I think that we lead our lives on the principal that induction/statistics yields true results. We can be 100% sure of virtually nothing so we rely on induction/statistics: should I get on that plane today? The last one did not crash so I guess it's OK... induction/statistics.

I think therefore we can extend induction/statistics to provide an answer to many questions that are not amenable to the deductive method. I Admit with questions like the origin of the universe, any inductions we make have a wide margin of error.

I am roughly speaking agnostic leaning in the direction of deism.

Starting any any value other than 50%/50% would be arbitrary. Its optimal to assign 50%/50% - no bias at all for/against the proposition. — Devans99

Which is completely arbitrary with respect to what's the case without their being any epistemological justification for two options being equally likely.

Which is completely arbitrary with respect to what's the case without their being any epistemological justification for two options being equally likely. — Terrapin Station

I thought the justification given here was adequate:

https://thephilosophyforum.com/discussion/comment/270560

Most things follow the normal distribution. And if we were to take all know distributions and do a weighted average of them into a meta-distribution, I would warrant it would look normal too. So 50%/50% is not arbitrary; it is optimal in statistical terms.

I thought the justification given here was adequate: — Devans99

There's no more reason to choose 50/50 than 100/0. Both are just as arbitrary in lieu of any information.

But we have information on the distribution of answers to unknown boolean questions; we know its definitely not 100/0 or 0/100.

So the epistemic justification is?

...is the fact that we know the distribution of answers to unknown boolean questions is definitely not 100/0 or 0/100. We know that boolean questions have distribution spaces for their answers. We know that distributions on average follow the normal distribution. So therefore we know that choosing 50%/50% is statistically the most likely correct thing to do.

.is the fact that we know the distribution of answers to unknown boolean questions is definitely not 100/0 or 0/100 — Devans99

I'm asking for the justification of that claim that I'm quoting, as well as the justification for the claim that if that's known for x, y and z, then it's reasonable to assume for this particular question, too.

My justification is inductive from everyday experience. Questions that are boolean and that have an underlying boolean sample space, have in my experience not come out to be all 100% yes or all 100% no. I can't claim I know the exact number of boolean questions in my life that have come out as yes versus no but the most reasonable assumption is to assume 50% yes 50% no.

So when I am confronted with a new boolean question on which I have no evidence, I would start at 50% likelihood it is true.

Questions that are boolean and that have an underlying boolean sample space — Devans99

What are a couple examples of this?

'Did the murderer do it?' is a good example. In absence of statistics for how many people in court actually come out guilty, we'd start by assuming it is 50% likely that the accused is guilty and then modify that estimate as we hear the evidence for/against.

Say there was blood on his shirt. We might say that makes it 25% likely (on its own) that the accused is guilty. So the revised estimate is: 50% + 50% * 25% = 62.5%. And so on for any other evidence we have for/against the accused.

Another example would be 'is there life on Mars?'. Before taking in any evidence about our knowledge of the solar system and the kinds of places life can tolerate, it would be correct to start by assuming a 50% probability of life. We could then adjust this up/down in the light of the various evidence we have about Mars. For example, Mars has water so adjust upwards by an amount.

I don't believe the reason you mentioned, is an epistemic, it's logical, but not epistemic.

What you always identify to as an epistemic involves the presence and utilization of numbers and statistics. That alone, is pure evidence that an epistemic is absent, as we are looking for an answer without numbers, without involving statistics. It should not involve statistics as we, I assume, have agreed that the characteristics of the mathematical statement only extrapolates the logic behind the mathematical statement for the origins of the universe — not the reason behind that statement.

'Did the murderer do it?' is a good example. In absence of statistics for how many people in court actually come out guilty, we'd start by assuming it is 50% likely that the accused is guilty and then modify that estimate as we hear the evidence for/against. — Devans99

But there's no epistemic justification for assuming a 50/50 split on the question of whether someone committed a murder in that case. There would be no justification for assigning any probability to it whatsoever. Saying that there's a 50/50 probability for something is not the same thing as saying you have no idea what's more likely.--Especially if we use the 50/50 as a base for further calculations.

That alone, is pure evidence that an epistemic is absent, as we are looking for an answer without numbers, without involving statistics — SethRy

Well that relates to what degree you regard inductive evidence as true knowledge. Once you get beyond 1+1=2 nearly everything we know, we know inductively. Are we brains in vats? We answer no inductively. I don't see that we will ever get very fair without induction/statistics.

But there's no epistemic justification for assuming a 50/50 split on the question of whether someone committed a murder in that case. There would be no justification for assigning any probability to it whatsoever. — Terrapin Station

If you had to make an assumption without evidence, would you assume he is definitely guilty, definitely innocent, or somewhere in-between?

Say there was blood on his shirt. We might say that makes it 25% likely (on its own) that the accused is guilty. — Devans99

That's the problem. You suppose 25% defines the blood stain on the shirt, that doesn't mean it creates a definitive number if it makes up 25%. You particularly, used the word might, just implies it's completely arbitrary, you just used your intuition to tweak it's percentage for proving he is the murderer.

Once you get beyond 1+1=2 nearly everything we know, we know inductively. — Devans99

A mathematical statement is permanent, it cannot be tweaked unless it's former terms are altered. The creation of the universe isn't, so sometimes deductive argument is needed to create an epistemological conclusion; a conclusion that is not driven by mathematical principle, but by rational reasoning.

What I am doing is taking inductive statements like 'blood on the shirt makes it likely he was guilty' and fitting % numbers to them. This does indeed need some estimation and is not exact.

If you think about what your mind would do in a court case as you are presented with evidence, subconsciously it would perform a similar process: blood on the shirt so that makes him a little more likely guilty, prints on the knife so that makes him a lot more likely and so on.

That is all I am doing with my calculation; I estimate % likelihood for each piece of evidence on its own and then combine the results.

The creation of the universe isn't, so sometimes deductive argument is needed to create an epistemological conclusion; a conclusion that is not driven by mathematical principle, but by rational reasoning. — SethRy

Even if we had a deductive proof demonstrating creation of the universe; would anyone 100% trust it? The first cause argument is meant to be that; it uses only cause and effect as an axiom, yet not many people place 100% trust in it. So even in the presence of a deductive proof, there would still be a need for a meta-analysis to combine the evidence from the deductive proof with other available (empirical etc...) evidence.

If you think about what your mind would do in a court case as you are presented with evidence, subconsciously it would perform a similar process: blood on the shirt so that makes him a little more likely guilty, prints on the knife so that makes him a lot more likely and so on. — Devans99

Yes, you are correct.

Even if we had a deductive proof demonstrating creation of the universe; would anyone 100% trust it? The first cause argument is meant to be that; it uses only cause and effect as an argument, yet not many people place 100% trust in it. So even in the presence of a deductive proof, there would still be a need for a meta-analysis to combine the evidence from the deductive proof with other available (empirical etc...) evidence. — Devans99

Mathematician and Philosopher Edmund Husserl, proposed that although science aspires to find certainty in the cosmos - using empirical evidence (observations and data) subjects to biased assumptions. Experience, by itself is not science. In synergy, it requires rational thinking.

Mathematics does not rely on empirical evidence, it is full of assumptions to reach a conclusion. To identify something out of terms, it is a rational process. If you make a mathematical statement with terms that came from assumptions, it taints the certainty of the result you acquire. Husserl believed that having assumptions compromise philosophical investigation; this belief is called Phenomenology.

There has to be a reason why you argue something, not a 50/50 beginning.

If you had to make an assumption without evidence, would you assume he is definitely guilty, definitely innocent, or somewhere in-between? — Devans99

Crimes are a bad example of that re what I'm going to assume, because I agree that it's a good principle to assume that someone is innocent until we demonstrate, at least partially via "physical" evidence, beyond a reasonable doubt, that someone is guilty of the crime they've been accused of. That's a special case just because of the social upshots of assuming that anyone is guilty in lieu of that.

For other questions like this, though, I'd simply make no assumption whatsoever, because there's insufficient information. There's certainly no way to assign probabilities to something for which we have no information, no frequency data.

Please remember that the 50%/50% step can be placed at the end of the calculation rather than the start (it makes no difference). It can therefore be seen as splitting the remaining % unbiasedly AFTER hearing all the evidence (instead of starting at a neutral point BEFORE hearing the evidence).

For other questions like this, though, I'd simply make no assumption whatsoever, because there's insufficient information. There's certainly no way to assign probabilities to something for which we have no information, no frequency data — Terrapin Station

If you make no assumptions, you get no answers. I would like an answer so I choose to make assumptions.

↪Terrapin Station
...is the fact that we know the distribution of answers to unknown boolean questions is definitely not 100/0 or 0/100. We know that boolean questions have distribution spaces for their answers. We know that distributions on average follow the normal distribution. So therefore we know that choosing 50%/50% is statistically the most likely correct thing to do. — Devans99

I, for one, have no problem with starting at 50% - 50%. My problem is with the nonsense that atheists, theists, and "agnostic tending toward deism" add to the problem to get at something other than 50% - 50%.

There is absolutely nothing of substance...nothing unambiguous...that can be added to the initial 50% - 50% assumption that makes sense.

YOU ARE CORRECT....it does start at 50% - 50%.

It ends up there also.

I, for one, have no problem with starting at 50% - 50%. My problem is with the nonsense that atheists, theists, and "agnostic tending toward deism" add to the problem to get at something other than 50% - 50%. — Frank Apisa

For the God question, it's not as if there's no data to go on. 100% it's the case that there's no evidence of a God, not to mention that the very idea of it is absurd/incoherent.

I, for one, have no problem with starting at 50% - 50%. My problem is with the nonsense that atheists, theists, and "agnostic tending toward deism" add to the problem to get at something other than 50% - 50%. — Frank Apisa


For the God question, it's not as if there's no data to go on. 100% it's the case that there's no evidence of a God, not to mention that the very idea of it is absurd/incoherent. — Terrapin Station

Lemme ask you this, TS:

There is 100% NO EVIDENCE of any sentient life on any planet circling the nearest 25 stars to Sol. Not one iota of evidence that there is any sentient life on any of those planets.

What does that tell us about the probability of sentient life on any of those planets?

Possibility is different from probability. Remember that. It's fair for both extremes because there is no evidence to disprove a god, only theories. There is also no evidence to prove a god, only ideologies that are reinforced with compromising logic and semantics.

That constant argument that has been ongoing since the ancient times is monotonous because evidence is scarce for both sides. Therefore, starting at 50/50 is just as random as 100/0; just because 50 represents one extreme, which is definitely logical, does not make it 50% possible.

I have been saying this over and over again. The capacity for the Philosophy of Mathematics to comprehend metaphysical and epistemological rudiments is logically capable, but you're in the territory where both empirical evidence and rational thinking is required. Mathematics does not have the capacity to calculate the existence of a god, it does not have epistemic justification behind the 50/50 assumption.

100% it's the case that there's no evidence of a God, not to mention that the very idea of it is absurd/incoherent. — Terrapin Station

Not really the reason why 50/50 as a mathematical proposition is right, but in epistemic justification is wrong.

We can be 100% sure of virtually nothing — Devans99

So says much of what passes for wisdom.

And yet you are confident enough of your language skills to think that this is a thread about God, and not about rabbits.You are using English, despite any doubts you might have. You are using a computer, and the internet, and your fingers or voice or whatever you use to write. You are confident that there are folk reading your posts, that they will reply...

In sort, the things you take for granted far outweigh your doubts.

Further I hope you do not reach all these conclusions on the basis of some sort of Bayesian (not Boolean) analysis. And even if you did, doesn't that imply confidence in Baye's theorem, and in the ability of numbers to represent the world?

In what way, is Baye's theorem flawed?

Is it? Why would you think that? I think it works neatly.

Would it still be reliable, when you need rational thinking?

If so, how?

I don't really understand how Mathematics could just define absolutely everything. Please educate me.