, to bring wider perception to bear on its source. Even before that, reflex arcs lower down the chain in the leg itself have initiated local motor responses. My motor system responds with automatic implementation of these decisions: weight transfers to the other, unpunctured leg; the affected leg, thus temporarilyfreed of support requirement, is withdrawn; my whole body rebalances in preparation for generalised movement if required, and swivels to face the source of pain. Other, nonmotor responses are also swinging into action: my senses are heightened in anticipation of further threat; chemicals are released into my bloodstream, just in case fight or flight are called for in response to this breach of bodily integrity. And all of this happens before my brain is even aware of the pain.
For human beings, unlike other animals, it also works in the opposite. For anything which can wait still longer than the time taken to think about it (perhaps whether the fruit of this thorn bush will, later in the year, be edible) I can call upon knowledge stored, and information processing executed, elsewhere - not within myself at all but, for example, in books or on the web.
I've oversimplified of course. In real life, there is a whole continuum of response levels. Recognising a caller at the door requires less speed of response than a pain in my leg, but more rapidity than the decision whether or not to bother walking around a distant obstruction. But those three levels will do for now.
Human beings are products of evolution (well, I believe so anyway - but even if you hold that we were created at a stroke, my point holds) and are, for now at least, to a considerable extent constrained when deciding where to locate information storage and/or processing. Survival through all those millennia of pretechnological existence made us what we are: a fixed package of onboard facilities with expansion taken care of by social groupings. In the future may this may not be true, but for now remains so. The machines we build have no such historic constraints; we can choose to place the facilities wherever the resulting time delay doesn't interfere with function.
Suppose I took on a humanoid robot such as Honda's Asimo as a personal house servant. Asimo has the same need for balance during locomotion as a human being, and that requires a speed of feedback between sensor and motor systems which is only available if placed within Asimo's physical shell. The layout of my house, however, is not required onboard even if Asimo is to be my servant - it can be stored in capacity within the house itself. My culinary likes and dislikes could be stored on the moon without impairing Asimo's ability to cook for me.
It's in that context that Dr C quotes from my article as noted above:
It is my opinion, and I plan to spend a lot more time on this, that all functions of a biological organism are "instinctive." Actually, I would put the word "reflexive" there. (DrC)
I willingly acknowledge that "instinctive functions" was too imprecise - I was making a fuzzy reference to the submentational level of reflex responses (exemplified above by reaction to a thorn prick in the leg) in a higher animal with a brain.
I also freely acknowledge that very many functions of a biological organism (or a synthetic one, for that matter) are instinctive and reflexive even when we like to think (or habitually feel) that they are not. On whether that can be extended to all functions I am more cautious. First of all, I'm not sure that it is true - though I am not saying it is untrue; I just don't know; I await with real interest and anticipation the more time which Dr C plan's to spend on the subject. Secondly, I'm not sure that it will ever be possible to ultimately decide the question one way or t'other - the ability of a system to fully comprehend itself being a likely barrier. Finally, I'm not sure where it leaves us on the sort of moral issues over which Dr C and I frequently tend to exercise ourselves.
The first and second issues can wait until I have read what Dr C has still to say - I shall certainly learn from his depth of knowledge, even if areas of disagreement or uncertainty remain, and I'm only going to make a fool of myself if I speak too soon. On the third, however, I can at least make a general start here and now.
Dr C continues from the para I've quoted above to say:
Furthermore, I contend that this includes all of the so called "higher" functions of humans up to and including Free Will. Take that you intelligent designers.
Now ... while Dr C and I have many differences at a detailed level, we share a sense of outrage at a range of human actions which result in suffering to others. We also share an expressed feeling that such actions should cease. Note that I say should. That word "should" implies moral decision. From a purely functional point of view there is no reason whatsoever why I can't murder, torture, disable, rape, or do anything else to, my neighbour if it benefits me in some material way. To (as both Dr C and I do) say I shouldn't do so, when I can and when it produces no material disadvantage to me or mine, is to introduce a moral dimension which has no supporting rationale beyond its own terms of reference. I'm not going to say that this cannot be explained in terms of reflexive functions, but I certainly find it hard to see the explanation unaided.
It seems to me, admittedly on purely intuitive bases, that the mentation upon which the concept "should" is erected has evolved beyond reflexive constraint. Not, I hasten to add, uninfluenced by reflexive processes; and built from older reflexive components; but not, in its total gestalt nature, still purely reflexive nor purely constrained by reflex.
Whatever the truth of this one, I'm not sure that it really touches the issue of Intelligent Design. Either we believe (as I do, alongside Dr C) that is implausible and the chreostochastic processes of evolution are a better bet, or we don't. Even if it can be definitively shown that free will can arise from reflex, that doesn't prove that it did - the choice of which to believe will still be there and the ID brigade are as unlikely to change their belief as I am.
I'll look forward eagerly not only to more from Dr C in this direction, but also to more (mentioned in the final parenthesis of the post) on the physical locale of knowledge in relation to medicine.
No comments:
Post a Comment