|
Author
|
Topic: Joshua A. Smart: On the Application of Irreducible Complexity
|
RBH
Member
Member # 380
|
posted 08. June 2003 18:46
Mike,
Please read the paper. You're costing me a signficant amount of time to distill and summarize it, and I'm no less pressed for time than you are.
RBH
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 08. June 2003 21:16
RBH,
I understand issues of time. Since there are about a half dozen other people pushing the Lenski study as relevant to IC, perhaps they can address my questions concerning function.
Let me make some initial observations concerning the crucial issue of ‘parts.’ For years, I have made it clear that I use the concept of IC as it applies to molecular machines (a decision, BTW, that is not arbitrary). As such, I find there to be a good working relationship between Behe’s concept of a “part” and a gene product. But I’m afraid the EQU’s “parts” are not good analogs for gene products. Thus, its relevancy, from my perspective, is quite suspect.
You note that the EQU’s parts must have two features – the instruction itself and the positioning of the instruction: " 'q' is a "kind of part;" 'q as instruction #11' is a "part." It’s almost as if all the parts are two-part IC systems, where both the instruction and its positioning in the program are needed to declare it a part. Now, one of the features of a machine is that you can disassemble it into its component parts. I can go into the garage and disassemble a hair-dryer into all its component pieces and lay them out on the bench for all to see. Likewise with the molecular machines, biochemists can disassemble all the parts and run them out on gels for all to see. Perhaps this is why disassembly is called taking something apart . But this feature seems to be lacking in your EQU. For the moment we pry one “part” from the whole, it vanishes. The “parts” cannot exist apart from the whole. And this brings in the issue of modularity. Modularity is what makes cooption a plausible explanation for IC machines. But according to the definition of parts entailed by the EQUs, they cannot swap “parts” because the “part” only exists as part of the whole.
You also note: It's like a brick pillar made of a slew of bricks piled up one on top of another. All the bricks are identical - all are the same "kind of part" - but pulling out the brick on the bottom of the pillar has a considerably different effect than pulling out the brick on top. They are different "parts" in the context of the structure.
Thus, to make the EQU IC, we also have to envision a brick pillar as IC. That seems awfully controversial to me. If we consider the IC mousetrap, or flagellum, we don’t find all the “same kind of parts” where there are different effects depending on what gets knocked out. In the mousetrap, eliminating any one of the components causes of the same dramatic effect – complete loss of function. It’s quite unlike a brick pillar.
Perhaps this is why Yersinia prefers to map the “part” to the instruction alone. But then we’re back to the problem of IC, as many of the instructions can be either crucial or extraneous, depending on where the fit in the system. In the case of the flagellum, that’s like saying fliG plays an important role in motility, but it also interspersed in the filament where it plays no role. Anyway, there's an even more interesting perspective that follows from equating the instructions with parts.
Yersinia,
Ambiguity arises any time we try to map abstract concepts to physical reality. Take Behe’s mousetrap. Behe lists the staples in parentheses, but doesn’t include them as part of the list of IC components. Anyway, from my perspective, the ambiguity associated with applying the term “part” to the EQU is far more excessive than when dealing with a mousetrap or flagellum.
Since I don’t view the molecular machine “parts” as the individual proteins, I don't find your first point to be relevant. I do view the parts as gene products, thus your second point should be relevant. But again, I’m having a hard time seeing how. After all, there is all kind of fuzz associated with declaring similarities as a reflection of common descent (unless you are a non-teleologist). Thus, the fuzz you have in mind would not come into play until we eliminated the fuzz I have in mind. And it’s not as easy as you seem to think. However, I really don’t want to steer this thread into a lengthy tangent concerning the reasoning used to declare “gene duplication,” so let’s just deal with the fact, for example, that the alpha and beta subunits of the F-ATP synthase are quite similar. So what? They are still encoded by separate genes and both are needed for the synthase to function. In fact, this provides another example of the utility of IC. Back on ARN, in 2001, someone named “Wolf” cited a paper claiming it showed that a parrot had a functioning ATP synthease without the alpha subunit. Because of IC reasoning, I was immediately skeptical and decided to read the paper in question. It turned out that ‘Wolf” has misread the paper in that one of the alleles for the alpha subunit had been junked. But parrots, of course, are diploid and the other allele was as well conserved as any other alpha subunit. Would a Darwinian perspective have led to the same skepticism (that was validated)? Of course not.
As for the fact that pseudogene copies exist, again, so what? If we are to use the EQU as a valid analogy, this would mean pseudogene products would also be incorporated into something like the bacterial flagellum, as the “junk” is integrated into the EQU program, even to the point where RBH scores the junk as a part.
Finally, yes there is almost always more than one way to do something in biology. In fact, there is no need to tack on the “in biology” point, as there is usually more than one way to do anything. There is more than one way to get to work. There is more than one way to kill a mouse in the house. There is more than one way to make a point on these forums. I’m not sure why this is significant. Perhaps we can put some teeth to it. Are you implicitly acknowledging that when biology only uses one way to do something, that counts against a Darwinian origin?
Since like RBH, I am pressed for time, I probably won’t be able to get back to this until later in the week.
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 08. June 2003 21:34
Mike: Would a Darwinian perspective have led to the same skepticism (that was validated)? Of course not.
Begging the question Mike. Perhaps a better way to phrase it would "I doubt it". But similarly may I ask would an ID perspective have led to the research by Lenski et al? Should it?
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 08. June 2003 22:06
Mike,
This one is easy, so I'll give it 45 seconds. A mousetrap (or your disassembled hair dryer) is an IC system only when the pieces are assembled in a specific configuration relative to one another. (That may be what Behe meant by "well matched".) The pile of disassembled pieces is just a pile. The same holds for the assembly language programs that evolved: They comprise an IC system when the pieces (instructions) are assembled in a specific configuration, where order counts. Leave one out and just like the mousetrap without one of its parts, they're non-functional systems: they no longer perform the function of correctly evaluating the EQU logic operation.
The several instances of a given instruction are not identical in context: assembly of components in a specific configuration relative to other components really does count in all (allegedly) IC systems. That some of those parts are identical in form to one another does not mean that they can be lumped together and counted as one "part," since in the context of the functioning system they are all individually necessary in their individual places in the system.
RBH
[ 08. June 2003, 22:08: Message edited by: RBH ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 08. June 2003 22:28
Mike, maybe you could expand upon your comment about the non-arbitrary nature of restricting IC to molecular machines? I've been trying to understand this restriction based on the (seemingly irrelevant) complaints you've raised, but I've failed so far. What are the salient features of a molecular machine that make it the only suitable target for IC? After all, Behe uses a mousetrap as an example, and that is assuredly not a molecular machine in the normal sense of the phrase.
Also, while the analogy isn't perfect, I'm not particularly troubled by having "part" be "position plus function". Very many components of protein complexes are like this; there is a binding domain, which gives the part its position, and there is a catalytic or structural domain which gives the part its function. (E.g. flagellar ATPases, G proteins, kinase cascade components, etc..)
Envision your brick pillar as an arch, and you have a rough analogy with both the program and with molecular machines.
Again, the analogy isn't perfect, but what I'm failing to understand is why the imperfection apparently isn't viewed as relevant to irreducible complexity.
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 08. June 2003 23:16
Okay, I'll give it 30 seconds. ![[Smile]](smile.gif)
You say, " The pile of disassembled pieces is just a pile. The same holds for the assembly language programs that evolved.."
Not true. With the EQU, you have no pile of disassembled pieces because when you "disassemble" it, the parts vanish. This simply follows from the fact that you've defined the "part" as a function of a whole. Ask yourself this - do any of the 23 independently evolved EQU's share any parts?
Rex,
I understand what you are saying - I'm not trying to rescue IC as an infallible test of unevolvability. And I never said molecular machines are the only "suitable target" for IC.
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 08. June 2003 23:20
Mike wrote quote:
With the EQU, you have no pile of disassembled pieces because when you "disassemble" it, the parts vanish.
Huh? You mean the instructions don't exist if one lists them, say, alphabetically? Remember, "EQU" is not the structure that is being disassembled; it's the program that evaluates EQU that is composed of instructions. It's a list of instructions in a particular order. I don't follow you at all here.
RBH
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 08. June 2003 23:32
You said:
quote:
As to parts, in constructing your list of 14 you have conflated "part" and "kind of part." 'q' is a "kind of part;" 'q as instruction #11' is a "part."
I'm not sure what "kind of part" is supposed to be. But it sounds to me that q, apart from its position as #11 in the string, is not a part. If q, all by itself, is a part, fine. Again, do any of the 23 independently derived EQU's share parts?
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 08. June 2003 23:45
Mike asked quote:
Again, do any of the 23 independently derived EQU's share parts?
All 23 programs used the same set of 26 primitive instructions. Each used them in a different sequence.
RBH
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 09. June 2003 00:15
So the "primitive instructions" are the parts?
PS: Don't worry, I'm going to bed.
[ 09. June 2003, 00:19: Message edited by: Mike Gene ]
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 09. June 2003 00:59
Mike,
The program I posted on page 2 of this thread is 60 instructions long, and has 60 parts in the definition I deem appropriate: "instruction in sequential context." Of those, 35 form the irreducible core by the knockout criterion: replacing any one of those 35 with a null instruction eliminates the program's ability to correctly evaluate the EQU logic function. The other 22 programs were composed of varying numbers of parts, with varying numbers of parts forming their irreducible cores.
All the programs were evolved drawing on a common pool of 26 primitive instructions.
The name of a part when the program is 'dissasembled' includes the letter designator (a - z). Any two parts of the irreducible core that have the same letter designator may be swapped with no effect on the functioning of the program.
Genes are sequences of primitives: G's, T's, C's, and A's. They are also sequences of triplet base-pair codons. At least some genes are IC - whack a given base pair and the gene's function is disrupted or eliminated. Whack a codon and the gene's function is disrupted or eliminated. What are the "parts"?
When your hair dryer is disassembled, are the two (physically identical) bolts that secure the handle to the body one part or two?
RBH
[ 09. June 2003, 01:03: Message edited by: RBH ]
IP: Logged
|
|
Argon
Member
Member # 276
|
posted 09. June 2003 09:20
Considering parts...
I think that most, if not all of the derived EQU programs also perform other functions. Some of the code from these functions might serve not only as precursors but also as components of EQU.
FWIW.
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 09. June 2003 10:58
Argon,
Yup, that's the case. And in many instances, the knockout procedure that eliminates the ability to correctly evaluate EQU also eliminates some of those other functions. In a few instances knocking out EQU actually restores the ability to perform some other function.
RBH
IP: Logged
|
|
Argon
Member
Member # 276
|
posted 09. June 2003 14:19
So, above the level of individual machine instructions, one could make the case that the EQU functions were at least partially composed from higher-level, independently-derived sub-functions (sub-programs). Thus the enumeration of "parts" or "components" should include these organized structures. Might these be roughly analogous to "functional domains"?
[ 09. June 2003, 14:20: Message edited by: Argon ]
IP: Logged
|
|
Jerry D. Bauer
Member
Member # 756
|
posted 09. June 2003 21:24
Let's throw this out and see how long the duck can swim. I've taken Thomas' work, expanded it, honed out what I believe to be irrelevant tenets and added some math. I suppose we can call this the Thomas SIBO methodology:
1) The object must be an irreducibly complex system. An ICS is defined as a system that cannot be reduced beyond a primary core of multiple parts; those parts being system-dependent and/or system-specific and existing in a minimally complex state. This system must show a positive function in the formula: Ft = (FM – Fm) + .001, where Ft is total function, FM is maximum function defined as a system with all it’s component parts intact, and Fm is minimum function defined as the number of parts of a system that are irreducibly complex.
2) The ICS must demonstrate molecular punctuated equilibrium.
3) There must not be a historically appropriate continuum that demonstrates ‘sloppy simplicity’ as part of the organization of the system under study.
4) The ICS must demonstrate specificity and complexity specificity must be over 500 bits in the formula: Cs = ln(W), where Cs is complexity specification and W is the component parts of the system.
5) As change occurs in this system over time, CsT must not drop under 500 bits: CsT = Cs1 –/+ Cs2, where CsT is total complexity specification, Cs1 is the complexity specification before the system change and Cs2 is the complexity specification after the system has changed.
CONCLUSION: If the above tenets are met, inference of design may be utilized to describe the origin of the system under study.
IP: Logged
|
|
|
Printer-friendly view of this topic