Marlowe

Given the new discoveries of complexity theory, there is another option besides random chance vs. an intelligent agent. Complexity shows that the properties of the basic structure of the physical universe naturally give rise to new levels of complexity and emergent structures and behavior - including planets, chemical soups, autocatalytic sets and life. See Waldrop's Complexity: the emerging science at the edge of order and chaos and Stuart Kauffman's At Home in the Universe. The mathematics of complexity, which are admittedly still new, appear to dissolve the problems with the random chance option - ie that there has not been enough time for the complexity of physical life to occur in this universe without an outside agent.

Thomas Metcalf

Originally posted by Tercel :

I was hoping to get your response to this post. Let me try to alter my argument to accomodate your version of FTA.

Good so far.

I think my argument would step in here. We have absolutely no way, as I see it, to figure out what sort of universe a god would want. It makes sense to think it would want some specific results, but I really don't think we can say much about what the intended results would be.

So my position is that A' and B' are on equal ground in:

-What a priori probabilities do we assign to the possibilities of God creating a universe (A') vs God* doing so(B')?

Now to recast your last two questions:

-Given that God was creating a a universe, what is the probability of F* occurring (C')?
-Given that God* was creating a universe, what is the probability of F* occurring (D')?

It was demonstrated in the first post that D' is significantly greater than C', even moreso than D is greater than C. Therefore, I think the real question is evaluating A' versus B'. And I confess I really don't see any justification for rating them anything other than equal. God* might be a particularly strange sort of being, or even not intelligent at all. Our probabilities are wide open, as far as I can see.

I'm looking forward to your response.

Thomas Metcalf

Originally posted by the_cave :

God* is kind of defined that way. So God* makes a better explanation than God.

My position is that the purpose of the universe is obviously for {F*} to be true, not {F}. (Because {F*} are less likely than {F}.) So anything that contributes to that fact (say, life being possible) is maybe just as important, but the point of the universe is to guarantee the truth of {F*}.

I don't think the second disjunct follows. God didn't finetune the universe for those consequences themselves, just for physical life. So they're not as important; they're kind of irrelevant. And further, there seem to be plenty of facts that don't have anything to do with whether {F} are true. God* explains these further facts, whereas God only explains {F} and the ones that are necessary for {F}.

Good analogy, but I think you have to say that if I didn't know anything about humans or houses, I'd be just as justified in supposing you cared about sawdust as I did in supposing you cared about right angles and planes. And in addition, I'd also reason that you cared about the pattern of knotholes in the wood you used, because that improbable result is better explained with the supposition that you do indeed care about them.

Kenny

I think that the matter turns on the a-priori probabilities of God versus God* existing.

Remember that in terms of the classic fine tuning argument, ‘God’ is not being defined in terms of a specific set of physical constants or large number of independent goals, but in terms of a particular goal, namely God is being defined as a being with sufficient power to create universes relevantly similar to our own (i.e. with a similar structure of physical laws) and as one who desires to create a universe relevantly similar to our own that is capable of sustaining intelligent biological life. Now as it so happens, goes the fine tuning argument, it turns out the probability of a universe relevantly similar to our own that is capable of sustaining biological life is extremely improbable on the hypothesis that God (as defined above) does not exist and (of course) extremely probable on the hypothesis that God does exist. Furthermore, the antecedent probability of God, as so defined, is sufficiently high (though, it need not be very high) so that the hypothesis that God exists gets confirmed on the fact that the universe is capable of sustaining intelligent biological life. Is it so unreasonable to suppose that there is an high antecedent likelihood that an intelligent being capable of creating universes might be interested in producing a universe in which other intelligent beings arise? Well, I don’t think so.

Now, given the very principles you outlined above, the more independent predicates that we add to our definition of God, the less the antecedent probability of God existing becomes. We could imagine several sub-hypotheses in which additional predicates are tacked on to ‘God’ as defined above, with each of these predicates being consistent with observed facts, and delineate a set of ‘gods’ (god1, god2, god3…) which constitute the set of these sub-hypotheses. The confirmation given to the hypothesis that God exists then gets distributed among each of these sub-hypothesis, but with the confirmation being significantly less the more predicates are added, until the confirmation becomes negligible.

Now, with respect to God*, God* is being defined, not with respect to one goal, but with respect to many largely independent goals. This means that God* is being defined by at least as many independent predicates as the set of facts God* is being defined to explain (contrary to God who is being defined in terms of a single goal and then just so happens to explain a large number of other facts about the world). This, in turn means, that the antecedent probability of God* existing is incredibly low – too low, in fact, to be confirmed on the available evidence.

God Bless,
Kenny

mattbballman

metcaf

Interesting Argument. Unfortunately, it depends on a notion of probability that no advocate of the fine-tuning argument would likely accept.

According to William Dembski, fine-tuning constants (f) established during the expansion of the big bang exhibit "specified complexity," - that is, they are (1) extremely improbable AND (2) they instantiate a meaningful pattern. The
pattern being their overall function in esablishing the conditions for the emergence of life. Events in set (f*) are highly improbable
but instantiate no such pattern. Thus, despite their immense improbability, we are not surprised that they occurred.

An analogy might be a poker game. A random hand is just as improbable as a royal flush - given equi-possibilities alone. But this leads to the counterintuitive conclusion that I should not be
surprised at the royal flush. The solution is to posit "specificity" as an addition criterion. What matters is the meaningful pattern relfected in the organization of the royal flush hand.

The poker example shows that improbability is not in itself sufficient to infer design. "Specification" is also needed (a concept well known to information theorists).

Thus, for supporters of the design argument, since fine-tuning does have specificity, the design inference still holds.

matt

Thomas Metcalf

Originally posted by mattbballman :

Well, no easily-describable pattern, but a specific pattern nonetheless. There is a particular set of facts {F*}, and the pattern is that every result happens to work toward the obtainment of the facts in {F*}.

I don't think any of these analogies work because they're presupposing someone in a position to change things, and this someone caring about the results. Royal flushes matter to poker players, surely, but if a group of dogs was distributing playing cards, and one of them happened to end up with five that composed a royal flush, I don't think humans would be quick to decide one of the dogs knew what a royal flush is and decided to cheat.

God* loves a specific pattern of constants, the one that will give the result that every member of {F*} is true. This is meaningful to God*, the way "physical life" is meaningful to God.

Thomas Metcalf

Originally posted by Kenny :

I think so, too.

I think this is the only response that is at all promising. I'd say we need not even define God* based on all these independent conjunct-facts; we can just say He's the sort of being that cares about everything except constants that have something to do with permitting the existence of physical life. God* doesn't seem much more complex than God, if this is the case. God* just has a specific pattern of constants in mind because of its aesthetic value to him.

See, I'm thinking God*'s goals aren't independent. God* has a specific pattern of constants that he likes very much, and so everything he does is to bring about this pattern of constants. The fact that physical life is possible is just an interesting consequence, although not interesting to God*.

tronvillain

At this point, someone usually drags out Bayes' Theorem. I will simply point out that the prior probability of God (as usually defined in the fine tuning argument for the existence of God) must be vanishingly small. After all, in the absence of any evidence there are a virtually infinite number of hypotheses, and "God" occupies only a fraction of them. Now, it is true that the "God" hypothesis will be confirmed by fine tuning, but it is far from the only hypothesis which is confirmed. There are so many alternatives that the confimation is marginal at best. Besides, the whole argument depends on the possibility of prior probability, and it is not clear that there was any such thing.

Theli

I just did that, but it's a point never taken by creationists.
:boohoo:

Tercel

Yes and no. We are very familar with the process of intelligent design. The design paradigm goes as follows:

-Analysis. Considerations are made about what is trying to be acheived. Goals are established. eg "I want to build a computer system capable of doing X,Y,Z" or "I want to order the playing cards in ascending order by suit". In the FTA's case: In the analysis phase, a suitably intelligent designer will consider the question of "do I want my universe to be able to sustain life?" (If the designer is not that intelligent or does not consider the question, then he is put in the "random" bucket for the purposes of the argument) He might well have other purposes too (though what an intelligent being would want with an empty universe I have to wonder: Pretty rock patterns?), but it is assumed he at least considers the question in his goal analysis. Now knowing nothing about the designer save that he is intelligent and considers the question, what is the probability that he will select a universe that can sustain life over one which cannot? I would suggest: Fairly high eg 0.9. But it hardly matters: If we insist that the designers purposes are entirely unknown then 0.5 is reasonable to represent the two equally unknown possibilities.

-Design. The designer then works out how to implement the design decisions. The designer searches amongst "all possible implementations" for an implementation that would meet the design requirements. In the FTA's case this would mean searching the possible worlds for one capable of sustaining life.

-Implementation. And then the selected implementation is actioned.


Looking at this paradigm we can see how, given an implementation we can decided whether it was designed or not. There are two effective probability distributions being dealt with. One even distribution where every possible implementation has an equally very low "chance" probability. One fluctuating distribution depending on each implementation's appeal as a solution to various design questions for an intelligent being.
eg if we walk into a room and find a deck of cards ordered ascendingly, we instantly assume that they were deliberately sorted by an intelligent agent rather than shuffled into that order randomly since that sorting is likely to be of disproportionate "interest" to an intelligent agent.

In the FTA, I have said above that the group of possible world that allow life and the group that don't are of equal interest to an "unknown purposes" being - ie probability 0.5 each. Yet the FTA asserts that the number of possible worlds which don't support life is significantly greater than the number that do. Hence the probability distribution for random chance is significantly lower and we see that the worlds that do support life would be of disproportionate interest to an intelligent designer. Hence the observation of a life-supporting universe provides confirmation of the design hypothesis.


I don't see your point. F consists of a small number of facts about the world all summed up in the one singular design decision that the world be able to sustain life hence the chance that a god would want F is reasonable. There are an almost infinite number of arbitrary and unrelated facts in F* (you are talking about a bizillion design decisions made exactly one way instead of just one) which makes the probability of a God* who wants exactly F* to obtain is vanishingly small.