The question is: who gets to choose my null hypothesis?
Chad asks whether there are "reasons" for being an atheist.
This is an axiomatically incorrect question: the bigger question is what is the null hypothesis?
Do we assume that there is a God and that she has some attributes, until evidence to the contrary is presented?
Or, do we assume that there is no such thing as a God, until some evidence to the effect is presented?
It would certainly seem simpler, ab initio, to start with the premise of No God, until and unless there is positive evidence to the contrary.
A counter-argument would be to note that the Bayesian prior is not ab initio reasoning, but that we are born and raised in deistic societies, and therefore the correct prior is to go with the flow and accept God because, well, since most everyone else does, did, or at least make vague obeisances in the general direction of a God like thingy.
Most people who default to deism are not in fact making reasoned arguments and wouldn't recognize an argument from Aquinas if it bit them.
Ah, but which God?
And there is one of the many rubs. One Dawkins nailed well.
Auðhumla? Óðinn? Athena? Mithras? Flying Spaghetti Monster?
God seems to be a bit of a contradiction. (Hm, I suppose that is a least consistent with the omnipotency bit, eh?)
One could retreat to the God of Spinoza, but, honestly, that was a total copout and Spinoza knew it.
So, which to choose?
Or is there no choice.
- Log in to post comments
Personally, I'm choosing the Flying Spaghetti Monster; no other God is that delicious.
Lately I've been thinking about ignosticism. I think it provides some sort answer to this question, or at the very least, skirts around it.
That's what I think this is about: a counter argument. Chad would like the PZ squadrons to give his blog's hit counter a spin.
I'd suggest that the correct prior distribution must incorporate the facts we have learned about the physical Universe — all the facts, as far as we have managed to discover them. A remarkable theme in this collection of discoveries is the idea that the overwhelming majority of causes do not resemble the actions of conscious, thinking beings. Therefore, when confronted with a new phenomenon, my null hypothesis should be that its cause is also not the action of a conscious, thinking being.
It has gone beyond ludicrous to imagine that beyond the fiery birth of the Universe eons ago, behind the physical laws whose unfamiliar character we can only appreciate through intense study, is a human face.
All religions in the Western tradition make claims about the material world. They tell us what the world is like, how it used to be, and what sort of stuff goes on within it. They are not free-floating ethical or philosophical systems; they have baggage. Like any other explanation for what, how or wherefore, they have to stick their head out and make statements if they're to have any merit whatsoever. When they do, they are quite frequently wrong. (Where in the Bible do we find a round Earth with America on it?) If a hypothesis cast in terms of electromagnetic fields and non-abelian gauge transformations performed this badly, we'd throw it aside without hesitation.
I fail to see why we should do any different for explanations which come to us through mangled translations of Bronze Age fables.
Neat post, Steinn. I've linked to you here (for some reason Haloscan won't recognize your trackback URL).
It might seem at least not totally ludicrous to consider what type of god or gods or god-like beings our current understanding of the universe implies, if one makes the generous and unwarranted assumption that any evidence implies the possible existence of anything like a god. If one started from that assumption (that the nature of a god-like being should be reflected in its creation, assuming the universe is its creation), then one could at least eliminate the biases of prior religious belief. I suspect such an exercise would not sit well with the conventionally religious.
Mark:
Scott Aaronson did an analysis of the kind you propose, in an essay entitled, "A Euclidean theater of misery". The conclusion:
Isn't someone who defaults to anything not making reasoned arguments?
Since representing "exists" takes less symbols in both English and logic than "not exists," I submit that "exists" is simpler and should therefore be the null hypothesis.
In theory maybe the null hypothesis should be picked to balance bias (which means us atheists pick "God exists" as the null hypothesis and get a result of "insufficient evidence to disprove"), but in practice of course nobody does that. In practice, nobody can even agree on how to evaluate the evidence. In practice one starts with the beliefs one has when one first asks the question and goes from there. If that happens to be some kind of theism, then that's because human children are credulous information sponges by dint of millions of years of evolution, not because they are intellectually lazy.
Scott misses the interesting question: namely which Calabi-Yau space is Her favourite, and which irreducible representation of E8
We gots to know.
Tim: almost all reasoned arguments start from some default assumptions.
There has been a lot of work in trying to reduce them to the simplest possible subset of assumptions, which turns out to be intractable.
Also, the string coding "exists"/"not exists" is not the level of simplicity, it'd be trivial to shorten and reverse the representation - but arguably part of the measure of the simplicity of the assumption is the implication. God hasn't done so well there, historically.
Or we can go back to Cartesian axioms and go from there.
If there are an infinite number of possible supreme spooks, then the probability that any particular one of of them exists is zero.
Q.E.D.
BUT, if there are only a large but finite number, like 101000 or so, do you walk away, or devote your life to finding the One True One?
Not to mention that inordinate fondness for beetles.
Tim said,
"Since representing "exists" takes less symbols in both English and logic than "not exists," I submit that "exists" is simpler and should therefore be the null hypothesis."
I'd argue that the theist position is "God exists", and the atheist (default) position simply has nothing to say on the subject. (I'm thinking of the definition of athieism as strictly a-theism, or 'without' theism here, not the assertion that a god does not exist).
It's the old teapot-orbiting-Jupiter question. The default position has nothing to say on the subject. If we had a nation of teapot-orbiting-Jupiter believers in this country, we'd probably have to invent the term ateapotist, just to describe those who don't believe the same way the rest of the masses do.
Of course if you make the assertion that something does not exist, then you have something to prove. If you take the position, however, that you simply have not been convinced by the evidence that something exists, then the person making the exitensial claim (god exists) is the one with something to prove.
Great post! When I am asked how I can possibly say "there is no god" I reply that I don't say that. I say, "I see no reason to begin to believe that there might be such a thing as a god." It takes longer, but it's a much more correct statement of my thoughts on the matter.
I made the same null hypothesis point on Chad's post but it seemed to go unnoticed. I didn't choose to be this way, there just was never a point where I was anything but an atheist. My parents did not raise me specifically as such, they simply chose not to force religion on me therefore I remained in the natural state - no belief in gods.
I believe that the vast majority of religious people do not choose to be such, they simply have that system of belief forced upon them from an early age. A number of them seem to change religions at some point, is this down to better PR? Special offers? Fancier costumes for the head honchos?
The question of reasons for choosing to follow a religion when one previously followed none seems odd. Can there be any rational reasons? As you mention it lapses into farce when you then think about reasons for choosing one god over another.
Anyone got any reasons? Anyone actually know any atheists who decided to follow religion and their reasons for doing so?
BUT, if there are only a large but finite number, like 10^1000 or so, do you walk away, or devote your life to finding the One True One?
Whatever you do, don't make a list of those names! ;)
It's hard to know, since "I was a self-centered, nihilistic atheist before I found Jee-sus" is standard evangelical boiler-plate. And what do you mean by "reasons"? How about a frozen tripartite waterfall, would that qualify? That is allegedly what tipped Francis Collins over the edge.
I think rationality ultimately fails with asking about the existence of God. It seems like asking if He exists is like asking what color are His socks.
The reasons I often see for atheism is:
God is a jerk, but God should be all loving, so He doesn't exist.
God can't exist because I can't see him.
I like the thought of a true null hypothesis, etc.--I would call this hard working atheism because there's some deep thought here, but these two reasons at some level are what I would call casual atheism because they are lazy reasons. If you look at the world around you, the death, destruction, etc. kinda sucks. How could God allow this? But if we look at what sucks about life, it boils down to natural disaster and human made disasters. God supposedly isn't responsible for humans if we have free will, and I think most scientists agree that what's good for a species might not always be good for the individual--that is, individuals may suffer even in an optimal outcome for a group or an ecosystem. Finally, if we have a supreme being who is ominipotent, omniscient, etc., and doesn't want to be seen, how can we prove otherwise? There's always a way to slip out of the net of rational investigation.
Ah, but which God?
I love that argument. Obviously, if god(s) exist, only one religion can be right. Which one?
Everyone is an atheist. Some just believe in one god less...
Of the two, the one that makes the most sense and is self-consistent is to "assume that there is no such thing as a God, until some evidence to the effect is presented". It's that simple, really (at least to me).
OldScratchHimself:
Well, to be fair, given a reasonable probability density function over each divinity attribute xi, one can have a finite probability for the divinity to have its i-th attribute in the range [xi, xi + dxi]. Of course, to truly capture the diversity of gods which can be invented by the human mind, p(x) will have to be infinite-dimensional.
Not to mention that a good many of the divinities tossed out into the conversation are logically inconsistent; that is, the only sensible position one can take toward them is ignosticism, and p({x}i) is zero for those values of {x}i.
Dang it, I can't discuss theology without LaTeX.
Oops: The index i should go within the curly brackets. I'll be quiet now.
Ok, so I get how the many incompatible things about religion justifies the attitude that all religions must be wrong. I actually think that most careful theologists will admit that not everything in the Bible, or Koran, etc. is strictly true (shh--don't tell the fundies!). At some level this is a discussion about _what_ religion to follow rather than whether God actually exists.
Furthermore, doesn't that kind of throw the baby out with the bathwater to claim that since there are multiple religions they're totally wrong? For all the bad, religion served a useful function for human society and in some ways protected experimental thought over a large fraction of humanity. Astronomy in particular owes much to early neolithic cultures who recognized as much as they might want to dance for the gods all the time, they still needed to pay attention to how stars were moving to eat well.
I also think the point is that we all choose our own null hypothesis, though it beats me how you prove it one way or the other.
There is of course the same problem that "most theory is wrong" - we know this because most theory is mutually contradictory.
There are two ways out of that: one is to find a meta-theory that unifies apparently disparate theories, or to become an observer...
Of course how you get there also matters.
The whole point about null hypotheses is that they are hypotheses, you are free to choose whichever and go from there. At that point the questions are whether to apply reason to the implications of the hypotheses, or merely observation - and you can ask meta-questions about hypotheses, if you care to: like are they consistent, are they constraining, are they falsifiable.
Ultimately it is a matter of taste.
Doesn't your analogy presume that there ever was a baby to begin with? Show us the baby, JohnD.
Mustafa,
The baby is the positive aspects of organized religion rather than God, per se. That's totally separate. I have no ability to prove God exists or even put forth a rational argument for why someone should believe He exists. I think that's a personal, spiritual, completely superrational choice.
I would certainly stand by my analogy and say that there are positive aspects in belonging to an organized religion, just as I feel there are positive aspects to being an atheist and rationally questioning the existence of God. I think the state of our current world provides a stable position to argue that having something like organized religion might be beneficial for society, such as a codified mechanism to pass along cultural and moral ideas, a structure that encourages community participation and a sense of belonging, etc. You certainly don't need a religion to do that, but the prevalence of fundamentalist groups, cults, etc. speak to a need among a significant portion of our population that need those things to feel whole. I for one would be happy if people wanted to be Methodists rather than join a suicide cult.
That is highly arguable. But since you support belief in things that are not at all provable simply to manipulate people's behaviour, I will simply oppose it on enlightenment libertarian grounds.
The reasons I often see for atheism is:
God is a jerk, but God should be all loving, so He doesn't exist.
This is basically an example of not so much trying to disprove the existence of "God" as trying to disprove the existence of some particular god. There may technically be ontological problems with disproving the existence of any and all gods, but if someone is positing a specific God (and they usually are), and that specific God is logically inconsistent, it's quite fair to point that out.
Finally, if we have a supreme being who is ominipotent, omniscient, etc., and doesn't want to be seen, how can we prove otherwise? There's always a way to slip out of the net of rational investigation.
Personally I think that if an idea is so inherently able to slip out of any attempt to rationally pin down its attributes, this is usually more a sign of a problem with the idea than it is a sign of problems with the idea of rational experimentation.
religion served a useful function for human society and in some ways protected experimental thought over a large fraction of humanity
I think it is clear that it is not at all valid, when trying to evaluate a specific factual claim like "God exists", to cite the social benefits of believing such a thing is true.
What if I were to pay you a dollar if you agree to believe there is no such thing as Global Warming? You would benefit from your belief, but this would not make the belief itself valid.