Is atheism a scientific perspective?

[Angrboda]

I appreciate the fact that you are actually evaluating the argument as opposed to name calling. 

Yes, I am aware that there are problems with the definition of specified. It is incredibly difficult to quantify information in the way that ID opponents demand. What I think Dembski is right about is that the quantity of information of a given sequence is directly proportional to the functionality that arises from said sequence. This is coupled with the fact that gene sequence exhibit an extremely high degree of functionality. While we don't know how to quantify it, this does not mean that we can't draw conclusions based on the qualitative features. I think he was premature in his attempt to draw probabilistic conclusions about these sequences when we don't know how much information we are talking about. However, I think the central argument that DNA contains information and that that information leads to specific functions is true.

Your objections are a part of a larger argument that I have developed. I would instead of dealing with them ex parte, include the whole.

---

Contra Design -- Against the argument for design from biology

For many, the question of design is as simple as Justice Stewart's observations on obscenity, to wit, "I'll know it when I see it." They start from the presumption that certain things look designed and go directly to "was designed" (Do not pass Go, do not collect $200). But it takes a little more than that to make an actual argument. There has to be something connecting the premise that "It looks designed," to the conclusion, "Therefore it is designed."

Schematically, it goes something like this:

P1) It looks designed;
P2) . . . .
P3)
P4)
C1) Therefore it was designed.

Now a first gander at P2, etc. is to suggest the following:

P1) It looks designed;
P2) If it looks designed, then it was designed;
C1) Therefore it was designed.

Unfortunately we know that P2 is not true. There are things which look designed that weren't designed and vice versa.

So we try a different tack:

P1) It looks designed;
P2) Things that look like an intelligence designed them, are designed;
P3) It looks like a thing an intelligence designed;
C1) Therefore it was designed.

But the key question here is what does it mean to say that it looks like a thing an intelligence designed? This is entirely too vague to be of use when debating whether something like the DNA in a cell was designed, wherein the target is clearly removed from any direct traces of a designer. And we still have the problem of false positives; we can't infer design if our argument is only 'sometimes' right.

So we attempt to narrow in on what it means for something to look like it was designed by an intelligence. Perhaps:

P1) It looks designed;
P2) Things that look like an intelligence designed them, are designed;
P3) It looks like a thing an intelligence designed if it is similar to the way humans design things;
P4) It is similar to that;
C1) Therefore it was designed.

This brings a little focus to the question, but again it's rather vague. We have two problems. One, it's not specified in what ways the item must be similar to count for a design inference; obviously the color of an object is irrelevant. The other problem is that for compositions as complex as a cell, we don't have similar things from human designers -- we're not that intelligent, so it leaves open the question of what we mean by similar if there are no similarly complex works of human design. As Hume remarks on the relevant rule of analogy, "wherever you depart in the least, from the similarity of the cases, you diminish proportionably the evidence; and may at last bring it to a very weak analogy, which is confessedly liable to error and uncertainty" (Hume, Dialogues, Part II). What aspects of human design are we comparing to a cell?

This is where complex specified information, ala Dembski comes in:

P1) It looks designed;
P2) It looks designed because it has CSI;
P3) Things that have CSI, are designed;
C1) Therefore it was designed.

Unfortunately, as you seem to be admitting, and I'm claiming, Dembski jumped the gun in terms of a rigorous, usable definition for CSI. So what happens if we adopt your language that "The specified part is indicating that the information is used to accomplish a desired function." So being specified alludes to the item having specific functional significance. Here's the problem with that. Consider a bird's wing. Its function is to allow the bird to fly. It's not information, supposedly that's in the DNA for the creation of the wing, but ceterus paribus, the cases are parallel. Whether you agree with evolution or not, it is the case that we have mapped out how it is possible for this function to have arisen naturally. Function isn't specific only to designed systems. As I said before, function is in the eye of the beholder. If there is a possibility that the function of the wing arose naturally, then obviously function cannot be used to split the baby. For if it is even possible that specified information can arise naturally, it's no longer a flag for design. Now you may think the situation is different with abiogenesis, but it's not. All that has to be shown is that a possible sequence from a simpler organism without that function could lead to that more developed organism, all the way back to the first cell and beyond. (Not directly relevant, but think of the bacterial flagellum and the Type III secretory system.) We don't have to show probability or even have a complete map of the process to conclude from the evidence of the past 100 years that abiogenesis is a significant possibility.

So, in a nutshell, talk of "used to accomplish a desired function" doesn't work as function can be attributed to intelligent and natural causes. It's not a divider.