Dogmatic Skeptics

[liamconnor]

Here is a (rather lengthy) quote from G.K. Chesterton's Orthodoxy:

Somehow or other an extraordinary idea has arisen that the disbelievers in miracles consider them coldly and fairly, while believers in miracles accept them only in connection with some dogma. The fact is quite the other way. The believers in miracles accept them (rightly or wrongly) because they have evidence for them. The disbelievers in miracles deny them (rightly or wrongly) because they have a doctrine against them. The open, obvious, democratic thing is to believe an old apple-woman when she bears testimony to a miracle, just as you believe an old apple-woman when she bears testimony to a murder. The plain, popular course is to trust the peasant's word about the ghost exactly as far as you trust the peasant's word about the landlord. Being a peasant he will probably have a great deal of healthy agnosticism about both. Still you could fill the British Museum with evidence uttered by the peasant, and given in favour of the ghost. If it comes to human testimony there is a choking cataract of human testimony in favour of the supernatural. If you reject it, you can only mean one of two things. You reject the peasant's story about the ghost either because the man is a peasant or because the story is a ghost story. That is, you either deny the main principle of democracy, or you affirm the main principle of materialism-- the abstract impossibility of miracle. You have a perfect right to do so; but in that case you are the dogmatist. It is we Christians who accept all actual evidence--it is you rationalists who refuse actual evidence being constrained to do so by your creed. But I am not constrained by any creed in the matter, and looking impartially into certain miracles of mediaeval and modern times, I have come to the conclusion that they occurred. All argument against these plain facts is always argument in a circle. If I say, "Mediaeval documents attest certain miracles as much as they attest certain battles," they answer, "But mediaevals were superstitious"; if I want to know in what they were superstitious, the only ultimate answer is that they believed in the miracles. If I say "a peasant saw a ghost," I am told, "But peasants are so credulous." If I ask, "Why credulous?" the only answer is--that they see ghosts. Iceland is impossible because only stupid sailors have seen it; and the sailors are only stupid because they say they have seen Iceland. It is only fair to add that there is another argument that the unbeliever may rationally use against miracles, though he himself generally forgets to use it.

Do you agree or disagree with the thesis that Naturalists are dogmatic about their exclusion of the miraculous?

[marco]

Anyway, simple or not, the binary situation applies where we are concerned with a statement being true or false.

I see a potential problem with the above statement. This line of thinking appears to be viewing a given statement in a complete vacuum. In other words, from a purely philosophical perspective, given a statement in isolation where we have no other knowledge, then what you say above is clearly true because we have no additional information to consider.

And that is the entirety of what I am claiming. The argument is about a statement with no support to it. Once we start to look at other things, then our opinion will change. In religious matters we could say there's a 50:50 chance there is a God, if we are in a vacuum of information. As you say, we're not devoid of information, especially with Yahweh, so we can reach other conclusions.

G.K. Chesterton is talking about Christian beliefs and Christian "testimonies" being considered as "evidence". With that additional knowledge and the a very long history of Christian theists making all manner of provably false claims, suddenly a claim of having a miraculous experience coming from a Christian is far from having a potentially binary truth value. It's far more likely to be false. Even if nothing more than a sincere belief on the part of the Christian that is simply mistaken.

I have said as much elsewhere. I was examining only the proposition that we have no information. I agree with you that once we add information, we change our view.

A horrible auto crash occurs where 6 people were in the car. Five of them die and one walks away without a scratch. "It's a miracle!", the Christian proclaims.

Sorry, but that's hardly evidence for a miracle.

If there are 5 atheists and one Christian, the probability that the Christian will be the sole survivor (given we as yet don't know which one) is exactly 1, from a Christian point of view.
In reality God destroys churches with as much fervour as he destroys brothels, babies with baddies. And this suggests God and Chance are interchangeable.

[marco]

[quote="Furrowed Brow"]

These numbers convey the basic asymmetry. From a position of ignorance we start with the probabilities: Deliberately true (1/n), Lie (1/n), Error (1/n), Coincidently true (1/m, where 1/m is always less than 1/n), and then: 1/n + 1/n +1/n +1/m = 1.

Assuming this is intended to be serious, I should point out that the addition rule applies only to mutually exclusive events. The best we can say for this quasi-mathematical play is it looks good but has moved a continent away initial conditions of ignorance.

[Furrowed Brow]

Assuming this is intended to be serious, I should point out that the addition rule applies only to mutually exclusive events. The best we can say for this quasi-mathematical play is it looks good but has moved a continent away initial conditions of ignorance.

Maybe the play at math didn't move much for you but it was meant to clarify an a priori asymmetry built into the meaning of the four terms.

• 1] Deliberately True i.e. intentional or considered
  2] Coincidentally True i.e. lucky guess
  3] Lie (Deliberately False)
  4] Error i.e. accidental deviation from accuracy

These four concepts are contrary terms. No two terms may both apply without causing a contradiction. Which is to say they are all mutually exclusive evaluations of an event. As a list they also exhaust all the possibilities. So when an agnostic says a statement is either true or false they may not realise but they are invoking this set of four possibilities.

Remember we are considering a witness statement so before reading the statement it is already less likely the witness has made a lucky guess. This is why I ignored the option in the earlier three part formulations as it was an insignificant technical distraction from the thrust of what I was saying which was: lacking anymore information the other three options 1, 3, 4 are each equally likely.

But here is the important point: the asymmetry I have been pushing for, viz., a priori any witness statement is more likely false than true, would collapse if option 2 were just as likely as the other three options.

If you wish to continue to contend lacking more information it is still 50/50 whether a witness statement is true or false, it will come down to whether option 2 is considered a priori less likely than the other options. This we can knock back and forth and I doubt minds will be changed but we wild have clarified one aspect of the gulf between the sceptic and the agnostic.

[marco]

Maybe the play at math didn't move much for you but it was meant to clarify an a priori asymmetry built into the meaning of the four terms.

• 1] Deliberately True i.e. intentional or considered
  2] Coincidentally True i.e. lucky guess
  3] Lie (Deliberately False)
  4] Error i.e. accidental deviation from accuracy

These four concepts are contrary terms. No two terms may both apply without causing a contradiction. Which is to say they are all mutually exclusive evaluations of an event. As a list they also exhaust all the possibilities. So when an agnostic says a statement is either true or false they may not realise but they are invoking this set of four possibilities.

Hmmm. True and false have been subdivided. The binary idea remains. Allocating equal probabilities to the supposed subdivision is rather arbitrary. I think the mathematics is just a dubious distraction.

Should you wish to continue to contend lacking more information it is still 50/50 whether a witness statement is true or false, it will come down to whether option 2 is considered a priori less likely than the other options.

I don't believe that supplying reasons for true and false, and then attempting to assign a probability to each, is useful. The obvious a priori situation is that the statement on which we have no information may be right or wrong. The allocation of reasons - it might be a hoax, a mistake, a coincidence does not help at all.
As I said, it is an interesting idea.... extracting multum ex nihilo.

[Furrowed Brow]

Hmmm. True and false have been subdivided.

Yes. And I'm hoping you are able to agree the subdivision are exhaustive. You may not think them useful but they are correct when assessing the truth of a statement. If so, what is on offer should be no more than tautology or if my phrasing has been too loose it should be obvious they can be reformulated as a tautology. A claim to truth may be deliberate or accidental, or it may be a lie or an error. These concepts are built into what it means to be true or false if we dig a bit. So you may say but it could be right or wrong...well yes....it could also be a considered truth or a lie or an error or just a lucky guess. We'd both be right but our analysis swerves off in different directions given our starting points.

Allocating equal probabilities to the supposed subdivision is rather arbitrary.

We are dealing with definitions that characterise how a statement may be true or false. A priori the probability of three of which are not distinguishable.* Given a statement we are yet to read and have not been informed of its provenance we do not know if it is the truth, a lie, or an error. For the moment let's go back to the three categories. There are two responses to this:

• 1] accept the concept of probability is here meaningless and therefore leave this an unknown (agnostic)
  2] count the categories and note there are two false categories compared to one (sceptic).

How to assess the two approaches? If you imagine an arcade game where a metal ball might drop into one of three slots. Given no more information the agnostic refuses to guess where the ball will land, the sceptic says my money is on false. Eventually the agnostic is forced to say this is false metaphor.

Now lets open this up to the full set of four possibilities. We have the same arcade machine to consider but this time the metal ball may land in one of four slots. We do not yet know which one. There are now three strategies

• 1] accept the concept of probability is here meaningless and therefore leave this an unknown (agnostic)
  2] accept the four categories are indistinguishable and the odds are equal (agnostic)
  3] point out one of the slots is slimmer than the other three (sceptic)

I am advocating 3. So the question is whether this is legitimate. The point rests on whether one of the four terms has a definition that makes it less probable than the other definitions.

The idea that some concepts are by definition less likely is not unusual. A miracle which is defined as a extraordinary or exceptional event is less likely than not - by definition. A long shot is by definition less likely than not. A fluke is less likely by definition. A windfall is less likely by definition.

The thrust of this argument is that there is a quirk in our definition of true and false that means by definition there is an asymmetry in those concepts. And it amounts to this: by definition any claim to truth is more likely an error than is it true by lucky guess. (Truth is hard to fluke). This asymmetry is built into the expanded definitions. - which unless you are going to argue otherwise are correct.

The obvious a priori situation is that the statement on which we have no information may be right or wrong.

Yes this is obvious and if we do not delve into definitions that inform the logical geography we may well think the two probabilities are a priori indistinguishable. But they are distinguishable.

Of course definitions do not determine the facts of the matter. But they do shape our starting point and whether we start out agnostic or sceptical. This all boils down to me claiming that scepticism is the logical starting point for any enquiry not agnosticism.

As I said, it is an interesting idea.... extracting multum ex nihilo.

Hate to be boring.


*(I was trying to make sense of the idea Error is more probable than deliberate truth but that could be a rabbit hole we don't need to go down).

[marco]

Yes. And I'm hoping you are able to agree the subdivision are exhaustive. You may not think them useful but they are correct when assessing the truth of a statement.

I believe one cannot subdivide the True/ False into x categories and say each category is equally likely. If one is told the unknown answer to a sum is an even number, there is no need to consider HOW this was obtained. It is sufficient to allocate equal probabilities to True and False. And this is a correct way of working.


The way in which an answer was wrong does not affect the 50% probability of a wrong answer. There could be a dozen ways, but that does not increase the 50% a priori probability.

The anomaly you claim - that falsehood is more likely than fiction - is simply based on the way you've chosen to evaluate probabilities. There are three outcomes when we throw two coins: 2 heads, 2 tails, a mixture. You would probably allocate equal probabilities to each outcome. You would be wrong. The correct evaluation is 1/4, 1/4/, 1/2.

To complicate matters more, if you were dealing with Bose-Einstein statistics, then the answer would, in that rarefied field, be 1/3, 1/3, 1/3. But quantum theory is sometimes counterintuitive.

As far as I see there is no reason to depart from the simple binary model. The fact you have hit an anomaly should suggest your method itself is at fault. Best wishes.

[otseng]

[Replying to post 44 by Mithrae]

Moderator Comment

Please debate without making personal comments about others.

Please review the Rules.


______________

Moderator comments do not count as a strike against any posters. They only serve as an acknowledgment that a post report has been received, but has not been judged to warrant a moderator warning against a particular poster. Any challenges or replies to moderator postings should be made via Private Message to avoid derailing topics.

[JoeyKnothead]

From the OP:

Do you agree or disagree with the thesis that Naturalists are dogmatic about their exclusion of the miraculous?

Naturalists are dogmatic about the natural. It's only when folks set to declare their claims of the non-natural do the nauralists set to fret.

Or, it's when those who claim the "miraculous" occured, and can't show it did, well how 'bout that.

Show me a god exists and I might can believe me in miracles. Until then, it'll truly be a miracle that I'd set to believe me in the "miraculous" claims of them that can't show they was.

[Mithrae]

But for the sake of argument let's assume that position of ignorance is a valid concept to apply in the case of communication: Is it correct to suppose that there are three logical options? Since first encountering C. S. Lewis' trilemma I have generally supposed so. But really, aren't the logical options
intentionally false (deception)
unintentionally false (error)
intentionally true
unintentionally true (coincidence)?

Good point. You are paying attention. I like the way you think.

I left out the fourth option because in the end whilst it adds a wrinkle I don't think it makes a significant difference to the basic argument. In natural language there is likely a good reason why the ideas of lying and error are natural concepts easy to grasp whilst the idea of a witness making a claim that was coincidentally true looks odd if not spurious.

Let's assume Joe just guessed and does not know Moe and has no idea whether Metallica were playing a concert last night. Joe's note could then well be false, but what are the chances he got a lucky hit. Coincidences by definition carry with them a smaller probability. So: there is a greater probability of a lie than a coincidence, there is greater probability of an error than a coincidence, and there is a greater probability of a deliberate truth than a coincidence. Whilst there are now four option to consider one option by definition is of less value than each of the other three options.

These numbers convey the basic asymmetry. From a position of ignorance we start with the probabilities: Deliberately true (1/n), Lie (1/n), Error (1/n), Coincidently true (1/m, where 1/m is always less than 1/n), and then: 1/n + 1/n +1/n +1/m = 1.

Naturally, we would then rightly suppose that in most cases coincidence is actually much less than a 25% possibility,

Compare the possibility of an error to coincidence. Given the number of things that can go wrong of which only one thing is needed to go wrong for there to be an an error compared to the number of things that must go right for a coincidence. In the probability model coincidence sits on top of the sombrero hat of improbabilities and an error is found at the bottom.

That's not necessarily the case; you're introducing an evaluation beyond the simple logical possibilities. And as I said, if we're going to do that, we should also note that reporting perceived truth is much simpler and much more common than making up a lie. Or alternatively, we could try to shoehorn 'truth is simpler than fiction' in as a "logical" principle alongside 'error is more likely than coincidence.' Either way, if unintentional falsehood > unintentional truth but intentional truth > intentional falsehood, there is still no good reason to consider a claim more likely false by default.

Then let's go back to my first two posts in the thread:

• Post #10:
  - Any single source of information has a non-zero possibility of being incorrect
  - Multiple converging sources have a much smaller possibility of all being incorrect
  - Information can be considered reliable relative to the breadth and unanimity of the sources confirming it

This analysis follows on after any a priori presuppositions. My basic argument is that scepticism is a priori before looking at the evidence.

But now here there is room to question your analysis. Let's first talk about something more mundane and consider a lottery.

Lottery winners have to prove a win. We don't take their word. The ticket is the source of the winning not the words of the eventual winner. I think your analysis glosses over this distinction. Try demanding winnings from the lottery without proof of winner. Let's say Moe looses his ticket and can't identify the store where the winning ticket was purchased but feels sure he had the ticket with the winning numbers. Hmm. I'm sorry Moe but . So what if a larger group of people claim they won the lottery and none have a winning ticket. In this instance do multiple converging sources really have a much smaller possibility of all being incorrect? Really?

Yes, there is a smaller chance of them all being incorrect. They're not going to get any money for their efforts, and it's not like the difference between a 99.9% probability Moe is incorrect and a 99% probably that Moe + a dozen others are all incorrect is particularly significant. But it is a simple mathematical fact that however unlikely it is for one person to think they'd lost a winning ticket, or however absurd it is for one person to think they've got a chance of scamming the win, it will be slightly more unlikely for two people to share those perspectives.

This is clearly a bunch of chancer and maybe people in genuine error. So the larger this group the greater the demonstration of human fallibility. An increasing group that prove fallibility does not increase the probability of the real winner who lost his ticket coming forward. In this instance the size of the data set is a red herring and its size is to be ignored because this self selective group is dominated by people who are certainly in error. And so in this instant breadth and unanimity of the sources is meaningless.

Of course, you have chosen example in which all but one claimant must be wrong, so then pointing out that the group is dominated by incorrect claims doesn't exactly tell us much. Nevertheless, it is possible for someone to lose a winning lottery ticket, and it's possible for them to make some desperate attempt at getting the money anyway. If your scenario involved the winnings being claimed by someone who owns the ticket, then of course all the hypothetical claimants without tickets are in error. But if the winning ticket were no-where to be found, it would actually be very plausible to suppose that the genuine winner might try to claim the money without the ticket. And while a single claimant could very probably be a scammer, and a dozen claimants could still quite probably all be scammers, the greater numbers would make it more likely that the winner is in there somewhere, even if he's never going to get the money.

Here is a different line of thought - the probability of someone winning the lottery is 1/n. If there are x tickets sold we find the sum of all this probabilities and the solution to x over n i.e. we find the value of x/n. The answer gives the probability of a winner this week instead of a roll over.

Any single source of information has a non-zero possibility of being incorrect

Ok this means every source could be a wrong claim. That is like saying any lottery ticket could be a loser. But if convergence means a rising probability something else occurs. A series of summed probabilities that increase the probability of a winner is like saying every ticket could be a winner I.E

[center]1/n + 1/n + 1/n +1/n + ... = x/n, and x/n is greater than 1/n.[/center]

I'm not sure what you're getting at here. Of course 1/n + 1/n is going to be greater than just 1/n, regardless of whether or not it equals x/n.

If saying any ticket can lose is the same as saying any ticket can win this mean the probability of a miracle is bound to increase the larger the data set if and only if there are no tickets that can never win. With that restriction the math can never generate the zero answer. That is like saying the possibility of every ticket never winning is not considered. In other words the calculation ignores the question - what if there is no God?. This is like considering the possibility that there is never going to be a lottery draw and someone is defrauding everyone who buys a ticket. Not sure how we build that possibility into the calculation. But this is exactly the wrinkle the calculation does not permit. Unable to produce a zero result the calculation fails to be unbiased. Applied to miracles it already presumes strong atheism must be false but it does not presume theism is false. Hardly objective.

I have outlined an approach to human testimony which
A) Most importantly, matches the model which we use as a matter of course in assessing information we can't or won't personally verify (ie, looking for convergence and credibility of sources confirming it),
B) Is logically and mathematically sound, and
C) Does not preclude any possibilities unless they are known to be impossible.

If there is a perspective which does preclude such possibilities - such as a definitive, absolute insistence that such-and-such cannot happen - then of course it will not be compatible with this approach. But I would suggest that this is a problem with unjustified exclusion of possibilities, not a problem with consistent application of the epistemic approach we tend to use in general. Still, you're incorrect that this approach could never produce a zero result; if no-one ever claimed to witness a miracle, there would be zero reason to suppose that miracles occur on this basis. That argument is a little bit like climate science contrarians insisting that it is "unfalsifiable" simply because it has so far failed to be falsified in a myriad of ways.

Finally, throughout the thread I have numerous times highlighted the fact that this probabilistic approach is simply a more general confirmation of what is specifically known from much more thoroughly documented cases of medically-unexplained rapid cures of serious illness. Miraculous healing is a phenomenon which has the depth of supporting evidence as well as the breadth. As I suggested in my previous post, there are two tiers on which we might be sceptical - variables/uncertainties applicable to a group as a whole, and variable/uncertainties at a more discrete level - and a number of the responses to the examples from Lourdes were along the lines of the former; that no matter how thoroughly documented, the whole thing is tainted by virtue of being a shrine and process controlled by the Catholic church. Weak though that specific complaint may be, it's also worth pointing out the more general argument from the fact that miracle reports can be found from every region and culture of the globe and every period in history, and commonly enough that dismissing them all as fringe lunacy would be utterly irrational.

A proof of the impossibility of miracles could still invalidate or require reinterpretation of all that testimonial evidence of course. But unless and until this proof is forthcoming, the rational position is to consider it highly probable that somehow or other miracles do occur.