What is information?

[cypress]

My example is an example of evolution creating a new function. The function was not there, and then evolution created it.

 

Since when is reviving existing function defined as new function? You redefined the question to one easier to answer.

 

The reason I used that particular example is that you want an example that will occur within your lifetime.

 

No, any time period will do. Molecular biology discoveries have uncovered the precursor elements required for new functional information. They are new protien tertiary structures, new protein binding sites, new expression controls, process controls and developmental controls to name a few. These precursor steps on average involve 5-10 point alterations in gene primary sequences. If mutation and selection account for generation of biological diversity and thus novel functional information, then identifying generation of these precursor elements by this process should be common occurrence.

 

Compare the speed for evolution to create the new function in this example -- a few generations -- to the speed of a blind search (many times the lifetime of the universe), and you see it is much much faster than a blind search (but less thorough).

 

1) It's not new function, 2) It is a two step process that occurs at the same rate as a random walk, one step off of function and one step back to function

 

Information is generated at 1 bit per coin flip.

http://en.wikipedia.org/wiki/Entropy_%28information_theory%29

 

Yes, I believe you are confirming that information is generated by blind search and random walk but on average no faster. In this example where any pattern is defined as information, information is generated at the rate equal to the pure random process. When functional information is desired, one also finds that functional information is generated at a rate averaging blind search when random processes are involved.

 

So which human created us all then?

 

I don't make that claim.

 

If as you say only humans have been observed to create information (false, regardless of your modifiers),

 

Only design has been observed to create functional information faster than blind search. I don't limit this to humans.

 

then what created us must therefore have been human.

 

No.

 

You proposing a non-human intelligence with no evidence other than an argument from ignorance, compared to scientists arguing evolution from a common ancestor via known and verified processes.

 

No, I make no claim about the character of this intelligence. I note that biological systems contain vast quantities of functional information that thus far only intelligence has been shown capable of generating at sufficient rates. The challenge to those who advocate for natural processes (including myself) is to identify a natural process capable of generating functional information at a rate consistent with generation of life and diversification of life.

 

 

Although evolution is not directly seen at the largest scales (nor expected to be), the evidence is in the DNA sequences that it was so. I'd go with the option with the evidence.

 

The evidence from DNA is that diversification occurred, but the evidence does not give us any indication of how it occurred. Evidence from DNA also indicates that in addition to diversification, some processes also adapt and alter, and damage and repair, existing function, and insert and remove nonfunctional information. Mutation and selection explains the latter nicely, but there remains no natural explanation for original generation and diversification.

[pioneer]

The DNA has proof reading enzymes that go over the duplicated DNA to correct base pairing mistakes. The number of mutated genes starts higher, with the proof reading enzymes reducing this to only a few typo's. The improper base pairing means more potential in the hydrogen bonds, since they don't match perfectly. The proof readers feel along the DNA until they feel this extra potential in an improper base pair (mutation) and then lower this potential via proper base pairing. These enzymes lower entropy within the DNA, by taking away its source of energy. Now the deck of cards starts is less shuffled than before.

Edited August 30, 2010 by pioneer

[Mr Skeptic]

Since when is reviving existing function defined as new function? You redefined the question to one easier to answer.

 

No, per the problem I specifically state that the gene codes for a non-functional protein. Which according to your definition, is zero information since it is non-functional. So from zero information to an entirely new functional gene in (I estimate about a day for bacteria).

 

No, any time period will do. Molecular biology discoveries have uncovered the precursor elements required for new functional information. They are new protien tertiary structures, new protein binding sites, new expression controls, process controls and developmental controls to name a few. These precursor steps on average involve 5-10 point alterations in gene primary sequences. If mutation and selection account for generation of biological diversity and thus novel functional information, then identifying generation of these precursor elements by this process should be common occurrence.

 

Yes, you say any time period and then demand that it be observed. Which severely limits the amount of time. Anyhow, the example is good enough.

 

1) It's not new function, 2) It is a two step process that occurs at the same rate as a random walk, one step off of function and one step back to function

 

1) the function is not there, so it is a new function within the experiment.

 

2) You said blind search, not random walk. Blind search means you don't get to limit your search to those close to the original organism, but have to try every single base pair combination until you get the right one. To get that specific improvement blind search would take so many times the life time of the universe that it is pretty much guaranteed never to happen via blind search.

 

I estimate evolution would do it in about a day with bacteria (10^-8 mutations per base pair, ~20 mins generation time if allowed to replicate freely for 24 hrs would be 10^21 individuals), needs only a single specific mutation.

 

Only design has been observed to create functional information faster than blind search. I don't limit this to humans.

 

And what other than humans have been observed to design? Are you saying maybe an animal designed us? Or maybe that something that has never been observed is needed to explain something you claim has never been observed? Would it not be simpler to just cut out the middleman?

[cypress]

No, per the problem I specifically state that the gene codes for a non-functional protein. Which according to your definition, is zero information since it is non-functional. So from zero information to an entirely new functional gene in (I estimate about a day for bacteria).

 

By your measure, plagiarism would be a valid method of generating a new research paper or report.

 

 

 

Yes, you say any time period and then demand that it be observed. Which severely limits the amount of time. Anyhow, the example is good enough.

 

The example fails on two levels. 1) it is a copy of existing function and therefore not new. 2) Stepping off and back onto a target, which is something completely within the capability of a pure random walk is not outside what would be predicted from information entropy and information theory. The balance of your argument is irrelevant .

 

 

And what other than humans have been observed to design? Are you saying maybe an animal designed us? Or maybe that something that has never been observed is needed to explain something you claim has never been observed?

 

Generation of functional information by intentional design is observed daily. The question being asked is what other than design is known to be capable of generating functional information? I am not speculating about who or what designed life.

 

Would it not be simpler to just cut out the middleman?

 

No it is not simpler or more correct to suggest that because we might not know what caused a particular design activity we should instead assume that processes that have not been shown to generate functional information at sufficient rates, somehow achieved what appears to be unachievable.

 

A better approach would be to find a process that does produce the observed results, and thus the purpose of the question.

Edited September 4, 2010 by cypress

[Mr Skeptic]

By your measure, plagiarism would be a valid method of generating a new research paper or report.

 

You agree then with what I said above? Since your response seems irrelevant to my point, there's no need for me to reply.

 

The example fails on two levels. 1) it is a copy of existing function and therefore not new. 2) Stepping off and back onto a target, which is something completely within the capability of a pure random walk is not outside what would be predicted from information entropy and information theory. The balance of your argument is irrelevant .

 

1) There is no existing function and therefore the function is new. Per your definition, the non-functional gene has zero information.

2) Good, then you agree that evolution is consistent with information theory. (Incidentally, there is no step off -- the algorithm is improving on a pattern, in this case of human-made design).

 

Generation of functional information by intentional design is observed daily. The question being asked is what other than design is known to be capable of generating functional information? I am not speculating about who or what designed life.

 

 

 

No it is not simpler or more correct to suggest that because we might not know what caused a particular design activity we should instead assume that processes that have not been shown to generate functional information at sufficient rates, somehow achieved what appears to be unachievable.

 

A better approach would be to find a process that does produce the observed results, and thus the purpose of the question.

 

You are glossing over the point: if as you say only humans have been observed to create design, it is a leap of faith to think that anything other than a human could, deity or no. But as I have shown, an algorithm can create design as well, so algorithms like evolution are also a valid source for design.

[cypress]

You agree then with what I said above?

 

No.

 

1) There is no existing function and therefore the function is new. Per your definition, the non-functional gene has zero information.

2) Good, then you agree that evolution is consistent with information theory. (Incidentally, there is no step off -- the algorithm is improving on a pattern, in this case of human-made design).

 

Other lineage's of the original persist with the previous function. The revived function is a copy of existing function with no net increase in information just as copying a research paper adds no new information.

 

I do agree that what we observe of evolution by mutation and selection seems consistent with information theory. We observe that mutation and selection on average does not generate information faster than that predicted by information theory and thus many orders of magnitude slower than predicted by the geologic record. Therefore based on observation we can say that so far mutation and selection does not generate functional information fast enough. Only a mind has been shown capable of generating functional information at rates consistent with and greater than that required to explain observed diversity.

 

You are glossing over the point: if as you say only humans have been observed to create design, it is a leap of faith to think that anything other than a human could, deity or no.

 

Nonsense, a bird's nest is designed as is a beaver dam.

 

But as I have shown, an algorithm can create design as well, so algorithms like evolution are also a valid source for design.

 

The algorithms we have discussed are tools of a mind, and require active information provided by the designer in order to functions. The designer's mind is responsible for the accomplishments of the designer's tool the tool was just the medium.

[Mr Skeptic]

Other lineage's of the original persist with the previous function. The revived function is a copy of existing function with no net increase in information just as copying a research paper adds no new information.

 

Ah, so algorithms are omniscient, and will only do as described if the information exists elsewhere? But if algorithms are omniscient, how can you question that they can create design? What a tangled web you weave.

 

In my opinion, algorithms are not omniscient and anything outside the algorithm is irrelevant to it.

 

Nonsense, a bird's nest is designed as is a beaver dam.

 

Ah, so animals can do it too. So, people and animals are possibly our Creator, or even algorithms -- but not anything else, since nothing else has been observed to make design.

 

The algorithms we have discussed are tools of a mind, and require active information provided by the designer in order to functions. The designer's mind is responsible for the accomplishments of the designer's tool the tool was just the medium.

 

Back to the omniscient algorithms I see. I don't know any algorithm that behaves differently depending on how that algorithm came to be, nor any mechanism by which it could do so. Algorithms work regardless, so any designer for them is irrelevant. Also, there's nothing to suggest evolution was designed.

[cypress]

Ah, so algorithms are omniscient, and will only do as described if the information exists elsewhere? But if algorithms are omniscient, how can you question that they can create design? What a tangled web you weave.

 

That's a strange direction to take. I have no idea how you conclude that I believe algorithms are omniscient.

 

 

Ah, so animals can do it too. So, people and animals are possibly our Creator, or even algorithms -- but not anything else, since nothing else has been observed to make design.

 

Another odd conclusion you have ascribed to me.

 

Back to the omniscient algorithms I see.

 

No. But I am thoroughly enjoying your attempt at non-sequiturs. You have had ample time to offer a natural process that can be observed generating functional information at rates greater than what information theory predict of random walks and blind search and as I suspected you offered none.

[pioneer]

Say we have an equation, which we can also plot as a graph. We can use a random walk on the graph paper to generate that same locus of points (useful information). Or we can use the equation to plot the locus of points, directly, generating the useful information at a much faster rate. The point is, if there is cause and effect (equation implies this specific locus of points), random walking can never keep up with the rate of useful information generation.

 

If life developed and evolved with considerable cause and effect, useful informational generation can out pace a random walk any day. Therefore we need cause and effect sprinkled in with random at a given proportion.

Edited September 5, 2010 by pioneer

[Mr Skeptic]

That's a strange direction to take. I have no idea how you conclude that I believe algorithms are omniscient.

 

Well, I described a situation where within the confines of the experiment there is no protein with a certain important function, and the function evolves. So it would seem that the algorithm generated information. You objected on the grounds that, in my example, the same information existed outside of the experiment and so no information was created. But, that is entirely irrelevant, unless the algorithm keeps track of all information everywhere, which would make it omniscient. So either your objection is irrelevant, or you are claiming the algorithm is omniscient.

 

Or maybe it is a matter of scale. I'm saying that the algorithm created this new function within the confines of the experiment where it didn't exist before, not that the function itself is somehow new. It is irrelevant whether the function exists elsewhere since it does not take part in the experiment. While I agree that in this example, the universe's content of information need not increase, that really says nothing about the algorithm's capability to generate information since the algorithm does not have as part of it the universe's information, only that within the experiment.

[cypress]

Well, I described a situation where within the confines of the experiment there is no protein with a certain important function, and the function evolves. So it would seem that the algorithm generated information. You objected on the grounds that, in my example, the same information existed outside of the experiment and so no information was created.

 

Here is how you described it: "make a point mutation such that it disables an important but non-vital gene. Evolution will rather quickly revert it back to the functional state."

 

You offered mutation of genes in bacteria and noted that this back and forth occurs regularly. I agree it does but since the initial starting point was functional and it still exists the reverted gene does not represent new function and this back and forth falls inside of the constraints of information entropy of a closed system. If we return to the original scenario described by the words quoted, we have an initial condition that was deliberately chosen by the experiment designer to succeed. The experiment set-up made use of a gene configuration that they knew contained a path that would result in a functional gene that a random walk could navigate given the resources available. The designer of the experiment provided the information required to succeed.

 

Or maybe it is a matter of scale. I'm saying that the algorithm created this new function within the confines of the experiment where it didn't exist before, not that the function itself is somehow new. It is irrelevant whether the function exists elsewhere since it does not take part in the experiment. While I agree that in this example, the universe's content of information need not increase, that really says nothing about the algorithm's capability to generate information since the algorithm does not have as part of it the universe's information, only that within the experiment.

 

When we isolate our investigation to the confines of the experiment, then we must note that the experiment designer with foreknowledge, inserted active information into the experiment at the beginning, by selecting a configuration with a known pathway navigable by the algorithm's fitness function. As with all the previous examples, this example succeeds by virtue of a designer. To see more clearly how significant and information rich the set-up was, let's take the same functional gene and then randomly shuffle the base pairs. For a 450 base pair gene, the number of configurations with a starting point similar to the one chosen by your designer is about 1 in 10^70 based on Douglas Axe's work with functional protein configurations. We could set this experiment up with no active information once a second for the entire life of the universe and still have almost no chance (< 1 in 10^30) of any of these experiments having succeeded in generating any useful function by mutation and selection.

[pioneer]

You guys are working under the assumption of no cause and effect due to a bias in biology.

 

If we look at the DNA and random changes for new information, the odds are we should be generating bad information at a higher rate than good information, since there are more things we can randomly do to mess up a cell than to advance a cell. Say we have 10,000 functional proteins, each with 1000 animo acids. If we randomly change the animo acids, the odds are we should be messing up the functional proteins faster than we are adding improvements. There are more sour spots than sweet spots, especially in you consider protein trains. If we assume random, evolution should not occur and life should constantly abort faster than progressing. Even a functional change in one protein can mess up a train if the new functionality does not coordinate. But thanks to cause and effect, the sour spots are ignored at a higher than random rate, so good information is favored. Algorithms are fabricated based on assumptions that overlook the slow rate of bad information.

 

Turn the discussion to a random walk for bad information. One will see there are problems with fundamental assumptions compared to observation.

[Mr Skeptic]

Here is how you described it: "make a point mutation such that it disables an important but non-vital gene. Evolution will rather quickly revert it back to the functional state."

 

You offered mutation of genes in bacteria and noted that this back and forth occurs regularly. I agree it does but since the initial starting point was functional and it still exists the reverted gene does not represent new function and this back and forth falls inside of the constraints of information entropy of a closed system. If we return to the original scenario described by the words quoted, we have an initial condition that was deliberately chosen by the experiment designer to succeed. The experiment set-up made use of a gene configuration that they knew contained a path that would result in a functional gene that a random walk could navigate given the resources available. The designer of the experiment provided the information required to succeed.

 

Yes, in this example, the target information exists elsewhere. That is irrelevant to the problem. If you prefer, a similar experiment can be done, with a minor change: make additional mutations to the DNA that are not relevant to the protein's function, as well as the single point mutation that will make it non-functional. In this case, the new information will not exist anywhere, yet evolution will find the functional version just as easily.

 

Yes, this is an experiment set up to succeed (as are most experiments); it need not even be done to know that it will almost inevitably succeed. However, the functional information is not included in the experiment -- it evolves from non-functional DNA. This cannot be ascribed to the experimenter, since the information is not there and just evolves. The role of the experimenter in this experiment is nothing more than to show that this can happen, so it will happen almost instantly to the shame of all who say it cannot be done.

 

With more steps away it will take longer, and with many steps away (especially if left slightly functional) it is likely to find another local maxima instead. Now tell me, was the information there or was it not? As I have repeatedly told you, whether information is functional or not has no relevance on whether it is information. Do you now agree? You seem to be saying that the non-functional information has the same information content as the functional one.

 

When we isolate our investigation to the confines of the experiment, then we must note that the experiment designer with foreknowledge, inserted active information into the experiment at the beginning, by selecting a configuration with a known pathway navigable by the algorithm's fitness function. As with all the previous examples, this example succeeds by virtue of a designer. To see more clearly how significant and information rich the set-up was, let's take the same functional gene and then randomly shuffle the base pairs. For a 450 base pair gene, the number of configurations with a starting point similar to the one chosen by your designer is about 1 in 10^70 based on Douglas Axe's work with functional protein configurations. We could set this experiment up with no active information once a second for the entire life of the universe and still have almost no chance (< 1 in 10^30) of any of these experiments having succeeded in generating any useful function by mutation and selection.

 

Irrelevant. Evolution does not deal with completely random proteins, so what you say has nothing to do with evolution.

[pioneer]

There is a big difference between useful and useless information. For example, there are trillions of units of information on the internet, but only a fraction of this will be useful information for any given search. If we wanted to know about dogs, we would not randomly walk the internet. A random walk would take a long time to get a proper hit; too much useless information relative to need. We would narrow down by typing dog. This will bring up 101,234 hits in 0.01 sec.. You then filter this to get the most useful information. This might be done by random walk.

 

If you have no goal in mind, but are just surfing the web to learn random things, a random walk can bring up useful information every step of the walk. This is more like the existing theory of evolution that has no goal in mind. But if we did this over the process of the day, the sum of all that information would not form a coherent whole. We may get some random overlap of coherency but it would be mostly little piles of dissociated information, even if it was useful information by itself.

Edited September 6, 2010 by pioneer