|
Author
|
Topic: Tinkering Design
|
Noel Rude
Member
Member # 516
|
posted 16. December 2003 16:01
Interesting how we argue from a priori assumptions – maybe here we should chalk one up for Gedanken who says, “If thought is physical, then it must base all new output on previous state and input, and there really is no evidence to suggest that thought is not itself a physical process.” But this is the question (hopefully not the premise): Is the mind entirely physical (in the sense of classical physics)? Some assume it is, others that it isn’t.
Obviously the mind has trouble thinking without the brain – just ask a stroke sufferer! But if there truly is no evidence to suggest that brain isn’t all there is to thinking – let it also be said that strong AI has had trouble coming up with a coherent theory of consciousness (which surely is an integral part of our thinking). Each strong AI hypothesis boils down to complexity – just make the machine complex enough and it will be conscious – or stimulus-response (perception, learning, whatever they call it) – just incorporate enough stimulus-response into the machine and it will be conscious. No, I would say “evidence” is pretty scarce on both sides of this equation.
But as to tinkering – in defining coherence linguists talk about something very much like tinkering. To be coherent (which I’m usually not), each clause in a discourse must carry over something old and introduce something new. Without something old we are incoherent (just try stringing together some sentences which share absolutely nothing). But with only something old we get tautologies. As discourse advances – and holds the hearer’s (and the speaker’s, for that matter) attention – each foregrounded (another technical term) clause should contribute something new.
But is there really anything new under the sun? That’s the age-old question. As it concerns information – at least from the linguist’s perspective – we wonder not just about a universal logic but also about universal categories of meaning. Didn’t Leibnitz spend a lot of time on this question? It is a tricky one. The trend in the 1960s (and as developed in the semantic theories of Ray Jackendoff) was toward semantic universals. The multicultural trend today is au contraire. My suspicion is that the same categories that are needed in universal grammar (also under attack today) are also those that are fundamental in lexical meaning. Nevertheless – even if there is a finite set of universal categories that underlie all information systems (which I suspect is the case, just as it is probably the case in the material world) – novelty does accrue at higher levels. And when it comes to Natural Language – tinkering is the only way to get there.
But the novelty – ideas come from somewhere – true. But oh what a mystery! Perhaps a universal set of meaning components (features) is a much simpler issue than the source of those sparks of novelty.
[ 16. December 2003, 16:21: Message edited by: Noel Rude ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 16. December 2003 21:16
Steve claims that a deterministic supercomputer model of QM isn't a good model of QM because it's deterministic. I disagree, because the observers in the simulation do not have access to the random number generator algorithm and seed, only its output. There is no experiment we could do to distinguish the two. So my point stands: QM is just as bad, in practice, for causal efficacy as determinism. (Note for those who know something about detecting QM vs. determinism: random number generators typically violate locality.)
Noel seems to be defending a "mental vitalism" position, to which I have two responses. The first is that ordinary vitalism is no longer seriously considered since purely quantitative
and organizational differences seem to account well for the difference between living and non-living entities--a fundamental qualitative difference is unnecessary. It is true that consciousness is confusing to us now, but the nature of life was at least as confusing to us not terribly long ago, so one can infer (from this and many other instances of successful (semi)-reductionist explation) that it is safer to bet that there is no non-physical basis of consciousness.
The second response is that the failures of strong AI aren't really informative. We know that if consciousness is a property of physical systems then it is a property exhibited by certain immensely complex physical systems (simple ones might be possible, but we have no examples). When dealing with emergent properties, it can be difficult to know a priori what level of scale is necessary for emergence. Nonetheless, we have some other examples of information processing capability: we do not conclude that computers are nonphysical because a fan, three transistors, and two LEDs don't generate a word processor. We're still grappling with how sensory perception is input to the brain--it is not anything like a solved problem even though a lot is now known--and higher processing has barely been started on yet (since we lack the tools to perform most meaningful experiments).
Evidence is far from conclusive, but this doesn't mean that evidence is that scarce. If one has other reasons to believe (or want to believe) that mind and consciousness is inherently nonphysical--because of the causal efficacy problem being discussed, for instance--then certainly there isn't enough evidence in favor of a physical model to show that one's beliefs are surely false. However, if one doesn't really care, the empirical evidence we do have points in favor of a physical model (because of things like strokes impacting consciousness), and points to the lack of anything sophisticated enough to count as evidence of absence via absence of evidence (because of the disparity in scale and complexity between systems with consciousness, and systems we have designed).
Noel's tie-in to linguistics and tinkering is intriguing. I have a question about the mystery of the novelty of ideas, though. Tinkerers and linguists have to come up with new designs and new phrases out of somewhere. What if one simply does this at random? Not at the lowest level of abstraction ("let's add a lot of steel"), but at the highest level of abstraction that is well-understood. For instance, we understand how to build airplanes, and we understand how to build lasers, and we have a finite number of places we can attach lasers and airplanes at that level of abstraction. Why not generate novelty simply by trying them at random, or trying every single one? I will grant that this is much slower than the way that we seem to actually do it much of the time, but wouldn't it work as long as we could evaluate the result? In other words, isn't this a way to remove the mystery of inspiration and shift the burden to comprehension/analysis?
[ 16. December 2003, 21:22: Message edited by: Rex Kerr ]
IP: Logged
|
|
Mark Szlazak
Member
Member # 391
|
posted 16. December 2003 23:30
Henry Stapp responds to Rex Kerr.
-----------------------------------
Rex wrote: quote:
But I am having trouble understanding the justification for merely incorporating QM to solve thorny issues of human agency. The reason is that I can model any QM effect I like, in principle, using a sufficiently powerful supercomputer with a sufficiently good pseudo-random number generator. This supercomputer would be entirely deterministic yet could replicate real QM to any desired level of accuracy including observer effects (so long as observer-classification routines were built in). Does this not also mean that "the foundation of our lives, the causal efficacy of our conscious efforts, is an illusion"?
So it seems as though we either need more or less. If we grant that we have causal efficacy and deny that such is possible in a deterministic setting, then we need a more radical departure from traditional scientific thought. Alternatively, if we come up with some way to explain causal efficacy in the QM case, it seems as though it would also hold in the deterministic case.
The issue is not determinism vs nondeterminism but classical-physics-based vs orthodox-quantum-physics based explanation of the detailed structure of behavior in actual experiments, as well as the specific form of the relationship between the psychologically and physically described data described in SSB.
Certainly a huge variety of behaviors can be described "deterministically." But one question is whether our actual behaviors can be accommodated within the time and space requirements, and physiological realities, of human beings. We are not talking about "in principle", but about a physical model of human beings. Quantum theory, by accommodating huge numbers of parallel possibilities, and a dynamics that involves nonlocal projection operators P, and dynamically important "choices" that specify classically describable probings of the environment of the agent-system, can do much more with a limited resource than the associated classical system. The QZE effect is something that cannot be achieved with a classical physics analog of a simple quantum system that supports it.
Another question is the connection between our consciously felt efforts and their apparent effects. A general "deterministic system" can do practically any conceivable thing, or at least an awful lot. But our efforts cannot do everything. Quantum theory, with its well defined Process 1, can do some things, but not necessarily anything. We want a theory that has a natural place for conscious effort: a natural dynamically specified place that makes it conform to the actually existing capacities and limitations of our conscious efforts.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 17. December 2003 06:11
We don't have a (failed, classical) physical model of human beings, so introducing QM is creating a solution in seach of a problem. There are abundant hints that neuronal actions produce "mind", and there are scarcely any neuronal properties that could depend on QM effects without being too unreliable to be useful.
Also, determinism can't do anything, and it can create a connection between consciously felt efforts and their apparent effects, since those neuronal processes that lead to a feeling of effort can also lead to, for example, motor cortex and thus cause (large-scale) physical movement. However, neuronal processes cannot cause lead to turn into gold.
So I still don't see a justification for incorporating QM. However, I also don't see much of a connection to tinkering, so perhaps I'll stop here before we get even more off topic.
IP: Logged
|
|
nobody
Member
Member # 145
|
posted 18. December 2003 13:17
Rex says:
quote:
non-human design doesn't imply a complete lack of tinkering!
That's true, of course. However I am very doubtful that God ever needs to do any tinkering. The incredible programming of life was designed to compensate for many variations in environment. For example: the incorrectly named "Darwin's" finches. They're actually God's finches.
P.S. The way you word your post leaves you open to the alleged "alien" tinkerers. I think the Raelians beat you to that position though....
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 19. December 2003 12:35
When considering possible methods of design, I am open in theory to all conceivable designers--little green men, God, humans, six foot tall invisible rabbits, Darwinian evolution, etc.. I may think that some are less likely than others.
Whether God (or the others) need to tinker to design is a somewhat separate question from whether they would tinker to design a specific item, which is again somewhat separate from the question of whether they actually designed a specific item.
The "incredible programming of life" contains the ability to self-tinker, by having variability and letting those ill-adapted creatures die. If we explain details of the design of life by appeal to God, it seems just as problematic to me that God's non-tinkered design allows such a prima facie inhumane method of adaptation to the environment, as to say that God tinkered changes directly. If the God answer is the right one, then even though we may admit he doesn't need to tinker, it's probably unwise to assume that he never did/does.
IP: Logged
|
|
Steve Petermann
Member
Member # 884
|
posted 19. December 2003 13:19
I can see how the question of tinkering is a theological issue, but I'm curious how it relates to the scientific question of intelligent design. Does a tinkering methodology inform the mathematics in some fashion? Clearly natural mutations are an evolutionary form of tinkering so how would positing a tinkering designer faciliate resolution of the issue? Seems to me that the more one posits a design methodology that mirrors what science sees happening in nature, the less helpful it would be. Instead it would appear that a more fruitful path would be to ferret out design methodologies that science doesn't see in nature. Instead of tinkering which might presuppose the inability to plan ahead, a good intelligent designer would use methodologies that utilize the predictive powers of intelligence. If they haven't already been done, it looks like to me that computer models of design methodologies with varying levels of prediction could give some sense of what could be designed and how fast.
IP: Logged
|
|
David Bump
Member
Member # 1017
|
posted 20. December 2003 00:32
The problem here seems to be that by addressing "tinkering design," we're just adding a new level of complexity to an already complex subject, and it may end up being a moot point.
I see references to "design" by Darwinian evolution (BTW, nobody believes in truly Darwinian evolution anymore, right?) and "the history of life" supposedly "showing" all sorts of tinkering going on, and now that mutations are a form of tinkering. These are all arguable points, at best. Or, (as Steve seems to be saying) if you state them in a certain way, they pretty much define away any need of, or indication of, an intelligent designer -- so why bother?
In fact, some of these discussions seem to be heading in the direction of calling into question the usefulness of terms such as "intelligence" and "design" themselves.
When raindrops fall on my car and leave circles on my car as they dry, are they desigining? Or is it the rain cloud? Ultimately it's part of God's design, but at a level that's not available to human attempts to recognize it.
When a truck ahead of me kicks up a pebble and it cracks my car's windshield, is that "tinkering" with my car's design? Would any amount of that add to the complexity of my car?
These are just analogies, but if we could focus on just exactly what we do observe and really know by hard data and experimentation, and specific examples of indubitable tinkering, we might find some use for this line of investigation.
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 20. December 2003 18:20
David Bump: BTW, nobody believes in truly Darwinian evolution anymore, right?
What do you consider to be truly Darwinian evolution?
My perspective is that selection remains a very important contributor to evolutionary theory but that neutral mutations are very important since they provide for evolvability and robustness. As far as the sources of variation, Darwin seemed mostly mute on this since he was not aware of the gene/DNA. I see various interesting sources for variation.
So while Darwinian evolution remains an important theory, much has been learned about the actual details which paint a fascinating picture.
As far as mutations being tinkerers, I believe that the argument is that variation and selection can give the appearance of tinkering.
Do they explain away any need for or indication of an intelligent designer? That would only be when such designer cannot exist in the presence of tinkering nature.
[ 20. December 2003, 18:22: Message edited by: Pim van Meurs ]
IP: Logged
|
|
David Bump
Member
Member # 1017
|
posted 21. December 2003 00:49
What I mean by "truly Darwinian" is just that -- what Darwin believed about evolution. Darwin didn't even understand some important basic facts about heredity, let alone genes and DNA. Not that it was beyond finding out in his time, as Mendel demonstrated. Darwin's theory depended on the idea that heredity was analog, not digital.
No scientist now could believe as Darwin did, so at best, if we're going to specify the form of evolution being considered, it would be Neo-Darwinian, or perhaps some other variation.
Of course, while there are various forms of variations, unless we're going to consider that God directly (miraculously) tinkered with genomes at certain times, the only mechanism to accomplish the posited evolution of highly specialized colonies of eukaryotic cells with degenerate endosymbiotes (aka humans) is mutation.
If we define "tinkering" to include any variation in anything (a simple design or even a geometrical pattern), then any variation from any source will have the appearance of tinkering. As the seminal ID papers noted, intelligent designers can produce works that appear chaotic, and natural forces can reproduce patterns like those produced by some intelligently-guided actions. The whole point of ID is to recognize those phenomena which stand apart from the overlapping areas.
Thus, as I said earlier, if this thread is to prove possibly useful, we need to look at the "high end" definition of tinkering and see if, at least there, we can find a demarcation between what intelligent tinkerers do, and what occurs when no known IT is involved.
This would seem to suggest an IC/SC definition: An IT may be invoked if a system is modified in several specified ways to a more complex, functional form (which agruably could not be replicated by slight, unguided changes with each stage being at least equally functional).
A designer could exist in the presence of another designer. Even if nature designed itself without any input from a designer, the designer *could* exist, but as far as I can tell, there's no reason to believe in such a designer.
However, as Charles Babbage pointed out, a system could be designed that would seem to run by certain mathematical "natural laws" (or calculations, series of numbers fed into a formula, etc.) and yet after any given number of examinations of the output, produce something unexpected, but not to the designer (or anyone able to thoroughly study the mechanisms behind the output.)
IF there is an appearance of tinkering in nature, then:
1) Which of these cases (if any -- or is it all?) can be determined are well within the range of purely natural forces?
2) Which of the cases may suggest an IT? If any, does the tinkering appear to require a more or less direct intervention?
And of course, we have to look at each case that appears to be worth further pursuit to see the basis for the claim of appearance of tinkering, to save us chasing wild geese, as indicated in an earlier post.
[ 21. December 2003, 00:57: Message edited by: David Bump ]
IP: Logged
|
|
|
Printer-friendly view of this topic