You know, strictly speaking, that's *incorrect*.
The thing that is required for a genetic algorithm and/or evolutionary approach to work is:
A well defined "optimization" -- some choices must be clearly better than others.
Fairly stable conditions compared to the time over which adaptations would occur.
A large number of trials.
For the moment, treat it purely mathematically. Let us say you are trying to map out the lowest location on a golf course.
The "God's eye" view would *know* the answer, intuitively.
A person might know from studying the landscaping diagrams.
Another enterprising person might remember, "Oh, that's right. Golfballs roll downhill." And then they'd walk to some spot on the course, drop the ball, and see where it rolls.
They'd find the lowest spot, of course. But only the lowest spot -- near them -- that *didn't* have a hill in the way.
To accelerate the process, you could do one of two things.
Intelligent selection: Look at maps of the golf course to get the lay of the land, and pick likely spots to drop the golf balls at.
Evolutionary selection: start with a bunch of golf balls in a certain area of the course and drop them. For each time you drop a set of balls, keep track of the best results from the set. Slightly change the locations of the best results in the next trial. Apply, lather, rinse, repeat.
IF the topographic features of the golf course fulfill certain conditions: no sudden small holes surrounded by hills, fairly flat, etc., then the evolutionary approach will work.
A "simulated evolutionary" approach will not start in one area of the golf course, but instead will "rain golf balls" all *over* the place. In this way it can sample areas which are inaccessible to small random changes to prior steps.
The murky parts of current evolutionary theory (to this abject layperson) have to do with how quickly environmental changes happen in comparsion to the time for new generations to produce mutations; the "intrinsic" rate of mutations, and its influence on the conservation of some features(*), and the fact that some mutations appear to happen much more quickly than others; why certain features appear to have developed independently, but others which would *seem* to be useful never got the chance; and the crucical sizes of populations to keep going, whether genetic diversity, or simply overwhelmed by ordinary disease or predation.
One of the problems is that evolution seems to be the intersection of biology (a science) and history (a discipline) -- just as we can't go back and re-create the Battle of Waterloo, we can't go back to when the first fish developed lungs and see what it specifically was about the environment that so favored lungs *then*, but not other times...
Anyhow, the upshot is that if you take enough random shots, and throw away the failures (like Hillary's Cattle Futures), the end results can look darn good -- just *as if* things had been planned.
(...as God says, "NO! Get your *own* divot!")
Cheers! (*) why the hell doesn't the coccyx just *go away*? Ditto for nearsightedness.
[[Evolutionary selection: start with a bunch of golf balls in a certain area of the course and drop them. For each time you drop a set of balls, keep track of the best results from the set.]]
What are you ‘keeping track of’ the results with? and how do you know the results are goign to work with hte whole system?
Did you read the thread on ‘life’s irreducible structures’? Where did the metainfo come from that organizes all this? Simply changign somethign doesn’t cause metainfo to arise- nor would htrowing a whole bunch of golfballs at the situation. You are implying htere is soem sort of intelligent agent at work selecting out the right ‘descisions’, and hoping further right descisions mesh with hte previous ones until some sort of change occures.
The thread I spoke about asked for input as to why the 5 point heirarchal/metainfo hypothesis was wrong, but it got no responses, because in order for life ot exist, all 5 points, and hte metainfo that correlates everythign must be inplace first before any changes can take place- and nature ismply hasdn’t got an answer for the points or hte metainfo-
Mutations just point out how encessary this metainfo is quite frankly
The thread I mentioned, it was brought up whether piling info on info (via random mutaiton selection) could result in metainfo, and hte answer appears to be no for several reasons. The link I gave above, talks about htese hierarchal points by both ID science and Macroevolutionsits, and discusses whether or not metainfo needs to be inplace before any selection or changes could take place.
the bottom line to what you are suggesting is that throwing a bunch of mutaitons at the ‘species’ can result in macroevolution, and htis simply isn’t true- mutaitons can only change, they can’t add non species specific informaiton. The meta info I spoke of contains all the info needed for a species- it contains species specific info, and it allows change- metainfo is info about information, and as such contains everythign needed specific to that species, both present and future to account for hte changes due to mutaitons.
Right off hte bat you have hte problem of the metainfo. It is impossible to get this metainfo by evolving chemicals into species because chemical info is far below metainfo, and cell info is also far below the metainfo
Here’s the link for tomorrow: http://www.freerepublic.com/focus/f-news/2163122/posts?q=1&;page=551
“Autopoiesis provides a compelling case for intelligent design in three stages: (i) autopoiesis is universal in all living things, which makes it a pre-requisite for life, not an end product of natural selection; (ii) the inversely-causal, information-driven, structured hierarchy of autopoiesis is not reducible to the laws of physics and chemistry; and (iii) there is an unbridgeable abyss between the dirty, mass-action chemistry of the natural environmental and the perfectly-pure, single-molecule precision of biochemistry. Naturalistic objections to these propositions are considered in Part II of this article.”
Andfrom Betty Boops excellent analysis which you’ll find abotu 1/2 way down the page, where hte discussion startsto really take off discussing whether naturei s capable of creating htese hierarchy’s or not:
“First a recap of the AP hierarchy: (i) components with perfectly pure composition (i.e., pure elements); (ii) components with highly specific structure (e.g., molecules); (iii) components that are functionally integrated (i.e., can work cooperatively toward achieving a purpose or goal); (iv) comprehensively regulated information-driven processes (DNA, RNA); (v) inversely-causal meta-informational strategies for individual and species survival (we’ll get to that in a minute).
But what does this hierarchy mean?
An interesting way to look at the problem, it seems to me, is to look at the available potential “information content” of each of the five “manifolds” or “dimensions” of the hierarchy.
(For the present purpose, we’ll assume that “algorithmic complexity” is a function that yields “information content.”)
(i) gives us the atomic numbers of all the elements: their natural fingerprints or IDs by which they can be identified. It also gives us all the laws of physics. (i) is the physical foundation of all life forms. [As you may recall, Chaitin estimated the algorithmic complexity of the physical laws as ~103 bits.)
(ii) gives us the elementary bonding laws; indeed, it gives us all of physical chemistry. It is based on (i), but it is not reducible to (i). Indeed, irreducibility is a property of each of the hierarchical dimensions (i)–(v) with respect to the one(s) prior to it in the hierarchy. [I’m not aware that anyone has ever tried to calculate the algorithmic complexity of the chemical laws. But it must be greater than 103 bits; for chemistry rests on physics, plus its own unique “value-added” which level (i) does not anticipate.]
(iii) is where things really get interesting. This signals the first evidence that a life form actually exists. It should be obvious that such a phenomenon cannot be explained on the basis of the information available at levels (i) and (ii). It’s as if (i) and (ii) “didn’t even see it coming,” though both were necessary to (iii).
[[why certain features appear to have developed independently, but others which would *seem* to be useful never got the chance;]]
All these myriad mutations happen fully within species specific parameters, and can be demonstrated scientifically- which also points very heavily toward the idea that there is a system of metainfo at work, controlling, directing, allowing, and not allowing etc. But again, these all work within species specific parameters that is already designed and built into the whole species- a system that ‘forsees’ problems, plans for htem, allows them, and works through them to try to help preserve the species. And again, these allowances are species specific- think of breeders who are bound by species limitations, and can only go just so far when messing with species trying to modify them.
Anyways, this system of metainfo is present right from the simpelst cells all the way to fully functional creatures, and idneed must be present or hte species, whether just a ‘simple’ cell, or a complex species, would perish before it could even get started.
The htread on ‘life’s irreducible structures’ is quite fascinating, and it presents a serious challenge to darwinian macroevolution, while fully accountign for and allowing for microevolutionary adaption and change within species kinds.
Macroevolution demands a ‘bottom up’ cretion of metainfo, which is impossible (somethign we discussed in the article), and it’s especially impossible if we’re to concider life coming from ‘dirty chemicals’ (Life is created from pure chemicals and cells)
Anyway- there it is. Thel ink I gave to Discovery.org is on a paper talking abotu this concept a bit further, which I posted exeprts to- but hte whole paper is worth a read as it layws out why mutaitons can’t account for hte complexities in heiararchy that we know to exist in species.