Just another Dallen topic that doesn't fit anywhere (unless it does)

[dallen68]

You know how in some of the videos Shane is on about how creationists and statist and such are always on about you can't prove a negative, or absence of evidence is not evidence of absence?

I'm wondering if that's actually true... I'm wondering if there is ever a time when you can take an absence of evidence as evidence.

[BogosityForumUser]

Since it is late, I'll give you the quick philosophical answers:

1.  You can prove a negative, just not all negatives.  Similarly, you can prove a positive, just not all positives.  The simple reason is that all statements are positive and negative due to the rule of double negation.

2.  Absence of evidence is evidence of absence; HOWEVER, while it does lend to strength, it cannot by itself be enough to surmount the Gettier problem with justification (if you are a JTB epistemologist).

[MrBogosity]

Absence of evidence IS evidence of absence, if you look where the evidence is supposed to be and it isn't there.

[Lord T Hawkeye]

You know how in some of the videos Shane is on about how creationists and statist and such are always on about you can't prove a negative, or absence of evidence is not evidence of absence?

I'm wondering if that's actually true... I'm wondering if there is ever a time when you can take an absence of evidence as evidence.

I'm careful about my wording on these matters.

Most of the time to woo claims, I'll say "there is no evidence for X" and I need not elaborate.  If they can't provide evidence, then my point stands.

Now, if something is contradictory and thus cannot be true, THEN I will say "X cannot be so" and back up my statement with an explanation as to why.

[dallen68]

I'm careful about my wording on these matters.

Most of the time to woo claims, I'll say "there is no evidence for X" and I need not elaborate.  If they can't provide evidence, then my point stands.

Now, if something is contradictory and thus cannot be true, THEN I will say "X cannot be so" and back up my statement with an explanation as to why.

Ok, along the same lines, how do you go about evaluating evidence? I'm not asking for what it says in the textbook here, I already know that. What I'm asking is how each of you, individually and collectively determine what is and is not evidence.

Let's say Dr. Woo comes up with Woo's Cure for the Common Cold. You challenge Dr. Woo's claims. Dr. Woo presents the following evidence:

A long list of letters behind his name.
A long list of academic looking papers, with abstracts, and many-lettered-authors, and impressive sounding "university medical facility" associations (for the purpose of this exercise, pretend you've actually heard of at least one of them).

How do you go about determining if the presented evidence is real, or something somebody made up?

In addition, please pretend the following:

Your Lexus search determines that there is a Dr. Woo associated with Random State Medical University, and this person is cited as a secondary or tricery author once or twice a year in the Journal of Bacterial Epidemiology.

[MrBogosity]

Ok, along the same lines, how do you go about evaluating evidence? I'm not asking for what it says in the textbook here, I already know that. What I'm asking is how each of you, individually and collectively determine what is and is not evidence.

I do an informal Bayesian inference. This is where you kind of guesstimate things without doing any math. You ask the following questions:

How likely is the hypothesis? (Extraordinary claims require extraordinary evidence.)
Have they provided evidence for their hypothesis? (If not, we're just left with how likely the hypothesis is—and again, if it's extraordinary, it's rejected.)
How well does that evidence explain the hypothesis? (Is it extraordinary enough for the extraordinariness of the claim?)
What are the chances we would see this evidence even if the hypothesis were untrue? (If the chances of this are big, it makes the evidence less extraordinary.)

Let's say you and I are the only two people on Earth, and we know everything humans know now EXCEPT how the solar system works. You are trying to convince me that the Earth rotates, as opposed to the sun going around it. I ask you for your evidence.

Let's say your evidence is to flip a coin: heads the Earth rotates, tails the sun goes around it. Flipping heads is not extraordinary enough to meet the claim; further, there is just as much of a chance that the coin will flip heads even if the sun goes around the Earth: 50/50. (In a formal Bayesian equation, the former is in the numerator and the latter in the denominator, so if the chances are equal as they are here they cancel each other out.) What that means is the evidence is useless, and we're left with just how probable the hypothesis is absent any evidence.

But if your evidence is to make a Foucault pendulum, things change. There's no reason to expect a Foucault pendulum to work if the Earth is stationary and the sun goes around it. That probability is low. But if the Earth rotates, it should work, so the evidence is extraordinary enough to meet the extraordinariness of the claim.

Let's say Dr. Woo comes up with Woo's Cure for the Common Cold. You challenge Dr. Woo's claims. Dr. Woo presents the following evidence:

A long list of letters behind his name.
A long list of academic looking papers, with abstracts, and many-lettered-authors, and impressive sounding "university medical facility" associations (for the purpose of this exercise, pretend you've actually heard of at least one of them).

How do you go about determining if the presented evidence is real, or something somebody made up?

The letters beside his name are pretty much worthless for this. You want to look at the academic papers. What you're really looking for is, have other scientists replicated his results? You also need to know enough about studies to understand proper methodology and how to interpret the results.

A good example is on the podcast a few weeks ago. The new IPCC report calculates the human component of Global Warming to at least 50% to a 95% confidence. Lots of people—even skeptics—interpreted this to mean that we're 95% sure humans are causing Global Warming. It doesn't mean that. It means we're 95% sure the measured value lies within the error bars. To calculate how sure we are humans are causing Global Warming, you need a Bayesian analysis.

[MrBogosity]

Another example:

Let's say that someone claims to own a cat. Lots of people own cats, so the prior probability is high; you don't need that much evidence to meet it. You might just take his word, or if you're at all skeptical, ask for a picture.

Now let's say he claims to own a Bengal tiger. It's not unheard of, but it's rare enough that a picture might not be enough; to meet the prior probability you might have to go and see it for yourself.

Now let's say he claims to own a unicorn. The prior probability is VERY low. Even if you went over there and saw what appears to be a unicorn, this probably wouldn't satisfy you. Since no one has ever verified the existence of a unicorn, the evidence is actually more consistent with some sick vet grafting a horn onto some poor horse. Since it's evidence we could see if the hypothesis were untrue, it's not extraordinary enough to cover prior probability.

Get the idea?

[tnu]

the brony in me really wants to find osme truth to the last one.

[dallen68]

Another one you can help me with, because apparently, I failed a logic quiz:

Person has music lessons on Thursday. He has music lessons today.
I'm told that I should know that today is Thursday. (by those conducting said test)
Difficulty: I had music lessons as a child, the only thing this statement proves is it's not Sunday or an odd numbered Saturday.
What gives?

[evensgrey]

Another one you can help me with, because apparently, I failed a logic quiz:

Person has music lessons on Thursday. He has music lessons today.
I'm told that I should know that today is Thursday. (by those conducting said test)
Difficulty: I had music lessons as a child, the only thing this statement proves is it's not Sunday or an odd numbered Saturday.
What gives?

The person who says you failed the test does not understand the question.

The two statements:

1: Person A has music lessons on Thursday.
2: Person A has music lessons today.

Is NOT sufficient to conclude:  Today is Thursday.

The reason for this is nothing in the statements precludes Person A from having music lessons on some other day or days as well as on Thursday.  All that the two statements, as presented, allow us to conclude that today MAY be Thursday (that is, the conditions given do not preclude today being Thursday).

[MrBogosity]

The person who says you failed the test does not understand the question.

The two statements:

1: Person A has music lessons on Thursday.
2: Person A has music lessons today.

Is NOT sufficient to conclude:  Today is Thursday.

The reason for this is nothing in the statements precludes Person A from having music lessons on some other day or days as well as on Thursday.  All that the two statements, as presented, allow us to conclude that today MAY be Thursday (that is, the conditions given do not preclude today being Thursday).

Yes, this is exactly the sort of bad logic that Wason's Four-Card Task exposes. It's Confirmation Bias: the tendency is to look for something that confirms it's Thursday (having music lessons) while not looking for things that falsify it, or that test for it being other days of the week.

[dallen68]

From the same test:

You and a friend go out to lunch. You agree to split the bill evenly.
A week later, your talking with said friend, trying to settle the bill. (Why you didn't do that in the café isn't explained) At any rate, friend remembers that you had one more coffee than they did, and Mr. Random in the next booth paid his $18 bill with a twenty. Since it was your birthday, friend agreed to pay for a slice of pie which was (a given number). How much do you owe your friend?

What the fuck does Mr. Random's bill have to do with me?
Why are we splitting hairs about one cup of coffee, a week later?
Why didn't we settle this when we were in the café (okay, it doesn't specify it was a café)?
I need better friends.

[MrBogosity]

From the same test:

You and a friend go out to lunch. You agree to split the bill evenly.
A week later, your talking with said friend, trying to settle the bill. (Why you didn't do that in the café isn't explained) At any rate, friend remembers that you had one more coffee than they did, and Mr. Random in the next booth paid his $18 bill with a twenty. Since it was your birthday, friend agreed to pay for a slice of pie which was (a given number). How much do you owe your friend?

What the fuck does Mr. Random's bill have to do with me?
Why are we splitting hairs about one cup of coffee, a week later?
Why didn't we settle this when we were in the café (okay, it doesn't specify it was a café)?
I need better friends.

None of that is relevant, since you agreed to split the bill evenly.

What was the answer supposed to be?

[dallen68]

None of that is relevant, since you agreed to split the bill evenly.

What was the answer supposed to be?

"I don't know, because the total is not mentioned"

[evensgrey]

"I don't know, because the total is not mentioned"

I notice a theme emerging.

Did ANY of the questions provide sufficient information to determine the actual answer?