Home » Complex Specified Information, Mathematics, Psychology » Design recognition is possible in part because of finite human memory and limited human information

Design recognition is possible in part because of finite human memory and limited human information

Why is it that humans can recognize the designs of other humans even for token objects like a system of 500 fair coins? Why does life resemble designs? Answer: designs frequently conform to simple organizing principles rather than explicit patterns. Simple organizing principles are a way to understand large amounts of data with our finite human minds and limited information. Ironically, the fact that humans have finite memory and limited information is one reason humans tend to think and design in terms of organizing principles, and thus design creation and recognition is possible in part because of finite human memory and limited human information.

First, it would be helpful to compare and contrast design detection using organizing principles versus design detection using explicit pre-existing patterns. One clumsy way for a human to build an artifact which another human can recognize is to build artifacts that conform to pre-existing blueprints either in the human’s memory or stored on some medium available to humans. One such example of this method of design detection is done with Genetically Modified Organisms (GMOs). If we have a seed, and we don’t know if it is a GMO, we can determine if it is a GMO if the manufacturer provides blueprints to observers to help us identify the GMO. (See: Genetic-ID, comment for Alan Fox)

As another example, if I knew another human had memorized a 500-bit pattern like this :roll: :

01011 00010 11111 00111 01101
00001 10001 00111 01110 10011
01000 00101 11000 11110 10111
10100 00101 01010 00111 01011
01110 00100 00101 01111 11010
00010 01011 01000 01111 10010
00110 01100 00100 10100 01101
01000 00000 11111 11110 10111
10111 10001 01000 01100 01010
00000 01010 00001 11001 01110
10111 10101 00110 00101 01110
00101 00101 11001 00011 11111
01010 10001 11000 11000 11111
01010 00100 10101 00100 00100
11001 00101 10100 01110 10111
01110 10111 10111 11010 10000
11010 10110 01001 00100 11100
01001 10010 01111 01010 01110
00101 10001 11110 00010 10010
11011 00000 10110 01011 01100

I could put out that pattern on a table using uniquely numbered fair coins (where Heads =1, Tails =0), and he ought to recognize it as designed. As pointed out in the essay To recognize design is to recognize products of a like-minded process, Part I, the real probability in the question of design recognition isn’t the probability of a given sequence, but the probability that a sequence conforms to recognizable patterns special to the observers.

If I used all the Earth’s resources to write explicit blueprints of 500-bit patterns, even then, my collection of blue prints would be a drop in the bucket compared to all the possible configurations of 500-bits. The number of molecules on the Earth is on the order of 10^50, whereas the number of possible 500-bit patterns is 10^150. Thus pretty much any explicit pre-existing pattern available to a human can be used to identify design in 500-bits since 500 randomly scrambled bits aren’t expected to match the relatively small number of blue prints a human could record even if he devoted the entire Earth to the task of storing all his blue prints.

But such pre-existing blue prints are not available in biology, so how can we detect design in biology? To answer that question, first consider a 500-bit pattern defined by 500 uniquely numbered fair coins. Further, consider patterns where the first 250 bits are duplicated by the next 250 bits, the number of such patterns is 2^250 = 1.8 x 10^75. Ironically, as large as this number is, it is still a drop in the bucket compared to the Universal Probability Bound (UPB) of 2^500 = 3.3 x 10^150.

If I wanted to communicate design to another human using 500 uniquely number fair coins, I would probably use the simple organizing principle of duplicated or repeating patterns like the one in preceding paragraph. If we saw these duplicated patterns in coins, it would violate theoretical expectation of chance processes by several standard deviations, and if I really wanted to get the point across that a system was designed, I’d build the system with 10,000 coins not a mere 500.

So in addition to the finite memory of human observers, we have the problem the human observer may not be front-loaded with explicit patterns anyway. Therefore to communicate design to another human, it is best I resort to structuring an artifact according to organizing principles I presume a human observer will recognize rather than explicit patterns. The fact that humans can recognize designs they’ve never seen before but obey simple organizing principles was powerfully demonstrated in this case of casino cheating caught by the FBI. I also gave a classroom exercise to illustrate this fact here.

“All heads” is an extreme form of duplication and symmetry, but there are other forms of duplication and symmetry that can also convey design such as bit patterns where the first 250 bits duplicated in the next 250 bits as described above. Duplication and symmetry are organizing principles which the human mind is hard-wired to recognize, and a system evidencing such organizing principles will be recognized as designed if the system organization violates chemical and physical expectation in relation to the law of large numbers. The Law of Large Numbers is The Fundamental Law of ID because it provides the paradoxical fact that we can have almost definite knowledge about design in the face of maximum uncertainty.

Living systems have huge amounts of duplication that go against physical and chemical expectation. As I pointed out in relevance of coin analogies to homochirality and symbolic organization in biology the left-handed homochirality in amino acids, the right handed homochirality in DNA’s and sugars, the homogeneity of alpha-peptide bonds in proteins, the homogeneity of 3′-5′ links in DNA, etc. — all these violate chemical and physical expectation from pre-biotic soups, therefore design is strongly warranted.

But even at much simpler level, the fact that living organisms are close duplicates on many levels to other organisms suggests design. A Darwinists will argue, “Organisms are similar to each other because living systems make copies of themselves through reproduction.”

To which I’ll respond, “The process of reproduction is a copying process with 3-dimensional copying machines, and 3-dimensional copying machines are not the product of chance from something like a pre-biotic soup. Emergence of complex 3-dimensional copying machines like the first life from pre-biotic soups cannot be attributed to mindless processes.”

Just my personal opinion — it would seem the Designer of life went through a lot of trouble to make biology recognizable as a design in a fashion similar to the way a human might try to make a design recognizable to another human, except that God-made designs evidence a far greater level of intelligence. Even Dawkins noted:

it was hard to be an atheist before Darwin: the illusion [sic] of living design is so overwhelming.

Richard Dawkins

I’d even go so far to say, the Intelligent Designer of life knew far in advance our mental and technological capabilities, and designed biology in such a way that biological design would be eventually perceptible to us as our science progressed. As Gonzalez and Richards pointed out, it seems the universe was fine-tuned not only for life but for scientific discovery by humans. I speculate part of the reason for fine-tuning was so that humanity would eventually realize we were all created by God.

In addition to conforming to the simple organizing principles of duplication, biology also conforms to other more complex organizing principles such as functionality. Unfortunately, functionality is more difficult to quantify in terms of violations of chemical and physical expectation, but it can be done, but is beyond the scope of this essay. In this essay I’ve chosen to focus on the simplest cases of organizing principles: duplication, repetition, and symmetry.

Biology evidences both non-functional and functional organizing principles. Some biological examples of design might actually be classified as conforming to both functional and non-functional organizing principles. For example, homochirality and linkage homogeneity in one respect could be treated as non-functional evidence of design (just like 500 fair coins all heads is a non-functional design), but in another sense homochirality and linkage homogeneity are critical to functionality. I highlight homochirality and linkage homogeneity because they are among the simplest and clearest examples of design in biology.

A slightly more complicated organizing principle are strings that conform to a grammar which violates chemical and physical expectation. But like functional organization, organization according to grammars is beyond the scope of this essay. I’m working on an informal analysis of design detection via detecting grammars in biology. But hopefully this essay and the essays that are referenced in the links shows it is possible to identify design in biology.

One may wonder why I’m being so meticulous and rigorous to argue design with simple examples. The reason I do this is I want to negate the anti-Design argument that design detection is like seeing faces in clouds, that design n biology is a figment of our after-the-fact imagination. The argument of “imagined, after-the-fact specification” fails with 500 fair coins all heads and homochirality and linkage homogeneity in biology.

My aim with the recent series of essays is to blast into oblivion the false claim that design detection is an after-the-fact illusion of our minds or some abuse of statistics. That is why I appealed to expectation values and the law of large numbers. If we drop appeals to expectation, we might as well throw out the scientific method which relies heavily on the notion of expectation. Thus no one can say that I abused statistics or made after-the-fact specification in relation to 500 fair coins and the analogous patterns in biology.

NOTES

1. Design detection is NOT an argument from ignorance but rather a proof by contradiction against the chance hypothesis. See: The paradox of almost definite knowledge in the face of maximum uncertainty.

2. By the way, here is a 3-dimensional “printer”. It’s not even a 3-dimensional copying machine. Hopefully that drives the point home a little more regarding the challenge of building a 3-dimensional copying machine that not only makes copies of objects, but copies of the copying machine!

3d printer

That photo is of a Sony 3D printer. From the website:

http://www.consumeraffairs.com/news04/2012/08/could-3-d-printing-be-the-biggest-invention-since-the-internet.html

Okay, it’s Saturday night and you’re preparing to have a dinner party. The guests will be arriving in a few hours, and you’re racing around the house trying to pull both you and the dining room together.

The table is just about set, the last pot of food spews its last few bubbles before completion, and you’re thankful because you have just the right amount of your best plates for each guest.

Then tragedy: One of the plates falls and breaks in half. Normally this would be a catastrophe, but you just purchased a new 3D printer, so you’re able to calmly exhale.

You then print up a new dinner plate — identical to the original — place it on the table and finish getting ready for an evening of friends, conversation and food. And no, you don’t have to wear special glasses to see it.

Next big thing

PhotoThis way of self-manufacturing, and fixing things in your home is said to be the next global-changing-technology since the Internet or the television, some experts say.

“Personally, I believe it’s the next big thing, says Abe Reichental, president and CEO of 3D Systems, one of biggest companies that make 3D printing machines.

“I think it could be as big as the steam engine was in its day, as big as the computer was in its day, as big as the Internet was in its day. And I believe this is the next disruptive technology that’s going to change everything. It’s going to change how we learn, it’s going to change how we create, and it’s going to change how we manufacture,” he says.

3D printing and copying are non-trivial tasks. They require deep integration, computation, sensing, control — not any sort of machine we’d think could emerge from a random pre-biotic soup.

3. photo credits : rackcdn.com

4. The Privileged Planet hypothesis by Gonzalez and Richards says the universe is fine-tuned for life and scientific discovery by observers located in the privileged position in the universe known as Earth.

  • Delicious
  • Facebook
  • Reddit
  • StumbleUpon
  • Twitter
  • RSS Feed

2 Responses to Design recognition is possible in part because of finite human memory and limited human information

  1. Sal, I understand your 500 heads point, but it is only possible with reference to “fair coins” and defining what that means. wd400 raised an objection again recently and claimed that 500H could be obtained by a random walk. Now I’m not sure he is correct about that, but it is true that 500H could be the product of necessity (which is why you have to get into the “fair” definition of your coins, which detracts from the main point.

    I’m wondering if, rather than referring to 500H coins in your various posts, it might be better to refer to something that is clearly complex and specified, say the first 100 digits of pi or something or coins that spell out something in binary.

    Just a thought . . .

  2. but it is only possible with reference to “fair coins” and defining what that means.

    Even with a bias of 60% heads, the point would generally stand. Just to be sure I plugged it in the binomial calculator and the odds were still astronomical.

    I’m wondering if, rather than referring to 500H coins in your various posts, it might be better to refer to something that is clearly complex and specified, say the first 100 digits of pi or something or coins that spell out something in binary.

    Just a thought . . .

    That’s a good point, and I’m working on it.

    I wanted to start with something simple with the primary aim of refuting complaints of after-the-fact specifications. I think we have a good base of simple examples to work from which can be built upon. So you would be correct to say my examples are specified, but they are not complex in the Kolmogorov sense. They do evidence “specified improbability” (a term I much prefer to CSI).

    Now the bit patterns with modest duplication (like the first 250 bits duplicated by the last) can be considered moderately complex in the Kolmogorov sense if the first 250 bits are complex. Biased coins will actually make these modestly complex strings even more improbable. So then this class of strings have no need to assert the coins or bit probabilities are fair a priori, since bias will actually make them less probable if they are 50% heads or 50% 1′s.

    I think the next step is showing the design in strings or systems that evidence grammars which violate chance (i.e. computer languages and cellular languages in biology).

    The final stage is demonstrating chance cannot resolve proteins and functional systems like bat sonar, etc. That will be challenging. We know they violate chance, but the difficulty is formally showing this. You could see how hard it was just to drive the point home with the simple example of 500 fair coins heads.

    With the first 250 bits duplicated in the last 250 bits, and if the ratio of 1′s is 50%, we don’t even need to assume “fair” anymore. So now we’ve demonstrated specifications that have analogues in biology that are not after-the-fact. But these are still relatively simple.

    It will be challenging to describe design in more complex structures, and I think the ID community has not provided enough meticulous detail for probabilities involving grammars and functionality. I think the ID claims are correct, but the formalities of proof in ID literature are a little too sketchy for my tastes.

Leave a Reply