Talk:Irreducible complexity/Archive1

From RationalWiki
Jump to: navigation, search

This is an archive page, last updated 13 August 2014. Please do not make edits to this page.
Archives for this talk page:   , (new)(back)

[edit] Conclusions often follow from assumptions made

The article has this: "Imagine a situation where every gene has an equal chance of mutating within a certain period of time."

I have a problem with such an assumption, as I know other fields that such assumptions are often misleading (even when the assumption is valid). In signal processing, for example, it is often assumed that quantization error (using a uniform quantization scheme) can be modeled as Gaussian noise. However, in many cases, this assumption is invalid. For example, a 100 Hz sine wave sampled at 280 Hz and quantized to 8 bits will exhibit quantization errors that do not exhibit Gaussian distribution (in fact, the quantization errors will result in the appearance of "noise" that is highly colored, that is, correlated to the signal of interest and the sample rate). Even so, if you consider the effect of quantization for an ergodic ensemble of signals, the statistics of quantization error approach a Gaussian distribution as the number of signals in the ensemble increases. HeartGold tx 09:48, 25 June 2007 (CDT)

Smart people ROCK! User:PalMD
Well, I don't mean to show off. But in all cases, the origin and applicability (and dangers that lurk in applying) such assumptions are often glossed over. HeartGold tx 09:55, 25 June 2007 (CDT)

Doesn't matter in this case, you'd still get a probability curve if there was another kind of spread and with 3 billion DNA base pairs quantization isn't really an issue.

If you insist I can change this part, but it doesn't seem necessary, though I will change "genes" to base pairs, that really do have equal chances for mutation. MiddleMan 09:57, 25 June 2007 (CDT)

I'll think about how to reword this. I am not concerned about quantization in particular but rather how conclusions follow from assumptions, and whether or not danger lurks in those assumptions. HeartGold tx 10:02, 25 June 2007 (CDT)

Well, basically a copy error during cell division is random, so it doesn't target specific base pairs, neutron radiation just hit the cell and damages a chunk of DNA, also randomly, it could be that chemical contaminations do have a preference for one kind of base pair, but there are only 2 kinds of base pairs if I remember correctly. Maybe Palmd, or TimS can clear this up. MiddleMan 10:11, 25 June 2007 (CDT)

Well, there are many different kinds of errors, most of which are random, or "nearly random" (some areas may be more susceptible than other). There are 4 nucleotide bases (well, 5, but that's just a technicality), and any can be hit by radiation or other mutations. Various copy errors, like frame-shift mutations, occur as well as deletion. Most mutations are nonsense, but some actually encode proteins. --PalMD-yada yada 11:49, 25 June 2007 (CDT)

4 nucleotides and 2 base pairs, right? In any case, even if one kind of base pairs was more susceptible to mutations than the other, you could still graph it into a probability curve, notice g(x) part in the exponential, it leaves you free to change the steepness and such of the curve, but in the end you still have an exponential function that only approaches zero as g(x) approaches minus infininity. MiddleMan 11:59, 25 June 2007 (CDT)

Good point. I have to rethink the Diehard battery of tests. Philosophically, you cannot really prove that a source is not random even if it repeats the same sequence over and over (or makes errors in a predictable way). Even so, statitics are used to analize random number generators, and used to demonstrate that they are not random. Strictly speaking, such tests can only demonstrate that they probably are not random, but nobody would take somebody seriously who claimed that a coin that came up head 99.9% of the time in experimentation was a fair coin. You might argue that it is random, just biased. Applying this back to irreducable complexity, it seems that scientific conclusions can be drawn regarding the liklihood of compound (requiring two or more simultaneous mutations) adaptations occuring randomly. For example, and I am not arguing that this has been done, if it was never determined that the probability of a compound useful adaptation was greater than 1E-57, it would be unreasonable that all such adaptations were the result of truly random and normally distributed processes, even if filtered via selection. HeartGold tx 23:40, 29 June 2007 (CDT)

[edit] One Word

I have a tiny problem with one word "the guiding role of Natural Selection, " - this suggests intent and therefore reasoning and therefore intelligence = ID! QED. Someone think of another word - please. Keepit under your hat! 08:48, 2 August 2007 (CDT)

  • "Guiding2 doesn't have to be intelligent. The course of a river is guided by the lay of the land and the hardness of the rock it cuts through. But yes, guided for natural events should be avoided in this argument, otherwise the creationists will leap on it with a cry of "Aha!" (because they like Norwegian pop groups). Totnesmartin 08:56, 2 August 2007 (CDT)
Yes, it's a bad choice. I've removed it for now as it's only a hostage to fortune. If anybody has a better adjective then let's use it.

Well that's 3 UKers as one mind. (why aren't you two working - who's going to keep my pension coming? Mush, Mush!) Keepit under your hat! 09:17, 2 August 2007 (CDT)

Alright - you're in Spain but you could pay some taxes on your vast properties back home Bob) Keepit under your hat!

Actually off work for the month of August. :-) But it's a bit too hot to go out at the moment. (And my tax position is up to date in both countries.) :-) --Bob_M (talk) 09:22, 2 August 2007 (CDT)
sorry, off sick. Totnesmartin 09:28, 2 August 2007 (CDT)
Hope you get better soon. Out of curiosity, where do you stand on this UKers thing? I'm a Brit - what are you? --Bob_M (talk) 09:29, 2 August 2007 (CDT)
Brit is better, UKer seems like I'm playing a Ukelele. But then, George Formby was a Brit! Getting better soon doesn't appear to be on the table, but I can manage. Totnesmartin 10:05, 2 August 2007 (CDT)

It always seems to sound a bit derogatory to me - but OK I'll go with it.Keepgerbils(in a cage)! 10:13, 2 August 2007 (CDT)

Your misfortune seems to be RWs gain. Hope your ailment departs soon. Keepgerbils(in a cage)! 10:13, 2 August 2007 (CDT)

Thanks. What about "Airstrip one prole"? Totnesmartin 10:17, 2 August 2007 (CDT)
The reason that I don't like "UKer" is I'm ex-pat, and explaining what on earth "British" is takes long enough without having to talk about the UK. Naturally I could say that I'm English - but it wouldn't be true. "Airstrip oneer" - might work here. --Bob_M (talk) 10:20, 2 August 2007 (CDT)
Personal tools
Namespaces

Variants
Actions
Navigation
Community
Tools
support