Can ANYTHING Happen in an Open System?
| October 19, 2006 | Posted by GilDodgen under Intelligent Design |
In a previous UD post I commented on an article by mathematician Granville Sewell, “A Mathematician’s View of Evolution.†Since then Granville and I have corresponded and he forwarded a follow-up piece entitled, “Can Anything Happen in an Open System?”
The essence of the thesis is as follows:
If an increase in order is extremely improbable when a system is closed, it is still extremely improbable when the system is open, unless something is entering which makes it NOT extremely improbable.
Here are some excerpts:
Critics of my Opinion piece, “A Mathematician’s View of Evolution,†have focused primarily on my first point, which deals with the question of whether or not major evolutionary improvements can be built up through many minor improvements. It is clear to me that they cannot, but this question is the traditional front on which most battles over Darwinism have been fought since 1859, and I did not imagine that my arguments would constitute the last word on this topic. I consider that the main point in my article was the second one.
…which is that the increase in order observed on Earth (and here alone, as far as we know) violates the laws of probability and the second law of thermodynamics in a spectacular fashion.
Evolutionists have always dismissed this argument by saying that the second law of thermodynamics only dictates that order cannot increase in an isolated (closed) system, and the Earth is not a closed system–in particular, it receives energy from the Sun. The second law allows order to increase locally, provided the local increase is offset by an equal or greater decrease in the rest of the universe. This always seems to be the end of the argument: order can increase (entropy can decrease) in an open system, therefore, ANYTHING can happen in an open system, even the rearrangement of atoms into computers, without violating the second law.
It requires only a modicum of common sense to see that it is extremely improbable that atoms should rearrange themselves into mammalian brains, computers, cars, and airplanes, even if the Earth does receive energy from the Sun. We will see that the idea that anything can happen in an open system is based on a misunderstanding of the second law; that order can increase in an open system, not because the laws of probability are suspended when the door is open, but simply because order may walk in through the door.
See also his article, “A Second Look at the Second Law.â€Â
42 Responses to Can ANYTHING Happen in an Open System?
Leave a Reply
You must be logged in to post a comment.
DvK
Orgel et al are on a wild goose chase. I’m sure it will be a lucrative area for research grants for some time to come but ultimately it will come up short. I can think of little more breathtaking than discovering that life was designed.
niwrad
There is zero CSI in snowflakes and other crystals. Snowflakes are complex but have no specification. Regular crystals are specified but are not complex. You don’t really have a good handle on CSI if you don’t understand that. A specification is an independently given pattern. A cubics crystal is specified (it adheres to the independently given pattern of a cube) but a cube isn’t complex. A snowflake has a complex shape but adheres to no independently given pattern. A gene is complex specified information. It is complex and adheres to the independently given pattern of the genetic code. Once something meets the criteria for CSI then one can examine probalistic resources available to form the pattern. Determining PR is the most difficult step in a design inference as it cannot be proven that all probalistic resources are known (lack of evidence is not proof of lack). However, one can become more confident that all possibilities are accounted for at some point and a tentative conclusion may be warranted. If the probability of the pattern emerging by unintelligent means is less than the universal probability bound (1 in 10^150) then according to Dembski a design inference is warranted.
DharmaBum
A specification in CSI is an independently given pattern. The snowflake patterns you linked were not given independently but taken from observation of actual snowflakes. The pattern must be independently given and these are not independent of snowflakes.
DaveScot:
Much of what you say about CSI has been superceded, if not obsoleted, by “Specification: The Pattern That Signifies Intelligence.” The notion of an independently given pattern was always vague, and I cannot find it anywhere in the paper. CSI is
Ç = –log_2[ 10^120 · Æ_S(T) · P(T|H)].
See the paper for what the factors mean. It is highly unlikely that there is zero CSI in any entity.
No, when Ç > 1, a design inference is warranted, and pattern T is called a specification. Note that the bound in the definition of Ç is 10^120, not 10^150, and that it is not really a probability bound, because it gives a threshold above which the product of specificity and “complexity” (probability) must rise.
The S in Dembski’s equation is a semiotic agent that uses its background knowledge to designate the pattern T. There is no restriction that the agent cannot have acquired background knowledge through observations of similar phenomena in the past. Furthermore, all that matters in a design inference is that one semiotic agent S be able to supply a specification.
Now if we observe a new snowflake and determine that it matches one of the 35 patterns, and we do not understand the physical process of snowflake formation well enough to give a detailed causal account (i.e., make P(T|H) non-small), then Ç will exceed 1 and we will draw a design inference.
It’s essential that you read the paper. CSI is not what it used to be.
“People on UD actually discussed far-from-equilibrium systems a while back, specifically the work of Ilya Prigogine.”
Where? I’ve searched on “Prigogine” and “far from equilibrium”, but haven’t found much (I don’t think).
DharmaBum
No, nothing I said has been superceded. I read the paper you mentioned a year ago. You evidently need to read it again. I quote
According to Dembski specification did not change in kind from the past as you asserted.
So you are also wrong in saying that detachability and independence are no longer part of specification. In the more recent paper Dembski calls it “prespecification” and says it must be part of any specification.
I reiterate that the snowflake patterns you provided are not independent. They were formed by observing snowflakes. They cannot be detached from snowflakes. A snowflake has zero CSI if it has zero specification.
DaveScot:
Rather than invoke Dembski’s authority in “Addendum 1: Note to Readers or TDI & NFL,” would you please show me in the paper itself the support for your claims?
By the way, the fact that Dembski says the new treatment is in “harmony” with earlier ones does not mean that it does not supercede earlier ones. He refers to “a fundamental
simplification to the account of specification that appears in this paper.” Why would this fundamental simplification not supercede past accounts?
If “such key notions as detachability, conditional independence, tractability, and universal probability bounds … no longer need to be made explicit,” and you read the paper a year ago, why are you still talking about them explicitly. That is, why have you not embraced the “fundamental simplification”? I think the paper is a huge improvement over NFL.
First, I never said a thing about detachability. Second, show them to me in Section 6, “Specificity.”
Even in Addendum 1, Dembski says:
Dembski clearly distinguishes prespecification (before an event) and specification (after an event), here and throughout the paper, and links the old conditional independence condition to prespecification. Again, if you can link conditional independence to specification in the body of the paper, please show me where. By the way, note that Dembski writes on page 15:
DaveScot:
Reiteration gets us nowhere. My claim is that there is nothing in the treatment of specificity in Section 6 that precludes the semiotic agent S having knowledge of snowflake patterns. You can resolve this easily by showing me the restriction.
The phrase “zero specification” is meaningless. A specification is a kind of pattern (description), not a quantity. Specificity
ÃÆ’ = –log_2[Æ_S(T)·P(T|H)]
is zero if and only if the descriptive complexity Æ_S(T) is the reciprocal of the “complexity” P(T|H). This simply is not going to happen very often at all. But let’s say that ÃÆ’ = 0. Then CSI
Ç = –log_2[ 10^120] + ÃÆ’ = –398.6,
which is far from zero.
DharmaBum
Rather than invoke Dembski’s authority in “Addendum 1: Note to Readers or TDI & NFL,†would you please show me in the paper itself the support for your claims?
No, I won’t. I can’t quote a higher authority on a paper than the author himself in the same paper. Dembski explicitely said detachability and independence are still inherent in the newer presentation. Snowflake patterns derived from the observation of snowflakes is simply not independent of snowflakes. It’s dependent on snowflakes.
IMHO Dave is right because of the presence of an obvious but implicit condition for the Ç formula
Ç = –log_2[ 10^120 · Æ_S(T) · P(T|H)].
When the patterns under consideration do strictly refer to a given amount of previous observations (and this is the case for snowflake structures), in a certain sense recognizing that a new pattern matches is simply a tautology. So this does not constitute a specification at all.
Maybe snowflakes that actually exhibit a detachable, independently given pattern will help DharmaBum distinguish between snowflakes whose pattern is that of a snowflake from snowflakes with a detachable pattern. In this example the foreknowledge of a semiotic agent is required to recognize the pattern but no foreknowledge of snowflakes is required.
http://tinyurl.com/ye4oto
DharmaBum Wrote (special emphasis by mattison0922:
Perhaps ‘too destructive’ is the wrong way to describe it. That cells require heat energy at least to keep water in a liquid state, and to permit biologically relevant reactions, is undeniable. Furthermore that certain extremophiles can tolerate such temperature extremes as those you cite is undeniably a remarkable adaptive feat.
In any case, the problem with heat energy is that there’s nothing to harness per se. It’s the final fate in some energies conversion to the ultimate state of entropy.
There’s no potential in heat to harness, thus heat is an inadequate souce of energy for biological systems to harness. Biologicals appear to require a source of energy that can be harnessed in an incremental or stepwise fashion. Heat doesn’t fit the bill.
I don’t believe I understand the relevance of your final paragraph. I wasn’t implying that these organisms are thermoautotrophs, rather pointing out that there is no thermoautotrophic class of organisms known. IOW heat, thus far, appears to be an inadequate source of energy to power biological reactions.
I bolded a highly relevant point in your quote above.
Due to his refusing to recognize that snowflake patterns derived from looking at snowflakes is self-referential DharmaBum is no longer with us. He’s done wasting our time here.
http://search.yahoo.com/search.....8;ei=UTF-8
The 3rd link isn’t initially about Prigogine’s work but the discussion leads toward it. This quote from Prigogine from before he died is relevent:
“The results of my research in thermodynamics were to show that non-equilibrium systems may lead to complex structures. For a recent account, see my book “Modern Thermodynamics, From Heat Engines to Dissipative Structures†(D. Kondepudi and I. Prigogine, John Wiley & Sons, Chichester, 1998). However, this is still far from a theory of biological evolution. To my knowledge, we have still not discovered the mechanisms, which lead to the remarkable adaptation between life, and environments, which we observe in nature.â€Â
On a side note, this discussion we’re now in has been “educational” since it shows how some ID opponents don’t even comprehend what they’re opposing.