Uncommon Descent Serving The Intelligent Design Community

Orgel and Dembski Redux

Share
Facebook
Twitter
LinkedIn
Flipboard
Print
Email

 

A couple of months ago I quoted from Lesli Orgel’s 1973 book on the origins of life.  L. E. Orgel, The Origins of Life: Molecules and Natural Selection (John Wiley & Sons, Inc.; New York, 1973).  I argued that on page 189 of that book Orgel used the term “specified complexity” in a way almost indistinguishable from the way Bill Dembski has used the term in his work.  Many of my Darwinian interlocutors demurred.  They argued the quotation was taken out of context and that Orgel meant something completely different from Dembski.  I decided to order the book and find out who was right.  Below, I have reproduced the entire section in which the original quotation appeared.  I will let readers decide whether I was right.  (Hint: I was).

 

All that follows is a word-for-word reproduction of the relevant section from Orgel’s book:

 

[Page 189]

Terrestrial Biology

Most elementary introductions to biology contain a section on the nature of life.  It is usual in such discussions to list a number of properties that distinguish living from nonliving things. Reproduction and metabolism, for example, appear in all of the lists; the ability to respond to the environment is another old favorite.  This approach extends somewhat the chef’s definition “If it quivers, it’s alive.” Of course, there are also many characteristics that are restricted to the living world but are not common to all forms of life.  Plants cannot pursue their food; animals do not carry out photosynthesis; lowly organisms do not behave intelligently.

It is possible to make a more fundamental distinction between living and nonliving things by examining their molecular structure and molecular behavior.  In brief, living organisms are distinguished by their specified complexity.*· Crystals are usually taken as the prototypes of simple, well-specified structures, because they consist of a very large number of identical molecules packed together in a uniform way.  Lumps of granite or random mixtures of polymers are examples of structures which are complex but not specified.  The crystals fail to qualify as living because they lack complexity, the mixtures of polymers fail to qualify because they lack specificity.

_______

* It is impossible to find a simple catch phrase to capture this complex idea.  “Specified and. therefore repetitive complexity” gets a little closer (see later).

[Page 190]

These vague ideas can be made more precise by introducing the idea of information.  Roughly speaking, the information content of a structure is the minimum number of instructions needed to specify the structure.  One can see intuitively that many instructions are needed to specify a complex structure.  On the other hand, a simple repeating structure can be specified in rather few instructions.  Complex but random structures, by definition, need hardly be specified at all.

These differences are made clear by the following example.  Suppose a chemist agreed to synthesize anything that could describe [sic] accurately to him.  How many instructions would he need to make a crystal, a mixture of random DNA-like polymers or the DNA of the bacterium E. coli?

To describe the crystal we had in mind, we would need to specify which substance we wanted and the way in which the molecules were to be packed together in the crystal.  The first requirement could be conveyed in a short sentence.  The second would be almost as brief, because we could describe how we wanted the first few molecules packed together, and then say “and keep on doing the same.”  Structural information has to be given only once because the crystal is regular.

It would be almost as easy to tell the chemist how to make a mixture of random DNA-like polymers.  We would first specify the proportion of each of the four nucleotides in the mixture.  Then, we would say, “Mix the nucleotides in the required proportions, choose nucleotide molecules at random from the mixture, and join them together in the order you find them.”  In this way the chemist would be sure to make polymers with the specified composition, but the sequences would be random.

It is quite impossible to produce a corresponding simple set of instructions that would enable the chemist to synthesize the DNA of E. coli.  In this case, the sequence matters; only by specifying the sequence letter-by-letter (about 4,000,000 instructions) could we tell the chemist what we wanted him to make.  The synthetic chemist would need a book of instructions rather than a few short sentences.

It is important to notice that each polymer molecule in a random mixture has a sequence just as definite as that of E.

[Page 191]

coli DNA.  However, in a random mixture the sequences are not specified, whereas in E. coli, the DNA sequence is crucial.  Two random mixtures contain quite different polymer sequences, but the DNA sequences in two E. coli cells are identical because they are specified.  The polymer sequences are complex but random; although E. coli DNA is also complex, it is specified in a unique way.

The structure of DNA has been emphasized here, but similar arguments would apply to other polymeric materials.  The protein molecules in a cell are not a random mixture of polypeptides; all of the many hemoglobin molecules in the oxygen-carrying blood cells, for example, have the same sequence.  By contrast, the chance of getting even two identical sequences 100 amino acids long in a sample of random polypeptides is negligible.  Again, sequence information can serve to distinguish the contents of living cells from random mixtures of organic polymers.

When we come to consider the most important functions of living matter, we again find that they are most easily differentiated from inorganic processes at the molecular level.  Cell division, as seen under the microscope, does not appear very different from a number of processes that are known to occur in colloidal solutions.  However, at the molecular level the differences are unmistakable:  cell division is preceded by the replication of the cellular DNA.  It is this genetic copying process that distinguishes most clearly between the molecular behavior of living organisms and that of nonliving systems.  In biological processes the number of information-rich polymers is increased during growth; when colloidal droplets “divide” they just break up into smaller droplets.

Comments
fifthmonarchyman: Pi has meaning in a way that the algorithm 4*(1 – 1/3 + 1/5 – 1/7 + …) does not. http://www.smbc-comics.com/?id=3639#comicZachriel
February 20, 2015
February
02
Feb
20
20
2015
11:58 AM
11
11
58
AM
PDT
I see that keith s cannot support his claim.
To specify a cylindrical crystal of pure silicon, you would merely need to 1. Specify the unit cell. 2. Specify the spatial pattern for adding new unit cells to an existing crystal. 3. Specify the boundaries of the cylindrical crystal (to stop adding unit cells when the crystal has reached the desired size and shape).
That seems pretty complex, keith s. You lose, again, as usual.Joe
January 30, 2015
January
01
Jan
30
30
2015
05:21 AM
5
05
21
AM
PDT
Define specified. Without being circular. Microevolution -- a process many IDists agree happens -- modifies specifications. Yes or no?Petrushka
January 30, 2015
January
01
Jan
30
30
2015
04:52 AM
4
04
52
AM
PDT
There is no such thing as a simple living being. There is no such thing as a complex living being that is not specified. Orgel, or Dembski?
It's Mung.E.Seigner
January 30, 2015
January
01
Jan
30
30
2015
01:05 AM
1
01
05
AM
PDT
fifthmonarchyman: But “Pi” works just as well and has less K-complexity than the algroythym. Pi is not a useful answer to Please send the value of pi, the ratio of the circumference of a circle to its diameter.Zachriel
January 26, 2015
January
01
Jan
26
26
2015
05:40 PM
5
05
40
PM
PDT
fifthmonarchyman: I suppose you could treat the algroythym like a specification you could treat any symbol as such Pi = 4 * (1 – 1/3 + 1/5 – 1/7 + … ) fifthmonarchyman: Do you honestly think I was arguing that the algroythym that I provided to approximate Pi could not approximate Pi? What you claimed was that the algorithm based on 4 * (1 – 1/3 + 1/5 – 1/7 + … ) had unbounded K-complexity. That is false. It's K-simple. fifthmonarchyman: Do you think that algorithmic processes can’t be simultaneous? An algorithm is a step-by-step process.Zachriel
January 26, 2015
January
01
Jan
26
26
2015
05:27 PM
5
05
27
PM
PDT
Zac says, Equal means equal. The ellipses means infinite expansion, which is an abstraction independent of time. I say I suppose you could treat the algroythym like a specification you could treat any symbol as such But "Pi" works just as well and has less K-complexity than the algroythym. What you can't do is treat it like an algroythym and then proceed to measure it's complexity as if it were a specification. you said, Glad you gave up your misunderstanding of K-complexity I say, Do you honestly think I have changed my position about K=complexity in the slightest? Do you honestly think I was arguing that the algroythym that I provided to approximate Pi could not approximate Pi? Geez Zac said, To return to your original idea, it’s not clear that NV + NS (natural variation and natural selection) is algorithmic as events can be simultaneous. I say, Do you think that algorithmic processes can't be simultaneous? 5*2=2+2+2+2+2 after all you say, Nor is there a specific solution or end to the process. I say, The solution is what the algorithm explains in this case the Panorama of life. If you are conceding that Neo-Darwinism can not explain the Panorama of life. Then we are in agreement and my point is made you say, Furthermore, neither NV or NS are simple. I say, I completely agree and that just means that it will require even more K-complexity to specify the model. peacefifthmonarchyman
January 26, 2015
January
01
Jan
26
26
2015
04:35 PM
4
04
35
PM
PDT
fifthmonarchyman: The equal sign is a convention that assumes infinite time. Equal means equal. The ellipses means infinite expansion, which is an abstraction independent of time. fifthmonarchyman: Here is an approximation of Pi 3.14159. If you need an increased resolution you might try squaring a circle. It won’t get you to Pi but it’s probably good enough for government work. A person can get unbounded precision with the arithmetic series. fifthmonarchyman: infinite numbers can not be completely reproduced by finite algorithms in finite time. Glad you gave up your misunderstanding of K-complexity. To return to your original idea, it's not clear that NV + NS (natural variation and natural selection) is algorithmic as events can be simultaneous. Nor is there a specific solution or end to the process. Furthermore, neither NV or NS are simple. The latter entails not only the environment, but the relationship between the structure of the organism and the characteristics of the environment. (Consider that a sequence may fold into a complex three-dimensional shape with charges unevenly distributed along its surface, and how this shape then interacts with other such structures.) The former entails many types of variation, mutation, recombination, splicing, etc. Then there's also contingency, branching descent with all its exceptions, and so on.Zachriel
January 26, 2015
January
01
Jan
26
26
2015
03:01 PM
3
03
01
PM
PDT
zac says, A finite algorithm can produce an infinite sequence in infinite time. I say, fair enough I will amend the statement..... infinite numbers can not be completely reproduced by finite algorithms in finite time. How is that for you? you say, Please note the equal sign. I say, The equal sign is a convention that assumes infinite time. Don't you agree? You say, What is your reply? I say Dear sirs, "I can't send you an infinite sequence over this channel. Here is an approximation of Pi 3.14159. If you need an increased resolution you might try squaring a circle. It won't get you to Pi but it's probably good enough for government work" I would give a similar reply if someone asks me to send my wife over the internet. I might send a picture and a note on how they can learn more about her. That is all I can do. What I would not do is send a 2D picture or an evolutionary algorithm and claim that it "equals" my wife. peacefifthmonarchyman
January 26, 2015
January
01
Jan
26
26
2015
02:39 PM
2
02
39
PM
PDT
fifthmonarchyman: All I’m saying is that infinite numbers can not be completely reproduced by finite algorithms. That is incorrect. A finite algorithm can produce an infinite sequence in infinite time. fifthmonarchyman: 4*(1 – 1/3 + 1/5 – 1/7 + …) is an algorithm to approximate Pi. That is also incorrect. Pi = 4 * (1 – 1/3 + 1/5 – 1/7 + … ). Please note the equal sign. Someone transmits this request: Please send the value of pi, the ratio of the circumference of a circle to its diameter. What is your reply?Zachriel
January 26, 2015
January
01
Jan
26
26
2015
06:20 AM
6
06
20
AM
PDT
skram, This is non-controversial. It is also irrelevant. I say, Irrelevant for what? We were are talking about how different measures of complexity and how they are related so the k-complexity of an infinite number is probably relevant. Don't you agree? You say, K-complexity is defined for finite strings. I say, I'm not disagreeing So now we need to decide if it can be used for infinite ones. I'd say yes as long as we are agree that it is un-bounded for such strings. I'm willing to hear opposing arguments but they need to be informed arguments not just pronouncements you said, Piotr’s point illustrates well that a long sequence of digits of ? is a K-compressible object, perfectly in agreement with Kolmogorov’s definition. I say, That's why I agreed with him. Geez peacefifthmonarchyman
January 25, 2015
January
01
Jan
25
25
2015
02:12 PM
2
02
12
PM
PDT
fifthmonarchyman: All I’m saying is that infinite numbers can not be completely reproduced by finite algorithms. This should not be controversial This is non-controversial. It is also irrelevant. K-complexity is defined for finite strings. Piotr's point illustrates well that a long sequence of digits of π is a K-compressible object, perfectly in agreement with Kolmogorov's definition.skram
January 25, 2015
January
01
Jan
25
25
2015
01:41 PM
1
01
41
PM
PDT
skram says, I don’t understand what you are saying. I say All I'm saying is that infinite numbers can not be completely reproduced by finite algorithms. This should not be controversial you say K-complexity is about specification and we all can agree that the digits of Pi represent a compressible string since it can be specified in a more economical way than by writing out the string itself. I say again I'm not denying the compressibility I'm denying that the compression is nonlossy. Again this should not be controversial a finite algorithm can not completely reproduce an infinite string. peacefifthmonarchyman
January 25, 2015
January
01
Jan
25
25
2015
01:27 PM
1
01
27
PM
PDT
fifthmonarchyman, I don't understand what you are saying. The compression of the digits of π is certainly algorithmic. Zachriel's recipe gives an explicit algorithm. Your specification requires a recipe (i.e., an algorithm) through which the circumference is computed. (E.g., through a polygon approximation.) But anyway, K-complexity is about specification and we all can agree that the digits of π represent a compressible string since it can be specified in a more economical way than by writing out the string itself.skram
January 25, 2015
January
01
Jan
25
25
2015
01:11 PM
1
01
11
PM
PDT
Skram says, To specify and to compute are two different things. I say That is my point!!!! Tell that to Zac he is treating an approximating algorithm as a specification. you say Your own specification (the ratio of the circumference and diameter of a circle) does not give any digits of Pi, it merely specifies one recipe to obtain the number. I say, exactly!!!! I'm not trying to calculate Pi with my specification I'm merely specifying it. The K-complexity of the symbol Pi is simple and the compression it yields is non algorithmic and nonlossy but at the same time complete. That is the beauty of axioms, Axioms are not algorithms peacefifthmonarchyman
January 25, 2015
January
01
Jan
25
25
2015
01:03 PM
1
01
03
PM
PDT
fifthmonarchyman: Agreed the question is can a finite algorithm ever produce all the digits of Pi. No, that isn't the question. To specify and to compute are two different things. Your own specification (the ratio of the circumference and diameter of a circle) does not give any digits of π, it merely specifies one recipe to obtain the number.skram
January 25, 2015
January
01
Jan
25
25
2015
12:50 PM
12
12
50
PM
PDT
Hey Guys, This is all about lossy verses non-lossy compression. I agree that 4*(1–1/3+1/5–1/7+…) can compress 3.141592653589793238462643383279502884197169399375105820974944592307816406286…… But information is inevitably lost the string is infinite the product of the algorithm is finite. peacefifthmonarchyman
January 25, 2015
January
01
Jan
25
25
2015
12:50 PM
12
12
50
PM
PDT
Piotr says, the Kolmogoriov complexity of Pi is C (a constant, computable from the length of the minimal algorithm that generates the digits of Pi) I say, Agreed the question is can a finite algorithm ever produce all the digits of Pi. The answer is no you say, The complexity of the string of the first n digits of Pi is C + log(n), because in addition to the algorithm you have to specify n. I say, again agreed. It's important to say that we don't need to know all the digits to know n digits but the term Pi specifies all of them I'm not sure you intend to but I think we are saying the same thing here. peacefifthmonarchyman
January 25, 2015
January
01
Jan
25
25
2015
12:44 PM
12
12
44
PM
PDT
fifthmonarchyman, take a look at Piotr's last comment. Writing down n digits of π directly requires a string of length n. However, we can write an algorithm outputting the n digits as a program of length of the order of log(n). This program is much shorter than the original string. Hence the digits of π represent a K-compressible string.skram
January 25, 2015
January
01
Jan
25
25
2015
12:39 PM
12
12
39
PM
PDT
fifthmonarchyman, Formally, the Kolmogoriov complexity of Pi is C (a constant, computable from the length of the minimal algorithm that generates the digits of Pi). The complexity of the string of the first n digits of Pi is C + log(n), because in addition to the algorithm you have to specify n.Piotr
January 25, 2015
January
01
Jan
25
25
2015
12:33 PM
12
12
33
PM
PDT
hey Skram, Zac is attempting to show that a specification and a algorithmic model are the same thing. This is the what the discussion is about. the term "J's#" has more information inherent in it than this algorithm 8675000+400-89 Can't you see that? peacefifthmonarchyman
January 25, 2015
January
01
Jan
25
25
2015
12:30 PM
12
12
30
PM
PDT
skram says The definition on Wikipedia mentions the computational resources needed to specify an object, not resources to actually compute it. I say No I'm reading carefully 4*(1 – 1/3 + 1/5 – 1/7 + …)is an algorithm to approximate Pi. Pi is a term that specifies the ratio of a circle's circumference to its diameter Do you not see the difference? peacefifthmonarchyman
January 25, 2015
January
01
Jan
25
25
2015
12:23 PM
12
12
23
PM
PDT
Skram says, Kolmogorov’s complexity only asks the question of whether it is possible to write a short program that can output ?’s digits. I say, What? K-complexety can be given for a text string. It is not only interested in digits. Besides the N in question is infinite. We are not asking for an algorithm to compute Pi to the nth digit we are asking for a an algorithm to specify all the digits. No finite algorithm can compute an infinite number in finite time. Don't you agree? But with the symbol/axiom Pi I can specify the whole shebang at one time all at once peacefifthmonarchyman
January 25, 2015
January
01
Jan
25
25
2015
12:17 PM
12
12
17
PM
PDT
fifthmonarchyman, You are not reading carefully. The definition on Wikipedia mentions the computational resources needed to specify an object, not resources to actually compute it.skram
January 25, 2015
January
01
Jan
25
25
2015
12:13 PM
12
12
13
PM
PDT
Zac says Pi is K-simple. I say Here is the definition of k-complexity quote: the Kolmogorov complexity (also known as descriptive complexity, Kolmogorov–Chaitin complexity, algorithmic entropy, or program-size complexity) of an object, such as a piece of text, is a measure of the computability resources needed to specify the object. end quote: from here http://en.wikipedia.org/wiki/Kolmogorov_complexity the computability resources needed to specify the pi algorithmically is infinite. Pi is infinite you can't get to Pi by a finite process. I'm not sure why this is so difficult to grasp. you say. Notice the equal sign. They are the same number. If I Google Pi I get the following Pi=3.14159265359 Surely you don't believe that that short 13 digit number is all there is to Pi? You say, Try to avoid using words like Kolmogorov complexity and computability, as you clearly don’t understand their meaning in mathematics. I say Try to avoid making official sounding pronouncements about mathematical meanings unless you are able to give evidence of your claims. Please explain with evidence how a finite algorithmic process can fully specify an infinite number peacefifthmonarchyman
January 25, 2015
January
01
Jan
25
25
2015
12:08 PM
12
12
08
PM
PDT
fifthmonarchyman, you are wrong about π. Kolmogorov's complexity only asks the question of whether it is possible to write a short program that can output π's digits. It does not ask the question of how long it will take. The formula cited by Zachriel is such an algorithm. It will reproduce π to an arbitrary precision, given enough time. In fact, you can even tell how many terms in that expression must be summed in order to achieve a given accuracy. Thus every single digit of π is obtainable through that algorithm. Therefore π turns out to be K-compressible.skram
January 25, 2015
January
01
Jan
25
25
2015
12:02 PM
12
12
02
PM
PDT
fifthmonarchyman: The fact is I’m using the standard definition of Kolmogorov complexity No, you're not. Pi is K-simple. fifthmonarchyman: Pi has meaning in a way that the algorithm 4*(1 – 1/3 + 1/5 – 1/7 + …) does not. Pi = 4 * (1 – 1/3 + 1/5 – 1/7 + … ) Notice the equal sign. They are the same number. fifthmonarchyman: Pi can’t be produced by any algorithm simple or otherwise it can only be approximated. That is incorrect. Pi is K-simple per the definition of Kolmogorov complexity, because it can be expressed as a simple algorithm. It's the length of the shortest algorithm that determines K-complexity. ETA: Try to avoid using words like Kolmogorov complexity and computability, as you clearly don't understand their meaning in mathematics.Zachriel
January 25, 2015
January
01
Jan
25
25
2015
11:35 AM
11
11
35
AM
PDT
Zac says, Kolmogorov complexity is a proper noun. You can’t change its definition willy-nilly. I say, Why is it that any time I explore the implications of a concept you accuse me of changing definitions? The fact is I'm using the standard definition of Kolmogorov complexity you just don't like the implications I'm drawing You say, Pi is Kolmogorov simple because it is can produced by a simple algorithm. I say, Pi can't be produced by any algorithm simple or otherwise it can only be approximated. You say, Suppose you wanted to send someone your message, the sequence above, to some arbitrary limit. You could send the literal, but because it is non-compressible, it might take a while. Or you could send a short algorithm, and the recipient could calculate the expansion themselves. I say, You are still missing the point. I do not want to communicate the sequence to some arbitrary limit. I want to communicate the entire sequence all of it In fact to communicate to an arbitrary limit is to loose information that is central to what I want to convey. It's the difference between a 2D picture of my wife and my wife. 3.141592653589793238462643383279502884197169399375105820974944592307816406286…… is not compressible by any means algorithmic or otherwise unless you know the specification. Then it is highly compressible. Just like J's# Pi has meaning in a way that the algorithm 4*(1 – 1/3 + 1/5 – 1/7 + …) does not. A specification is not the same as an approximating algorithm. Not sure how many more ways I can state it. peacefifthmonarchyman
January 25, 2015
January
01
Jan
25
25
2015
10:10 AM
10
10
10
AM
PDT
fifthmonarchyman: It will only approximate the number to ever greater accuracy forever. Kolmogorov complexity is a proper noun. You can't change its definition willy-nilly. Pi is Kolmogorov simple because it is can produced by a simple algorithm. Suppose you wanted to send someone your message, the sequence above, to some arbitrary limit. You could send the literal, but because it is non-compressible, it might take a while. Or you could send a short algorithm, and the recipient could calculate the expansion themselves.Zachriel
January 25, 2015
January
01
Jan
25
25
2015
09:11 AM
9
09
11
AM
PDT
Zac says That is incorrect. Given arithmetic, there is a simple algorithm which can calculate the series. You provided one yourself. I say That algorithm will never get you to 3.141592653589793238462643383279502884197169399375105820974944592307816406286…… It will only approximate the number to ever greater accuracy forever. Peacefifthmonarchyman
January 25, 2015
January
01
Jan
25
25
2015
08:48 AM
8
08
48
AM
PDT
1 2 3 5

Leave a Reply