Absence of Evidence is Evidence of Absence

[MorpheusSandman]

Continued from This thread.

(An absence of evidence is not evidence of absence.)

Yes it is. I'd like to know who came up with that mathematically provable false aphorism.

Nope. You should probably read your own material before posting it incorrectly. From the link: "The absence of an observation may be strong evidence of absence or very weak evidence of absence, depending on how likely the cause is to produce the observation."

False evidence is no evidence at all, but if you want to continue the discussion, I suggest a new thread.

I have no clue how you think that contradicts my point, especially when that article's thread title says the exact same thing I did (unless you think Yudkowsky is contradicting his own thread title). The post you quoted says, when you elide the complicating bits: "the absence... may be STRONG evidence... or VERY WEAK evidence..." what he's describing there is the LEVEL of evidence any absence is. I don't know what you mean by "false evidence," and I don't think "false evidence" is what's meant by "absence of evidence."

I used this example on another forum a while back: Imagine that you haven't seen your neighbor for two weeks. Typically, you see him a few times a week just incidentally, such as when you're leaving for work and he's watering his lawn. Now, there are many things that could explain WHY you haven't seen him. One possible explanation is that he's dead. Now, the only evidence you can have of him being dead, without actually seeing his body or being told by a relative, is the ABSENCE of evidence that he's alive, and this absence of evidence IS evidence of his absence (death). It may be very WEAK evidence, given that there are other possibilities that account for your observation (not seeing him) with a higher degree of probability, but this absence of evidence that he's alive is, necessarily, evidence that he's dead according to Bayes' Theorem.

[Lokasenna]

To some extent it must depend on the context of the argument, though broadly speaking I'm more in favour of Atheist's view.

Medievalists like myself are used to scrabbling in the dust of history, trying to discern what meaning we can from the relatively small percentage of literary material that has survived into the modern world. The idea that 'absence of evidence is not evidence of absence' is a key tenet of the medieval scholar - because our absences are so large and glaring, we have to treat them with respect, and not attempt to fill them with baseless conjecture, however much we might like to imagine that the hypothetical evidence (or lack thereof) might support our conclusions.

[MorpheusSandman]

Honestly, this issue really isn't even debatable because Bayes' Theorem--P(A|B) = (P(B|A)*P(A)) / P(B) ...Where P = Probability, A = proposition, B = evidence, and | = given--proves it formally. For someone to continue debating it means either there's a misunderstanding of the term or a lack of understanding of Bayes' Theory, or perhaps both. I'll try to explain with an example from your own field

The idea that 'absence of evidence is not evidence of absence' is a key tenet of the medieval scholar...

Let's say it's possible that there was a book written by Julius Caesar that's now lost, but it's also possible that he never wrote such a book. If he never wrote it, there would be an "absence of evidence" 100% of the time because the book doesn't exist. On the other hand, if Caesar HAD written the book, there would only be an "absence of evidence" some percentage of the time less than 100%. We may not know HOW MUCH less than 100%, but it must be less. So any continuing "absence of evidence" would necessarily be evidence in favor of the absence of the book. It may not be STRONG evidence, given how much is irrevocably lost to history for a variety of reasons, but it is evidence nonetheless. If you wanted to fill in some hypothetical numbers:

A = 50% Caesar didn't write the book
B = 90% probability there would be an absence of evidence even if the book was written
B|A= 100% probably there would be an absence of evidence given the book wasn't written

Following Bayes' Theorem you'd end up with a new A of 55.5%. So the absence of evidence, given the above numbers, has increased the likelihood Caesar didn't write the book (the absence) from 50% to almost 56%.

Absence of evidence is evidence absence. The above offers incontrovertible mathematic proof. Any continued use of the phrase means one doesn't understand the phrase or the math. How strong the evidence is is all that's debatable, and it depends upon a variety of factors, including how much absence would be expected even if the something existed. Obviously, we rarely have such precise numbers of these, but knowing the math involved and what it says at least prevents us from making objectively wrong statements like "absence of evidence isn't evidence of absence."

[Lokasenna]

The difficulty with your argument is that you are assuming that ignorance is quantifiable - where do you pick these numbers from, hypothetical or otherwise?

An absence of something, whether it ever existed in the first place or not, is (to use your terminology) 100% - what do you mean by 'there would only be an "absence of evidence" some percentage of the time less than 100%'? Something is either exists, or it does not - and if it does not, then you can't make suppositions about it.

Using your mathematical model as a proof is as flawed as asserting that, thanks to Xeno, the athlete will never the tortoise - which, of course, is nonsense. I could, for example, use your model to argue for the existence of practically anything - for example, that Caesar wrote a first James Bond novel centuries before Fleming appeared on the scene...

[MorpheusSandman]

The difficulty with your argument is that you are assuming that ignorance is quantifiable

Our ignorance is quantifiable to a certain extent depending on the subject. I play poker for a living. On every hand I am ignorant of a lot of factors involved in figuring out the best move to make. Yet, the better I am at quantifying the combination of my knowledge and ignorance, the better decision I can make, and the more I will win in the long run. Sometimes our ignorance can be known precisely, such as in a coin flip; our knowledge concerns the coins two equal sides, while our ignorance is our inability to calculate gravity, force of flip, landing surface, etc. that would (all in combination) determine what side the coin landed on. The combination of our knowledge and ignorance makes the flip 50/50, rather than 100/0.

what do you mean by 'there would only be an "absence of evidence" some percentage of the time less than 100%'?

Two possibilities: the book exists or it doesn't.

1. If the book doesn't exist, we'd expect there to be an "absence of evidence" for its existence 100% of the time.
2. If the book does exist, we'd expect there to be an "absence of evidence" for its existence some (less than 100%) of the time.

We do not know the probability of 2 precisely, but it must be less than 100% because the book exists. Anything that exists will leave some evidence for itself even if it's only in the form of its actual existence (meaning it would be possible to find the actual book).

Let's use a different example. Instead of a book that may or may not exist, let's take a book you KNOW exist,s but that you've misplaced. You begin by searching your bedroom. The same idea applies: either the book is in your bedroom, or it is not.

1. If the book is not in your bedroom, there will be an absence of evidence 100% of the time.
2. If the book is in your bedroom, there will be an absence of evidence less than 100% of the time.

You may search and search your room and not be able to find it, and for every place you look and don't find it, this "absence of evidence" is "evidence for (the book's) absence" (in that room). While you may not can quantify this evidence in precise numbers, this is one case in which common sense and the math actually gel; or do you continue to search that one room until you die endlessly repeating the phrase "absence of evidence is not evidence of absence" or do you eventually say to yourself "given the absence of evidence, it seems probably the book isn't here"?

Something is either exists, or it does not - and if it does not, then you can't make suppositions about it.

Sure we can; if something doesn't exist then it leaves no evidence of its existence.

Using your mathematical model as a proof is as flawed as asserting that, thanks to Xeno, the athlete will never the tortoise - which, of course, is nonsense. I could, for example, use your model to argue for the existence of practically anything - for example, that Caesar wrote a first James Bond novel centuries before Fleming appeared on the scene...

Yeah, you're just plain misunderstanding me because there's nothing in my argument that allows for you to "argue for the existence of practically anything," unless there IS evidence for those things. What evidence is there that Caesar wrote the first James Bond novel?

[The Atheist]

I have no clue how you think that contradicts my point, especially when that article's thread title says the exact same thing I did (unless you think Yudkowsky is contradicting his own thread title). The post you quoted says, when you elide the complicating bits: "the absence... may be STRONG evidence... or VERY WEAK evidence..." what he's describing there is the LEVEL of evidence any absence is.

I'm amazed we're having this discussion, because all he missed was "or no evidence at all" after the "VERY WEAK" bit.

The blindingly obvious point is that every new discovery is made in the lack of evidence. Do you really not get that point?

If you can classify the level of evidence for non-existence, then there's obviously some evidence to the contrary, or you wouldn't know it didn't exist! I'll slip in Sasquatch just there.

I don't know what you mean by "false evidence," and I don't think "false evidence" is what's meant by "absence of evidence."

I was referring to you quoting something that does not, in itself, constitute of evidence. It seems to be a conclusion based on false premises, but that's fine, I see it a lot.

I used this example on another forum a while back:

And I'm glad you brought it out here, because it exemplifies my point entirely.

In the case of your neighbour, you start from a position of evidence - your neighbour does actually exist. When you have no evidence at all, you cannot draw any conclusion from it.

In the case of Mr Walter P Smyth, 24 Maple St, Canada, you don't have evidence that he exists, or whether he's been murdered. You can go and find evidence of his existence, but right now, you have no evidence whatsoever. That is exactly the position with the premise in relation to every thing we haven't yet discovered. I should imagine even Galileo would have laughed at you if you'd told him about black holes.

All your premise does is enable a false conclusion to be drawn out of a hat. Russell's teapot.

You may now put your case for the fact that no other forms of life exist in the universe. We've been searching for them - never mind stumbling over evidence, we've been actively searching for them for decades.

[The Atheist]

Our ignorance is quantifiable to a certain extent depending on the subject. I play poker for a living.

You should take up Bingo instead, because that's the answer and you've skipped past it again.

Depending on the subject....

Exactly.

[Ecurb]

One of my friends believes Sasquatches roam the forests of our native Oregon. He points out that there is no "absence of evidence" for his belief -- there are photographs, eye-witness reports, and footprints (casts of which he has made himself).

I don't buy it. That's because there is an absence of the evidence that I expect would have been discovered if Sasquatches exist. No skeletons, bodies, or live specimens have been produced, despite the fact that the Pacific Northwest has been well explored and is filled with people. The only "evidence" I have for the non-existence of Sasquatches is this absence of (a particular kind of) evidence to the contrary. The extent to which such "absence of evidence" is persuasive is our own estimate about the probablility of finding evidence given the nature of the evidence we would expect to find and the intensity of our search for it. If I glanced through my bookshelves for my missing book without finding it, I would have weak evidence that it isn't there. If I scoured the bookshelves meticulously for hours, the absence of evidence would provide stronger evidence that it wasn't there.

This is basic scientific method. My disbelief in Sasquatches is falsifiable. Produce a Sasquatch, living or dead, and I'll alter my beliefs. However, if "absence of evidence is not evidence of absence" I have no evidence whatsoever to support my belief that Sasquatches do NOT exist.

[Volya]

If there is no evidence for something then surely it is just common sense that it is more likely to be false than true.

[MorpheusSandman]

I'm amazed we're having this discussion, because all he missed was "or no evidence at all" after the "VERY WEAK" bit.

The blindingly obvious point is that every new discovery is made in the lack of evidence. Do you really not get that point?

If you can classify the level of evidence for non-existence, then there's obviously some evidence to the contrary, or you wouldn't know it didn't exist! I'll slip in Sasquatch just there.

These all sound like non-sequitors. It is never "no evidence at all" even if it sometimes "very weak." It can never, ever be "no evidence" because of how Bayes' works. In the case of absent evidence, B|A will always be 100%, while B will always be less than 100%. If the thing existed, B CAN NEVER BE 100%! And as long as it's less than 100%, absence will always continue to support A more than ¬A.

I don't know what "every new discovery is made in the lack of evidence" means or why/how you think it's relevant to this issue. So, we didn't have evidence for something, discovered it, and then we do. So, even there: it existed, and it eventually provided proof (not just evidence) of its existence. If it didn't exist, we never would've found it!

You can rarely classify the level of evidence of anything precisely, but there are certain abstract statements that can be made about such situations: and the AOEISNEOA is not one of them; it's patently false.

I was referring to you quoting something that does not, in itself, constitute of evidence.

I'm still lost: I was quoting something that was false evidence?

That is exactly the position with the premise in relation to every thing we haven't yet discovered.

Indeed. But let's consider two categories:

1. Things we imagine to exist, but do not.
2. Things that actually exist, but that we don't have evidence for.

Now in both of these scenarios we are positing knowledge that we actually don't have (for 1. we say we "know" they don't exist, and for 2. we say we "know" they exist), so imagine that we don't have that knowledge. For the things in category 1. there will always, now and forever, be an absence of evidence for them because they don't exist. Yet, for things in category 2. while there may be an absence of evidence NOW, there MAY not be later. So it doesn't really matter if we know something exists or not, the same rules applies for potentially existing things: The continuing absence of evidence is evidence for their absence. Perhaps strong evidence, perhaps weak evidence, but evidence nonetheless. Ecurb's example of Sasquatches is a perfect example: the evidence that exists is far weaker than the absence of evidence that exists for their absence.

Really, there's absolutely no reason why this formula would just stop working or not apply to the potential existence of things. Consider the scenario where you've lost a book. You start searching your bedroom. After 5 minutes, there is an absence of evidence that the book is there. Is this in itself not evidence that book is actually absent in the room? Most would say "yes," and be correct; why would the same thought not apply to the potential existence of something?

You may now put your case for the fact that no other forms of life exist in the universe. We've been searching for them - never mind stumbling over evidence, we've been actively searching for them for decades.

The evidence that other life exists in the universe is actually us. We know that life can exist given the right combination of elements and enough trials, and we know (roughly) how many planets there are where life could exist. Now, it's indeed possible to argue endlessly over the numbers, and it's a good example where we recognize our knowledge is far, far outweighed by our ignorance. So while the "absence of evidence" for ET is, indeed "evidence for their absence," it is extremely weak evidence and, indeed, their existence may be more probably given other the other facts involved.

[MorpheusSandman]

This is basic scientific method. My disbelief in Sasquatches is falsifiable. Produce a Sasquatch, living or dead, and I'll alter my beliefs. However, if "absence of evidence is not evidence of absence" I have no evidence whatsoever to support my belief that Sasquatches do NOT exist.

Absolutely, and your post does get into some tangentially related subjects like what, exactly, constitutes evidence and absence of evidence and what kind of evidence/non-evidence is stronger/weaker than others. Obviously, I think you're absolutely right regarding Sasquatches for the reasons you've given. Photographs, eye-witnesses, and footprints seem far weaker evidence than actually producing a Sasquatch (given how we know footprints can be faked, like Crop Circles, and that eye-witnesses can lie, be mistaken/confused, etc.), and the fact that nobody has produced one is quite strong evidence they don't exist given how populated and explored the relevant areas have been.

[Emil Miller]

Rummy said it best.

There are known knowns. These are things we know that we know. There are known unknowns. That is to say, there are things that we know we don't know. But there are also unknown unknowns. There are things we don't know we don't know.

http://youtu.be/_RpSv3HjpEw


The hilarious part is the guy at the end.

[AuntShecky]

Julius Caesar didn't write a book? If that's the case, I'm suing my Latin I teacher for foisting upon the class the work of an imposter! This litigation, of course, would entail borrowing the Way Back machine from Sherman and Mr. Peabody.

As far as evidence of absence/absence of evidence: a cable news channel last night featured a side-bar issue concerning the mystifying disappearance of Malaysian Air 370. The gist of that segment centered on the the old dichotomy of supply and demand: the lack of of facts (absence) vs. the world's clamor for answers (evidence.) In this case, there a chasm separates what is actually known about that flight and the pressure by the various media and their audiences to know exactly what happened. Thus, self-appointed "experts" rush to fill that gap with wild speculation and preposterous conspiracy theories, from the somewhat plausible --hi-jacking or an elaborate suicide plot on the part of a crewman or passenger-- to the outlandish and scientifically specious --aliens, black holes, as if a black hole did surgical strikes, sucking in a single aircraft rather than the whole solar system! Supplanting the disturbing mystery with a theory, no matter how unreasonable or irrational, somehow alleviates fear by allowing a small measure of control over the uncontrollable.

If absence of evidence creates a barrier between reality and one's hopeful perception of reality, then there are only two things one can do: drop the whole matter or continue the search for evidence. (Those courses of action would hold through for most cases EXCEPT matters of religious faith, because faith, by definition, is intuitive and thus beyond the scope of "proof." Cf. Douglas "The Hitch-hikers Guide to the the Galaxy" Adams, who said that if your faith depends on evidence, then your faith is nothing. )

[Hawkman]

What seems to me to be a flaw in the argument is that we have no reason to suppose that anything exists without some kind of evidence for it. Even for the debate over whether God exists is based on the primary evidence of the universe and everything in it. God is therefore one of many theoretical explanations for it. Consequently, evidence for the existence of god cannot actually be considered incontrovertible proof that he exists. As for the adventures of James Bond being written by Julius Caesar there is no reason to suppose that such a book ever existed, so why would one look for it? If he had written such a book then it would be possible that we would know of its existence from secondary evidence in the form of contemporary or near contemporary references. If evidence should turn up which stands up to verifiable historical and scientific testing, then it would be reasonable to accept that such a book existed but has subsequently been lost. Whether or not the book is ever found, it is reasonable to accept that it had existed. If the evidence is proved false and a hoax, then there is no evidence for the existence of the book.

[The Atheist]

2. Things that actually exist, but that we don't have evidence for.

....Yet, for things in category 2. while there may be an absence of evidence NOW, there MAY not be later.

There you go - now you understand. We're not talking about the future, or at least you weren't.

I find it strange that you can type that - which completely negates the statement, yet you continue to try to defend it.

Ho hum.