Argument against all of science

[Darth Wong]

Those of you who debate anti-science types may be familiar with the following argument:

There are only two possible truth values: 1 or 0, true or false. And nothing in science can ever be proven to have a truth value of 1. In fact, since science is primarily inductive (generalizations from discrete observations) rather than deductive, none of its conclusions are even rational. As Hume pointed out, it is irrational to assume that the Sun will appear to rise in the East tomorrow just because it did every day before.

This is a very common argument which I have received many times on my site. I've thought about it for a little while, and I propose a counter-argument; let me know what you think of it.

I start by pointing out the two main premises of the argument:

1. "The only possible truth values are 1 and 0". This premise is assumed to be "self-evident", and like all such a priori statements, no actual justification is ever given. For people who delight in pointing out that no rational justification is ever given for the premises of empiricism, this is a rather ironic flaw. In the past, when dealing with this argument, I have usually contented myself by simply pointing out that this is an assumption.
2. "All inductive reasoning is invalid." Once again, the people who propose this argument consider it so manifestly self-evident that it requires no justification, so they do not provide one. Yet I must ask: if it has no justification, how do we know that this premise is true? Certainly, one can find an example of an inductive argument which was falsified, but how can one generalize about all inductive reasoning based on this example without relying on the same sort of induction which is explicitly denied by this very premise?

So now that I've pointed out my doubts about the argument's two main premises, how do I take the next step and show that the argument itself is actually invalid? I combine the two premises deductively and show that they can produce a demonstrably false conclusion. To wit, if all inductive reasoning is invalid, then we can deduce that any inductive conclusion is just as invalid as any other, so the conclusion "the Sun will rise in the East tomorrow" is no more or less invalid than the conclusion "the Sun will rise in the West tomorrow". Moreover, if the only possible truth values are 1 and 0, then we should assign precisely the same amount of certainty to both of these conclusions!

This is patently absurd; if we are just as certain that the Sun will rise in the West as the East, then by any reasonable definition of probability, the two events have equal probability. It is trivially easy to show from experience that the two events obviously do not have equal probability, despite the prediction from the argument's premises that they should. Ergo, the argument against science is invalid.

Does this make sense? Is there some glaring flaw in my reasoning which escaped my attention? Is there some more efficient method of attacking this tiresome argument? Comments would be appreciated.

[FireNexus]

The argument is entirely unnecessarry. The only need to defeat this argument (since theoretically speaking it logically flows assuming the premise is correct) is to ask for justification of the premises.

Since they make the claim, they have to justify it. If they're unable, they'll lose.

As nice as it would be to convince fundies of truth, the fact that they aren't already at least halfway there is evidence enough that they use the wall of ignorance. Nothing you say will convince them. So why spend the extra effort.

Of course, that's just my opinion, and I'm kind of a slacker.

[Kuroneko]

Those of you who debate anti-science types may be familiar with the following argument:

Oh, yes. I remember these words verbatim. By the gods, I wish I did not.

There are only two possible truth values: 1 or 0, true or false. And nothing in science can ever be proven to have a truth value of 1. In fact, since science is primarily inductive (generalizations from discrete observations) rather than deductive, none of its conclusions are even rational. As Hume pointed out, it is irrational to assume that the Sun will appear to rise in the East tomorrow just because it did every day before.

This is a very common argument which I have received many times on my site. I've thought about it for a little while, and I propose a counter-argument; let me know what you think of it.

I start by pointing out the two main premises of the argument:

1. "The only possible truth values are 1 and 0". This premise is assumed to be "self-evident", and like all such a priori statements, no actual justification is ever given. For people who delight in pointing out that no rational justification is ever given for the premises of empiricism, this is a rather ironic flaw. In the past, when dealing with this argument, I have usually contented myself by simply pointing out that this is an assumption.

This assumption is only justified if one assumes traditional Aristotelian-based logic. However, there are many systems of logic besides, some with many more truth-values. Product systems, a particular subclass of many-valued logics, for example, are useful for evaluation of sentences like "the present King of France is bald" vs. "the present King of France is not bald". Under traditional treatement, both sentences are false (contrary to intuition), because they contains terms which do not refer. Many-valued logics can get around such problems, and many more besides, rather easily.

A more commonly-known many-valued logic is termed 'fuzzy logic'. Also a fairly prominent example is the Bayesian treatment of science, which treats scientific statements as explicitly probabilistic, with truth-value being a real number in the interval [0,1], with probabilities adjusted according to evidence. Personally, I find the Bayesian account unconvincing, but I'm using it as more counterexamples as to why the assumption of two-valued logics is not justified, because simply there are so many alternatives.

1. "All inductive reasoning is invalid." Once again, the people who propose this argument consider it so manifestly self-evident that it requires no justification, so they do not provide one. Yet I must ask: if it has no justification, how do we know that this premise is true? Certainly, one can find an example of an inductive argument which was falsified, but how can one generalize about all inductive reasoning based on this example without relying on the same sort of induction which is explicitly denied by this very premise?

Ah. Well, this stance comes from reading Hume and ignorance of the philosophy of science that came after Hume. Hume's treatement was basically thus: arguments in the form "Some Fs are Gs" to "All Fs are Gs" is not a deductively valid argument, and there is no deductively valid argument that justifies the Principle of Induction [PI]. Hence, PI must be justified inductively. This gives rise to an obvious problem: the only justification for PI is circular. Hume then simply assumes PI is true more or less 'just because', or rather because without it science was thought to be impossible. This is not in itself a problem--after all, we take some things as axiomatic all the time (mathematical systems, and some physical principles like conservation of energy). What's one more axiom?

Then came Sir Karl Popper, who eliminated the need for the use of inductive arguments in science in the first place--through falsification. Under his doctrine, science doesn't really need induction. Instead of the inductive "observation->(probably)hypothesis", science only really needs "observation->hypothesis (not known to be false)", combined with a rigorous testing and revision should the hypothesis become known to be false. It is here that the scientific requirement of falsifiability became generally accepted. (Sir Karl was grinding and anti-Freudian ax when he formulated this account.)

So now that I've pointed out my doubts about the argument's two main premises, how do I take the next step and show that the argument itself is actually invalid? I combine the two premises deductively and show that they can produce a demonstrably false conclusion. To wit, if all inductive reasoning is invalid, then we can deduce that any inductive conclusion is just as invalid as any other, so the conclusion "the Sun will rise in the East tomorrow" is no more or less invalid than the conclusion "the Sun will rise in the West tomorrow". Moreover, if the only possible truth values are 1 and 0, then we should assign precisely the same amount of certainty to both of these conclusions!

In defense of Hume, I feel compelled to point out that he did actually assume the Principle of Induction. He simply did it without knowing of any logical justification for it, with full recognition that he did not have such justification--exactly because he saw the only way to justify it would be in a circular way.

This is patently absurd; if we are just as certain that the Sun will rise in the West as the East, then by any reasonable definition of probability, the two events have equal probability. It is trivially easy to show from experience that the two events obviously do not have equal probability, despite the prediction from the argument's premises that they should. Ergo, the argument against science is invalid.

Your argument is makes sense. I like it; it uses the assumptions of the anti-scientist against that position. However, your argument is abductive rather than deductive or inductive. It essentially reduces to the following:

• Anti-science assumptions -> Observation O [e.g., Sun rising] cannot be justified.
• O bloody makes sense!
• Therefore, not (Anti-science assumptions).

On purely logical grounds, you begged the question. Nevertheless, the argument is very compelling, because it shows that (Anti-science assumptions) make us unable to leave the semi-solipsist pit of ignorance.

Does this make sense? Is there some glaring flaw in my reasoning which escaped my attention? Is there some more efficient method of attacking this tiresome argument? Comments would be appreciated.

Yes, it makes sense. However, it should be noted that not only is the first assumption unjustified, but there is the simple fact that induction is not critical to the practice of science.

[Kuroneko]

Methink I need to amend some things, since I did not do your argument full justice it deserves.

Your argument is makes sense. I like it; it uses the assumptions of the anti-scientist against that position. However, your argument is abductive rather than deductive or inductive. ...

The argument stands as is. Now, if you were to argue for induction in that manner [(Induction->Prediction P)&(P is a reasonable prediction) -> Induction], then it would've been abductive reasoning. But that's not quite what you're doing.

Now, a strict Popperian scientist wouldn't even need induction to conclude Prediction P (e.g., Sun rising in the East tomorrow); rather, s/he would appeal physics combined with astronomical/geographical facts predicting such an event. The anti-scientist, however, does not have that option, and would be forced either accept the principle of induction or judge the prediction unreasonable. And really, if one is not allowed to predict the Sun rising in the East instead of the West, what kind of reason is that?

It's a very good argument, Mr. Wong. I apologize for my earlier misunderstanding of it.

[Nova Andromeda]

There are only two possible truth values: 1 or 0, true or false.

--This premise is false. In fuzzy logic you can have truth values between 1 and 0.
-In addition, I bet there are other systems of logic that extend beyond two values (true and false) to three or more values, but I haven't looked into the matter (I would, however, be interested in knowing if anyone else has).

[phongn]

--This premise is false. In fuzzy logic you can have truth values between 1 and 0.
-In addition, I bet there are other systems of logic that extend beyond two values (true and false) to three or more values, but I haven't looked into the matter (I would, however, be interested in knowing if anyone else has).

IIRC, there are ternary logic (and the USSR did some experiments in ternary computing). As for 'fuzzy logic', a more concrete example would be spam filtering -- using the Bayesian treatment that Kuroneko noted earlier.

[Lagmonster]

Kuroneko is a seriously underrated poster. I literally have nothing of helpful substance to add after him. Well done.

[Darth Wong]

Thanks for the feedback guys, especially Kuroneko.

[Rye]

[*]"All inductive reasoning is invalid." Once again, the people who propose this argument consider it so manifestly self-evident that it requires no justification, so they do not provide one. Yet I must ask: if it has no justification, how do we know that this premise is true? Certainly, one can find an example of an inductive argument which was falsified, but how can one generalize about all inductive reasoning based on this example without relying on the same sort of induction which is explicitly denied by this very premise?

Yeah, that's an example of the stolen concept fallacy. The reasoning that composes the statement denying inductive logic relies genetically on inductive logic. At least, that's how I understand the statement, something like this:

"This inductive argument x is invalid." infers the general proposition: "All inductive arguments are invalid." would be an example of inductive reasoning then subsequent stolen concept fallacy. The concept is used as a basis, i.e. taken as "valid" in order for the following reasoning, why denies it's own basis. So put that concept back right now, you crazy antiscience computer users!

This is patently absurd; if we are just as certain that the Sun will rise in the West as the East, then by any reasonable definition of probability, the two events have equal probability. It is trivially easy to show from experience that the two events obviously do not have equal probability, despite the prediction from the argument's premises that they should. Ergo, the argument against science is invalid.

Does this make sense? Is there some glaring flaw in my reasoning which escaped my attention? Is there some more efficient method of attacking this tiresome argument? Comments would be appreciated.

Makes sense to me; I mean, we all know that it's simply bullshit to say the Sun may rise in the west, because it never has, if it did one day, then we'd have to update that model of the universe with the new data, science can adapt to any facts you throw at it due to its fundamental nature, but there's no solid reason to believe it might until it's happened at least once. It'd probably be more efficient if you just outlined the second part of the reasoning as a stolen concept fallacy.

[Patrick Degan]

There are only two possible truth values: 1 or 0, true or false. And nothing in science can ever be proven to have a truth value of 1.

Smells like a huge False Dilemma Fallacy. The rest is babble —particularly relying on Hume to judge the validity of experience.

[Admiral Valdemar]

These people are two a penny on most religious boards. They only come out of the woodwork when they see you attacking religious logic (or lack of) and find it too tempting to attack science and how it works because they can't grasp the concepts we wield in that field.

Watch out for the odd nugget of gold from the rare breed that supposedly have an MSc. or other qualification in a science and yet can't recall basic 101 facts in said course. Jorge was one such archnemesis I had to constantly battle over at Theologyweb.com and Theologyforums.com.

[Alyrium Denryle]

Another problem with this argument is that it distorts the property of identity. While only 1 truth exists for any given problem or question, we cannot always know this absolute truth, and it is a matter of degrees as to how close we get to it. And we cannot always prove 100% whether our answer is correct, because there is always the posibility that something new will come up, and prove us wrong.

[Admiral Valdemar]

Another problem with this argument is that it distorts the property of identity. While only 1 truth exists for any given problem or question, we cannot always know this absolute truth, and it is a matter of degrees as to how close we get to it. And we cannot always prove 100% whether our answer is correct, because there is always the posibility that something new will come up, and prove us wrong.

But if science gets one thing wrong, science can get a LOAD of things wrong!

Why should we trust scientists? They keep updating and changing theories (which aren't fact!) and telling us lies! Obviously science is just as hit and miss as any other method of reason, so there's truth to the Bible.

And God exists.

[Alyrium Denryle]

Yes, science can get stuff wrong. However, when it does, it updates things to be closer to that ever elusive and probably impossible perfect truth.

The fact that sceince updates itself with new data is what makes it so great. It gets closer and closer to understanding the universe with each revision. If we took the bible at face value, pi would still equal 3, and we would still think the earth was flat.

Do you have evidence that god exists?

[Admiral Valdemar]

Yes, science can get stuff wrong. However, when it does, it updates things to be closer to that ever elusive and probably impossible perfect truth.

Wrong!!! Didn't science, like, say the Earth was the centre of the universe? We know that's not true! What else can they be wrong about? Makes you think.

The fact that sceince updates itself with new data is what makes it so great. It gets closer and closer to understanding the universe with each revision. If we took the bible at face value, pi would still equal 3, and we would still think the earth was flat.

The Bible is the word of the Creator though! Don't you see that he knows this stuff already and infinitely better than any human?! Humans get things wrong. God doesn't.

Do you have evidence that god exists?

Read the Bible and you'll see the light! I'm praying for you!

[Alyrium Denryle]

No, religion said the earth was the center of the universe, in fact, when sceintists proved religion wrong, the various churches did things like burn them at the stake, or put them under house arrest.

If the Word of The Creator is so perfect, why is pi 3.14159<infinity> rather than 3 liek ti says in the bible?

I have read the bible, and no amount of praying will convince me

[Admiral Valdemar]

And it goes on and on and on. Sometimes they may admit defeat on one point, other times it's like banging your head against a brick wall with a smile and prayer book.

These people are impossible to turn 99% of the time. But even if you get the odd one or two that e-mail you in secret saying you've opened their eyes, they'll not be listened to by the other fanatics.

Know thy enemy, people.

[salm]

Another problem with this argument is that it distorts the property of identity. While only 1 truth exists for any given problem or question, we cannot always know this absolute truth, and it is a matter of degrees as to how close we get to it. And we cannot always prove 100% whether our answer is correct, because there is always the posibility that something new will come up, and prove us wrong.

that´s pretty much this part of Kuroneko´s post:

Then came Sir Karl Popper, who eliminated the need for the use of inductive arguments in science in the first place--through falsification. Under his doctrine, science doesn't really need induction. Instead of the inductive "observation->(probably)hypothesis", science only really needs "observation->hypothesis (not known to be false)", combined with a rigorous testing and revision should the hypothesis become known to be false. It is here that the scientific requirement of falsifiability became generally accepted. (Sir Karl was grinding and anti-Freudian ax when he formulated this account.)

[Darth Wong]

Formally, as Kuroneko pointed out from the writings of Karl Popper, the scientific method works on the principle of falsification, hence it is theoretically unreliant on inductive reasoning. However, before a theory is formalized, it must first germinate in the mind of the researcher, and that process usually does involve inductive reasoning. You can't see someone happen repeatedly, verifiably, and reliably without eventually assigning a pattern to it, and in fact, one could not function as a scientist without the use of inductive reasoning on at least some level. While the method itself is not reliant upon induction, the process through which human beings go about formulating theories in the first place (and for that matter, living day-to-day) tends to be.

The "Sun rises in the East" argument is a good example of that: while philosophers could argue until they were blue in the face about whether it is logical to say with confidence that the Sun will rise in the East tomorrow, no normal human being is so incredibly skeptical of inductive conclusions in any practical sense as to honestly question that the Sun will rise in the East tomorrow. And given a large body of carefully collected experimental data, no scientist is going to seriously look at it and refuse to draw conclusions because he can't count on the universe functioning tomorrow the same way it does today. In short, the absolute refusal to employ inductive reasoning that some would recommend would be sheer madness.

In other words, while we cannot assign absolute truth values of 1 to inductive inferences, it would be even more ridiculous to assign truth values of 0 to those same inductive inferences no matter how much supporting data they use, which is what the extremists would have you do (and which is what my argument is designed to lay bare). If one accepts the existence of truth values between 0 and 1, and one accepts the possibility that one inductive inference can be more reliable than another by virtue of basing its conclusion upon more data points, then it stands to reason that as the number of corroborating data points approaches infinity, the truth value would approach 1, even if it never actually gets there.

Personally, I feel that the modern philosophical contempt for inductive reasoning has gone too far; "inductive" is virtually considered an insult in some philosophical circles. You can use that dreaded word to dismiss an argument wholesale, the way some politicians use the word "liberal".

[salm]

In other words, while we cannot assign absolute truth values of 1 to inductive inferences, it would be even more ridiculous to assign truth values of 0 to those same inductive inferences no matter how much supporting data they use, which is what the extremists would have you do (and which is what my argument is designed to lay bare).

istn´t that basicly what popper is saying?

- not assign value of 1 = there are only competing hypothesis but no proof

- it´s redicolous to assign value of 0 = the best hypothesis is the one that´s not proven wrong and tested the most.

[Admiral Valdemar]

To add to Mike's comments, the usage of metaphysics and other epistemological methods in such debates can confuse most newcomers to this area of debate. I myself have found it hard to understand what some well versed debaters in this area are going on about. They try to turn every straightforward answer into a minefield of off-topic thoughts or vague references. It doesn't help when you're talking about science and someone, rather than explain why they feel a protein is impossible to fold, goes on about the nature of how we see entropy and life.

[Darth Wong]

A lot of these people try to snowjob you by using a lot of arcane philosophy terms and name-dropping references, and some of them also seem to delight in wording their propositions in the most indirect possible way, so that it becomes a nearly insuperable challenge to actually figure out how their argument works in terms of a logic diagram.

[Kuroneko]

Formally, as Kuroneko pointed out from the writings of Karl Popper, the scientific method works on the principle of falsification, hence it is theoretically unreliant on inductive reasoning. However, before a theory is formalized, it must first germinate in the mind of the researcher, and that process usually does involve inductive reasoning. You can't see someone happen repeatedly, verifiably, and reliably without eventually assigning a pattern to it, and in fact, one could not function as a scientist without the use of inductive reasoning on at least some level.

Well, yes, but neither would the mathematician actually get anywhere without intuition. Semi-vague intuitions on one level do not invalidate the need for strict rigor at another. It seemly shows that intuition, or perhaps even more appropriately, imagination, is necessary for the actual practice of science (and mathematics). I very much doubt there's anyone sufficiently schizophrenic as to deny that.

While the method itself is not reliant upon induction, the process through which human beings go about formulating theories in the first place (and for that matter, living day-to-day) tends to be.

Ironically, somewhat echoing Hume's solution: let's do it because it is simply our way of doing things; it's our nature.

[...] If one accepts the existence of truth values between 0 and 1, and one accepts the possibility that one inductive inference can be more reliable than another by virtue of basing its conclusion upon more data points, then it stands to reason that as the number of corroborating data points approaches infinity, the truth value would approach 1, even if it never actually gets there.

Bayesianism attempts to do this explicitly. As long as the a priori probabilities satisfy the axioms of probability [P(~p) = 1-P(p); If p and p is analytic, then P(p)=1; If p and q are mutually exclusive, then P(pvq) = P(p) + P(q)], then everything is alright. New evidence adjust the probabilities via (a generalization of) Bayes's Theorem of conditional probabilities. Bayesianists then argue that over a long accumulation of evidence, the probabilities assigned to such statements will converge to the same value. Thus, it doesn't matter what the prior belief of the Sun rising in the East is (as long as it is nonzero and non-one). It could be 0.01% or somesuch, but as long as the opponent is behaving rationally, the cumulative addition of supporting evidence will make it converge to 1.

Personally, I feel that the modern philosophical contempt for inductive reasoning has gone too far; "inductive" is virtually considered an insult in some philosophical circles. You can use that dreaded word to dismiss an argument wholesale, the way some politicians use the word "liberal".

This is somewhat the case, but it's more to do with why we're doing induction. The principle of induction by itself is a metaphysical claim that has no justification but itself. It's not inductive reasoning that's the real problem; it's the squabble over how to justify it. Popper justified it by eliminating it. Some justify it by making an over-system with the same functionality--e.g., Bayesianism, as above. Yet others make a new breed of argument: the hypothetico-deductive, which is just abductive reasoning in a new dress.

Your accusation of contempt for inductive reasoning is more due to the influx of philosophy students who overstep their bounds in their ignorance. In real life, if an applicant for a PhD in philosphy who tried to argue such extreme positions seriously: say, all things are really one and differentiation is just an illusion, then the examiner would inspect his own office and observe: "One, two, three... F." Such extreme views are the problem, and unless the would-be philosopher is proposing a solution (or at least a critique of another's solution), it's a no-go.

[Darth Wong]

Well, yes, but neither would the mathematician actually get anywhere without intuition. Semi-vague intuitions on one level do not invalidate the need for strict rigor at another. It seemly shows that intuition, or perhaps even more appropriately, imagination, is necessary for the actual practice of science (and mathematics). I very much doubt there's anyone sufficiently schizophrenic as to deny that.

While the method itself is not reliant upon induction, the process through which human beings go about formulating theories in the first place (and for that matter, living day-to-day) tends to be.

Ironically, somewhat echoing Hume's solution: let's do it because it is simply our way of doing things; it's our nature.

True, although I am uncomfortable with treating inductive reasoning on the same level as raw intuition. That's my beef with the negativity toward induction; it does have more validity than gut feeling, even if it is not proof, per se.

Bayesianism attempts to do this explicitly. As long as the a priori probabilities satisfy the axioms of probability [P(~p) = 1-P(p); If p and p is analytic, then P(p)=1; If p and q are mutually exclusive, then P(pvq) = P(p) + P(q)], then everything is alright. New evidence adjust the probabilities via (a generalization of) Bayes's Theorem of conditional probabilities. Bayesianists then argue that over a long accumulation of evidence, the probabilities assigned to such statements will converge to the same value. Thus, it doesn't matter what the prior belief of the Sun rising in the East is (as long as it is nonzero and non-one). It could be 0.01% or somesuch, but as long as the opponent is behaving rationally, the cumulative addition of supporting evidence will make it converge to 1.

Cool, it appears I stumbled into the philosophical concept of Bayesianism without realizing it

This is somewhat the case, but it's more to do with why we're doing induction. The principle of induction by itself is a metaphysical claim that has no justification but itself. It's not inductive reasoning that's the real problem; it's the squabble over how to justify it. Popper justified it by eliminating it. Some justify it by making an over-system with the same functionality--e.g., Bayesianism, as above. Yet others make a new breed of argument: the hypothetico-deductive, which is just abductive reasoning in a new dress.

Your accusation of contempt for inductive reasoning is more due to the influx of philosophy students who overstep their bounds in their ignorance. In real life, if an applicant for a PhD in philosphy who tried to argue such extreme positions seriously: say, all things are really one and differentiation is just an illusion, then the examiner would inspect his own office and observe: "One, two, three... F." Such extreme views are the problem, and unless the would-be philosopher is proposing a solution (or at least a critique of another's solution), it's a no-go.

Granted; I suppose I should be more careful to distinguish between the smart-ass pseudo-philosophers that I keep running into (particularly in relation to creationism, where every Bible-thumper learns just enough of the terminology of philosophers to preach to the choir) and the genuine article.

[Kuroneko]

True, although I am uncomfortable with treating inductive reasoning on the same level as raw intuition. That's my beef with the negativity toward induction; it does have more validity than gut feeling, even if it is not proof, per se.

There is a way out of this even for one who holds that such inferences are mostly intuitive. The key difference is whose intuition it is. The first guess of an expert, even for a particular problem s/he has not encountered before, but nevertheless falls into the field of expertise, is more valuable than that of the layman. That's true no matter the field--mathematics, engineering, science, medicine, detective work, and whatnot. The former has a proven record of getting the job done, whereas the latter is truly no more than simple gut feeling. This is hardly guaranteed to be correct, but I doubt many would hold it unreasonable.

Also, if one wishes to be slightly more rigorous, one can introduce alternative explanations (rather than just treating it as guesses, even expert guesses) into this account. Suppose tomorrow you wake up, look out of the window, and observe that the ground and the surroundings are wet (suppose also that you live in the suburbs if that is not the case). What could be the reason? You quickly think up of a few possibilities:

1. Your neighbour, whom you suspect to be slightly off-balance, went crazy with the garden hose last night.
2. An accident in a firefighting helicopter caused it to dump water in your neighbordhood.
3. Aliens.
4. It rained.

Technically, the abductive argument [(Rain->Wetness)&(Wetness) -> Rain] is a logical fallacy, but there is no denying that it is the most plausible of the alternatives.

Cool, it appears I stumbled into the philosophical concept of Bayesianism without realizing it :D

Indeed. There are some conceptual problems with the Bayesian account, the most promiment of which is the relevance of evidence and simplicity of theories. At first glance it actually seems to abhor simplicity: Since probability requires that whenever p implies q, P(q) must be at least as great as P(p), and all circles are ellipses, it follows that Copernicus should have treated orbits as ellipses even though there was no known empirical evidence at the time to support it. This problem can be dealt with, but only by appealing to a principle which is a weaker form of induction. (Instead of the traditional account of particulars-to-universal inference, it only recognizes particulars-to-[next]particulars.)

Still, I have a certain amount of admiration of Bayesianism simply because it confronts the problem of induction head-on, rather than the Popperian "well, we don't really need it anyway" solution. Sir Popper's account of science is logically unassailable on this particular point (although not on others), but I wonder if cutting out induction completely is too high a cost.