Monday, November 19, 2007

The God Hypothesis

I'd like to make a few comments about my recent post, Intelligent Planting. I left the narrative without any comments there, as I figured this was an allegory that worked well enough on its own. I set up a parody of Intelligent Design to better illustrate all the leaps in logic design proponents expect people to make. It starts with jumping from "it doesn't look random" to "it was designed." Then it jumps from "it was designed" to "God/Pete designed it, and he also did all these other things recorded in the Bible." Now, of course, design proponents are all about hiding their religious affiliation, but it's there, and it is their ultimate goal, whether they'll admit it or not. I also then threw in some of the other doggerel they use to justify this for good measure, particularly mocking the appeal to faith.

That all being said, you might be somewhat surprised to learned that Intelligent Design wasn't my initial impetus for writing this story. Instead, this comes from a different argument for God which makes much the same leap in logic (from saying there was someone or something to saying it was God). This was what's known as the Cosmological Argument. When boiled down, it essentially becomes, "There was an ultimate cause for everything, therefore God."

You might want to take a moment to read through the Wikipedia article on this argument, linked above. What I'd like to call to your attention is the simple fact of how many variations on this argument there are. The Cosmological Argument is often presented as being strictly logical, but if that were so, then you wouldn't expect these variations on exactly how the Prime Mover/Uncaused Cause/God started things. Therefore, it would seem that most or all of these arguments are likely making some assumptions behind the scenes (or are just fallacious).

For example, let's take the argument of Thomas Aquinas, one of the more complete versions. From the Wikipedia summary:

1. Every finite and contingent being has a cause.
2. Nothing finite and dependent (contingent) can cause itself.
3. A causal chain cannot be of infinite length.
4. Therefore, there must be a first cause; or, there must be something that is not an effect.

Few would argue with points 1 and 2, but let's take a look at point 3. Why is it that a causal chain cannot be of infinite length? Presumably this is simply stated because the thought of it seems absurd, but is it really? Let's extend things into the future. Under most modern models for the universe and many religious models as well, time will go on infinitely into the future. This means that as long as it keeps going, we'll keep on having a causal chain. Thus, the causal chain will extend infinitely into the future. Ipso, a causal chain of infinite length. (Man, have I been itching to properly use "ipso" in a sentence...)

So, if it can extend infinitely into the future, what's wrong with having a causal chain extend infinitely into the past? It's at this point that it seems a bit more absurd instinctively, but logically it doesn't have to be. All laws of physics we know of are time-reversible, with a single exception that allows us to see order in time, the collapse of a wavefunction. If you compensate for collapse and run things backwards in time, you can see the same theme of causation occurring. Instead of a sperm and egg causing a zygote, you get a zygote causing a sperm and an egg, for instance. Running things this way, it doesn't seem so absurd that things might go on forever.

So here we have the problem with this particular argument: a false premise. The argument may still be technically valid (the conclusion can't be false if all the premises are true), but with a false premise, it's unsound, and we have no reason to believe the conclusion given this argument. Now, this doesn't mean that there wasn't actually some first cause, it only means that this argument doesn't prove it. So, let's entertain the idea that there was a first cause now, for completeness' sake.

What can we say about this first cause? Well, nothing, really. We can't claim it must have been intelligent, or even complex in any way, as it's easily possible for intelligence and complexity to arise from unintelligent, simple conditions, driven by a little randomness. However, we have a lot of religious people pointing to their own god and saying it fits the bill of a first cause. The argument for a first cause, even if it were valid, doesn't give us any reason to believe that the first cause is anything like their god, but that's not necessarily a problem.

What we can do is treat their god as a hypothesis to explain the first cause. A tactic like this is often done by scientists; we have a problem, so we hypothesize something to explain it which is a bit beyond what we know. Since it's beyond what we know, it often comes with the ability to predict other phenomena we haven't tested for yet. So, we then go and test for those phenomena. If they exist, we have evidence that this hypothesis is true. We can do roughly the same thing for the hypothesis that a god was the first cause.

The immediate problem is that invoking a god here is a gross violation of Occam's Razor. An explanation that invokes a particular brings in many interrelated claims, and has many, many predictions beyond the simple creation of the universe. This doesn't mean it isn't true, however; it just means that we're going to need a lot of evidence to support it. Otherwise, a simpler explanation (or less precise god) will be much preferred.

Now, there are many deities we can choose from, so I'll only use a couple examples here, positioned at extremes. Most other deities will fall on the continuum somewhere between these, and a mix of the applicable arguments will apply.

First up, I'm going to take the god believed in by many evangelical Christians in the US. This is the god discussed in the Bible, who did all the things claimed there. He's omnipotent, omniscient, omnipresent, and just generally omni. He also takes a role in our day-to-day lives. He listens to and answers prayers. He smites those who displease him in any way. When people die, their souls are judged by him. If they've led a ridiculously devout life, free from even the slightest pleasure (shadenfreude over thinking about sinners going to hell excepted), they go into heaven, a place of eternal bliss. If they're even slightly off, or believe in a slightly different god, they go to hell, a place of eternal torment. (Aside: I don't particularly care if anyone believes in exactly this god; I'm just using it as an extreme example.)

This the hypothesis to be tested. On the other end, we'll have the null hypothesis, which we'll be comparing this to. At the end, we hope to be able to reject either this or the null hypothesis. In this case, we can use the null hypothesis, "No god exists." If we find sufficient evidence for this god, we'll be able to reject this null hypothesis.

So, what of this is testable? If you say "none of it," scroll down a bit. I've got your untestable god there. This is a god who interferes with the world. If there are natural effects of supernatural causes, they can be tested for. Anyways, there are two big points here that we can test: Intercessory prayer and smiting the heathens. Let's start with prayer. This is something that actually has been scientifically tested. Repeatedly. And then some more. And again, because every time, the results weren't satisfactory. Whey weren't they satisfactory? Because the tests were either poorly done, or they didn't show any effect to prayer. Even if you don't agree that the ones who showed an effect were poorly done, it's still only a marginal improvement. It's nothing compared to the effect you'd think an omnipotent god like this could have.

But wait! "Absence of evidence isn't evidence of absence!" I hear you cry. Ah, well there's a big exception to that. Absence of evidence can indeed be evidence of absence when you've properly and thoroughly looked for evidence. I call it the Modus Tollens Exception, as you can phrase it in the following logical form:

P1: If A exists and we use method M to search for evidence, we will find evidence E.
P2: We used method M to search for evidence, and did not find evidence E.
C: Therefore, A does not exist.

Boiling it down to the simple logic, this is the valid structure:

P1: If A and M, then E.
P2: M and not E.
C: Not A.

This is essentially the Modus Tollens argument form, with a small complication of an extra requirement in the premise. Since we're using a logical form here, my argument that absence of evidence is in this case evidence of absence is also valid.

It's a bit harder to test the prediction that this god will go around smiting heathens, as we can't really control things here. However, if you look at what happens in the world, there isn't good evidence that this takes place. For instance, many evangelicals claimed that Hurricane Katrina was their god smiting New Orleans for all the debauchery that takes place there. The problem was that the French Quarter, which was where most of the debauchery took place, was one of the least-damaged areas. This particular claim, at least, doesn't hold water. (Ugh... I swear that pun was unintentional.)

In any case, testing prayer alone is sufficient here. We can, of course, add to the evidence for all the other claims about this god, though it's not necessary for the time being. Since we have evidence that prayer to this god doesn't work, we can reject the hypothesis that this god exists. And no, we don't go and accept the null hypothesis that no god exists; we just say that we fail to reject it. It could well be true, but we haven't shown that here.

Alright, now let's switch to the other extreme. Let's take a Deist god. This god created the universe, and then just kind of sat back and watched. Or maybe he went off to create another universe, or just took the next few billion years off to slouch around and watch TV. Or maybe he's a "she," or an "it," or some other gender we don't have a pronoun for. Not much is claimed about this god. In fact, aside from that he created the universe, not anything is claimed about this god. In contrast to the previous case where so much was claimed about the god that it was easy to find evidence against him, here, we don't have anything claimed at all beyond what we know happened. With this, we can't make any testable predictions about him, so we can't scientifically test for his existence.

This is still a hypothesis, though. The problem is that it's a completely useless hypothesis. There's nothing we can do to improve upon it, or get any further evidence that it might be true. There isn't even anything we can do to differentiate it from similar hypotheses, such as saying that instead of a god, a Flying Spaghetti Monster created the universe. Or we could say that the universe spawned yesterday from primordial slood with all the particles in just the right positions and velocities for us to be here with all this evidence for a past and memories of it.

In the end, there's no way to reject this hypothesis. But coupled with that is the fact that there's no way to get evidence for it. There's no reason for you to actually believe in it. The fact that it can't be disproven is no reason to believe, as following that logic would lead you to believe a million contradictory explanations for the beginning of the universe. This is why science doesn't do anything with untestable hypotheses; they're utterly useless.

In the real world, most gods people believe in fall somewhere between these extremes. They try to balance out not contradicting reality with having enough predictions to be useful. However, this doesn't really solve any problems. In order for their to be evidence of a god's existence, it has to make some predictions that later turn out to be true (and of course, can't be adequately explained without him. A god predicting gravity isn't a big deal). Simply throwing away disproven predictions and holding onto untestable ones still doesn't give anyone a reason to believe in this god.

However, I could well be wrong. If I am, if there's some god out there with good evidence for his existence, I would in fact quite like to know about it. I'd expect to have heard of it by now, but you never know. Maybe the right study just hasn't been performed yet. In which case, I challenge any believer who believes they have a testable prediction about their god to go out and perform a study to test it. Perform it well enough, and a positive result could be just what you need to convince me. Until then, I'm happy living my life accepting the null hypothesis as most plausible.


Techskeptic said...

This is a long post, and I dont have time. Quick gripe though.

P1: If A exists and we use method M to search for evidence, we will find evidence E.
P2: We used method M to search for evidence, and did not find evidence E.
C: Therefore, A does not exist.

This is exactly the argument that god jockeys and politicians use to close SETI.

You are missing some term to define completeness of the search. And once you put that term in there, then it will be used to simply say, we simply havent searched long enough for evidence of god, woo, or whatever.

How do we know when SETI is done (I cant imaine that there is really a time) and when the search for an omniscient, omnipresent, being who can control every single electron in the universe but still not be directly or indirectly detected in a quantitative means?

Infophile said...

The thing about SETI is that its search can never really be complete. We don't really know when and where an alien race might evolve and start sending signals we could detect, so we have to keep going at it forever just in case. The difference when it comes to gods is that people make claims that are testable here and now. People actually believe that prayer works, so we can test that. When they start to water down their claims so science can touch them, they venture into "useless" territory.

Mark said...

Levitating super turtles!

Infophile said...

They aren't levitating, each of them is standing on the one below it.

Arcanum said...

First, let me say 'interesting post' and I do agree that gods are an artifice by which to explain that which cannot (easily) be explained.

Let's look at Aquinas again:
1. Every finite and contingent being has a cause.
2. Nothing finite and dependent (contingent) can cause itself.
3. A causal chain cannot be of infinite length.
4. Therefore, there must be a first cause; or, there must be something that is not an effect.

I agree with you that our Newtonian worldview suggests intuitively that we should accept 1and 2, and that 3 might not be acceptable.

However, there are some real problems with the concept of infinity: how, for example could we reconcile the fact that July 4, 1776 and September 11, 2001 are not simultaneous unless time were circular, coming back upon itself? Obviously, all the empirical evidence for a cosmic 'start' time, negates the possibility of infinity for time. Events impose the dimension of time -- pre-event = pre-time. Since the cosmic clock begins with the bang, then 7/4/1776 and 9/11/2001 do not generate a paradox.

I think that the problem with Aquinas' argument lies in the assumption that everything must have a cause (1 and 2). What causes the cause?

Even if you say that the universe is contingent and the first cause of the cosmos is necessary a la Leibniz, this is a mere supposition from a definition that relies on Newtonian intuitions for our glib acceptance. If this supposition were factual, then what brought the necessary first cause into being? Their attempt to provide a solution to this dilemma by concocting yet another definition, just doesn't convince.

Arcanum said...

P1: If A exists and we use method M to search for evidence, we will find evidence E.
P2: We used method M to search for evidence, and did not find evidence E.
C: Therefore, A does not exist.

Hmmmm ... the problem is that we cannot logically ever disprove the nonexistence of 'A'. I agree that there are no supernatural agencies (the supernatural is merely another of those so-convenient apologetic definitions). However, that 'C' is a real problem because failure to find E does not demonstrate that A does not exist, it merely renders A highly unlikely.

I think that the actual post-science situation is more like:

P1: If M (mechanism) exists and we use method S (science) to test for an explanation for M, we will find D (data) -- either G (god-data) or Q (other-data).
P2: We used method S (experiment) to test M, and consistently find data that fits Q predictions, but not G predictions.
P3: We have ample explanations for the prevalence of G hypotheses and for the emotional reasons for inventing such hypotheses.
C: Therefore, it is increasingly unlikely that the G explanation is accurate.

Infophile said...

However, there are some real problems with the concept of infinity: how, for example could we reconcile the fact that July 4, 1776 and September 11, 2001 are not simultaneous unless time were circular, coming back upon itself? Obviously, all the empirical evidence for a cosmic 'start' time, negates the possibility of infinity for time. Events impose the dimension of time -- pre-event = pre-time. Since the cosmic clock begins with the bang, then 7/4/1776 and 9/11/2001 do not generate a paradox.

I'm sorry, but I can't quite follow your line of argument there. What does a potential beginning to time have to do with whether or not we can measure it? Humans have only been around a small fraction of the time of the universe, and yet we can set our clocks at some point and then later agree that two events happened at different times. Just because we can't find an absolute time doesn't mean we can't measure intervals.

Arcanum said...

I wasn't clear - blame time and space constraints plus my inability to express it clearly.

Let's start at time 0 and move forward to 7/4/1776, next we move onward to 9/11/2001. There is no problem with working out the time interval from our starting point to either date -- different intervals to different dates.

Compare this with the situation with infinite time -- the time from infinity to 7/4/1776 is infinitely long, as is the time from infinity to 9/11/2001. The paradox lies in the fact that 7/4/1776 and 9/11/2001 do not occur at the same time.

There's a similar argument for the paradox associated with an infinite series of numbers.

If time circled back upon itself, this dicrepancy would be removed, but we'd have the problem that events should then repeat themselves.

However, we know that the universe began at a finite time over 14 bya, so point 3 of Aquinus' 'proof' is supported by empirical evidence.

I think that even though we cannot disprove a god, we can say that supernatural entities did not bring this about (on several lines of reasoning), so the problem with the argument must precede premise 3.

Infophile said...

Well, that's only really a problem if you want it to be. Instead of comparing to time of negative infinity, compare to the time of, say, the supposed birth of the supposed messiah of a major world religion. One date's 1776 years after it, the other 2001.

Also, the beginning of the universe doesn't mean the beginning of causality. Most scientific models leave wide open the possibility of there being something before this universe.

Aquinas Dad said...

I do not mean to sound rude about this, but one way of determining if a person understands an argument is if they can summarize it; unfortunately, you fail to summarize the cosmological argument properly.

"There was an ultimate cause for everything, therefore God."

Is not correct. The cosmological argument (BTW, it is stated so many different ways in part because the terminology of formal logic has changed over the centuries; also, the development of thought concerning the CA has been going on for centuries) is far too complex for a perusal of wikipedia to be enough - for example, St. Thomas Aquinas' original discussion covered both if the universe is of finite existence and if it is of infinite existence, meaning that he addressed your main point a loooooong time ago.

The main thrust of the Cosmological Argument is rather simple; the cause of the universe is, by definition, supernatural. That's it. St. Thomas extrapolated that this supernatural cause is God, but he did *not* present the CA as "proof" of God's existence.

As for the use of Occam's Razor, there are two issues; first, Occam's Razor is not a law of logic, it is a guideline for making initial decisions. Galileo, for example, lampooned the idea of Occam's Razor as a rule rather than an initial tool in his Dialogues. The second problem with your invocation of Occam's Razor is the work of Richard Swinburne - a champion of Occam;s Razor he actually developed the so-called Inductive Cosmological Argument that argues that Occam;s Razor, if seen as a *rule*, demands the acceptance of God as the least complex theory of existence!