Saturday, May 11, 2013

Vallicella on believing beyond the evidence

He considers it rational in many instances. Here. 

29 comments:

Dan Gillson said...

I'm still Lutheran enough to say that faith is our basic orientation to world. One can always take skepticism one step further than certainty.

zillipede said...

Not sure if I'm misunderstanding anything or if the examples are as terrible as they seem.

"Absolute evidentialism implies that the desert wanderer (...) may neither believe nor disbelieve of either source that it is potable or poisonous"

I would drink one of the sources and hope for the best; this doesn't mean that I believe that the source I drank from was potable. My belief, until I had more evidence, would still be that it has a 50% chance of being potable.


"But note that my initial belief that my neighbors are decent people -- a belief that I must have if I am to act neighborly toward them"

Why? Why would it be impossible to act neighborly just because that brings about the best results? (Of course, acting neighborly may even be my best option if the neighbors are nasty people.)

BeingItself said...

As zillipede has pointed out, these are just terrible examples and indicate what a sloppy thinker that guy is.

Regardless, this exercise is a tacit admission that religionists do not have adequate arguments and evidence for their beliefs.

Crude said...

Zillipede,

Are you saying there are cases where it's justifiable to act as if a proposition is true, even if evidence sufficient to secure belief that it is true is not available? Or would you put it another way?

zillipede said...

Crude:

I simply don't think that the above examples need entail "acting as if a proposition is true". In the case with the water sources, it doesn't matter if there are 100 sources and 99 of them are poisoned - I'm justified in choosing one at random because that's the strategy that brings about the best consequences. Calling that "acting as if my chosen source is potable" seems weird to me, when what I'm really doing is "acting as if there's a 1% chance that my chosen source is potable". I can be convinced that my actions will probably poison me, and still rationally choose that action if the alternatives are worse.

B. Prokop said...

"this exercise is a tacit admission that religionists do not have adequate arguments and evidence for their beliefs."

Maybe some "religionists" (whatever that means). I do (have adequate arguments and evidence - mostly evidence).

Let the others speak for themselves.

William said...

I like Bill's neighbor illustration.

The results of a choice to trust or distrust often changes those whom we choose to trust or distrust.

Bayesians beware. This kind of interaction between us and our fellow man (or God) cannot be reduced to a Bayesian calculation because the act of acting on our estimate can change (in us or them) what we are estimating.

Crude said...

Zillipede,

Calling that "acting as if my chosen source is potable" seems weird to me, when what I'm really doing is "acting as if there's a 1% chance that my chosen source is potable". I can be convinced that my actions will probably poison me, and still rationally choose that action if the alternatives are worse.

Okay, let's go with the 99/100 scenario. You pick a source at random and you drink it.

At that moment, do you believe you're drinking poison?

If not, does this change for a 999/1000 scenario? 9999/10000? Will it change at any point?

zillipede said...

Crude: At that moment I'll believe that it's likely but not certain that I'm drinking poison.

Crude said...

Zillipede,

Crude: At that moment I'll believe that it's likely but not certain that I'm drinking poison.

Alright. And this doesn't change, no matter the proportion of poison to water? It's never 'I believe I'm drinking poison', it's always 'I believe it's likely I'm drinking poison'?

I'm not going on full defense for Valicella here or anything - I'm mostly curious, and think the criticisms of his take can get pretty interesting.

zillipede said...

Crude: I'd like to avoid a semantic discussion about the word "belief", but if you define "believing" as "estimating that there is a >50% probability of something" then in the above scenario, I'll believe that I'm drinking poison. On the other hand, if I was entirely certain that I was drinking poison then there would be no point in drinking it; the only reason I'm drinking anything at all is that there's a chance it will be water.

Crude said...

Zillipede,

Well, I think asking when a belief is actually a belief is unavoidable here. You talk about not being 'entirely certain', but hell, that's never available. You don't even need knowledge of poison in that case - just walk into a room with water.

I was asking this because I really was trying to figure out when 'belief' actually kicks in on your view.

cautiouslycurious said...

Crude,
“Okay, let's go with the 99/100 scenario. You pick a source at random and you drink it.

At that moment, do you believe you're drinking poison?”

No. Some beliefs in play would be “It is likely that I’m drinking poison”, “drinking a random bottle gives me the best odds of survival”, “it is likely that I’ll die at the end of this scenario,” etc. While I wouldn’t believe I would be drinking poison (or else I wouldn’t drink it), I would hold the above beliefs to be true, hence taking the risk of drinking poison (based on belief two and if this condition doesn’t apply, then the rational choice is to not drink). It doesn’t take much thought to see that we have a sound basis for each of these. In the event you find yourself facing certain death and a %1 chance of surviving, go with the %1. I don’t see where faith is required or how this is somehow at odds with rationality.

Crude said...

Cautiouslycurious,

No. Some beliefs in play would be “It is likely that I’m drinking poison”, “drinking a random bottle gives me the best odds of survival”, “it is likely that I’ll die at the end of this scenario,” etc.

Alright. Same question to you then: If not, does this change for a 999/1000 scenario? 9999/10000? Will it change at any point?

cautiouslycurious said...

Crude,
Alright. Same question to you then: If not, does this change for a 999/1000 scenario? 9999/10000? Will it change at any point?

The only belief that would change would be “drinking a random bottle gives me the best odds of survival” since that would depend on the circumstances. Increasing the probability of it being poison would still lead me to believe that it is most likely poison, that I’ll most likely die at the end of it, and it would have a slim chance of being water.

Think of it this way. Suppose a friend gives you a lottery ticket (since knowing the odds, you would never buy one for yourself). Do you believe that it’s the winning ticket? Do you believe that it’s the losing ticket? Or something in between? If you believe it’s the winning ticket, congratulations, you believe you’ve won millions of dollars, no need to wait until the drawing; you can go to the car dealership straight away. If you believe it’s the losing ticket, then you don’t need to bother checking the numbers, since you believe they won’t match anyway. So, do you head to the car dealership, or would you throw the ticket away? Or the third alternative, you don’t believe either option presented, and instead believe something else, like “it’s unlikely that I have the winning ticket, but I might, so I’ll still check the winning numbers”. Even though it's millions to one, you could still rationally check the ticket.

zillipede said...

I don't have a very stringent or well-thought out definition of belief, but what I wrote above ("estimating that there is a >50% probability of something") seems pretty good to me.

"You talk about not being 'entirely certain', but hell, that's never available."

The absolute certainty that one source was potable water was a part of the original example: "He knows (never mind how) that one is potable".

Crude said...

CC,

The only belief that would change would be “drinking a random bottle gives me the best odds of survival” since that would depend on the circumstances. Increasing the probability of it being poison would still lead me to believe that it is most likely poison, that I’ll most likely die at the end of it, and it would have a slim chance of being water.

Alright, so the proportion of poison to water does not matter: you never 'believe source X is poison' per se, it sounds: you just believe in an odds evaluation.

Followup question: do you believe anything at directly? Or do you believe nothing but odds calculations, no matter how crude (ha ha) they are? Do you believe you're looking at a monitor right now, or do you say 'No, I don't believe that! I believe there's a X% (probably very high) chance I am looking at a monitor.'?

Keep in mind, this isn't in the service of defending Valicella's examples - I think they lead to some problems. But I'm pretty fascinated by the process of belief as you and Zill describe it.

Crude said...

Zill,

I don't have a very stringent or well-thought out definition of belief, but what I wrote above ("estimating that there is a >50% probability of something") seems pretty good to me.

Sure, I can respect that. But that seems to introduce a problem: if estimating >50% probability means you believe something, then in your 99/100 example, you'd believe you're drinking poison.

The absolute certainty that one source was potable water was a part of the original example: "He knows (never mind how) that one is potable".

I'm talking practicality, not in terms of thought experiments. And even what I said goes too far - arguably cogito ergo sum or something close to it is with certainty.

zillipede said...

"But that seems to introduce a problem: if estimating >50% probability means you believe something, then in your 99/100 example, you'd believe you're drinking poison."

I don't see the problem. Given my definition of "belief", this just means that I estimate that it's more likely than not that I'll die from whatever it it that I'm drinking, something which is true given the premises. And it's still the best option, so there's nothing irrational about the belief or my actions.

Crude said...

Zill,

I don't see the problem. Given my definition of "belief", this just means that I estimate that it's more likely than not that I'll die from whatever it it that I'm drinking, something which is true given the premises.

Well, no. If you believe any given drink you pick up in that situation is poison, it would be irrational to drink it.

Now, you can go CC's route and say that you don't ever believe that straightaway, you only estimate percentages. So I'd ask the same question of you: is there anything you believe, full stop, 100%? Or is it all odds calculations?

zillipede said...

I wrote "Given my definition of 'belief'". It seems that you're ignoring that and reading what I wrote as if I had been using YOUR definition of "belief" (which you're welcome to explain further), and then - unfairly, I think - judging my behavior as irrational based on this redefinition.

Crude said...

Zill,

It seems that you're ignoring that and reading what I wrote as if I had been using YOUR definition of "belief" (which you're welcome to explain further), and then - unfairly, I think - judging my behavior as irrational based on this redefinition.

I meant no misunderstanding - I'm just going by what I'm reading from you and trying to make sense of it. However, I never said your behavior is irrational. I'm just cashing out what I took you to be saying: "You believe X is poison. But you're going to drink X."

Now, I get that it sounds like you're using 'belief' as shorthand for a likelihood calculation. Great, we can run with that - I'll ask the same questions I did.

cautiouslycurious said...

Crude,
“Followup question: do you believe anything at directly? Or do you believe nothing but odds calculations, no matter how crude (ha ha) they are?”

I take sensory input at face value and some other non-empirical claims (e.g. labels, “My name is X”, etc.). However, further interpretation of that input is subject to error. So, that brown pole-like object with brown limbs extending outwards would be taken at face value. However, calling it a tree made of wood that would burn when in the presence of heat and oxygen would be an extra interpretation of that data. I more or less determine beliefs by how they drive/effect behaviors. So, if you want to build a fire and head over to that tree for a fuel source, then you believe it’s a tree (although confidence intervals still apply) and not an illusion. So, for the poison-water scenario, the determination to drink depends on the circumstances so we are able to determine that the relationship between the chance of being poison compared to not drinking is the relevant factor in play and not simply the determination of whether the liquid is water or poison. If they would have drunk it regardless of the circumstances, then they would have had a different set of beliefs.

“Do you believe you're looking at a monitor right now, or do you say 'No, I don't believe that! I believe there's a X% (probably very high) chance I am looking at a monitor.'?””

I believe that I’m looking at a monitor right now, with extremely high confidence.

Crude said...

CC,

I take sensory input at face value and some other non-empirical claims (e.g. labels, “My name is X”, etc.).

Alright. So your subjective experiences as experiences is something you cannot be in error about, according to you? "I have the experience of looking at a monitor right now" - sure, you could be hallucinating, but you're you have 100% belief that you're having that experience?

How about the labels? You say you accept them - does that mean you don't think you could be wrong about a label?

I believe that I’m looking at a monitor right now, with extremely high confidence.

Is that just another way of saying what I said? I mean, my phrasing was "No, I don't believe I'm looking at a monitor. But I believe I have this very high chance that I'm looking at a monitor."? If there's a difference between the two, I'd like to hear what.

So, if you want to build a fire and head over to that tree for a fuel source, then you believe it’s a tree (although confidence intervals still apply) and not an illusion. So, for the poison-water scenario, the determination to drink depends on the circumstances so we are able to determine that the relationship between the chance of being poison compared to not drinking is the relevant factor in play and not simply the determination of whether the liquid is water or poison.

Alright, but the tree example seems a little odd compared with what follows. So even if the tree may be an illusion, in attempting to use it for fuel, you believe it to be a tree. The reference to 'confidence intervals' has me wondering again. With the poison/water case, surely you don't believe it to be poison. Do you 'believe it to be water' with some relevant confidence interval?

B. Prokop said...

This thread may not be the best place to post this, but just so no one wonders what happened to me, I feel the need for another "Internet Fast" (other than e-mail, which seems essential to conducting business nowadays). So I am signing off for the next 30 days (and maybe longer, if I still feel the need after that time has passed).

See you all in late June (or even July).

"Be still, and know that I am God." (Psalm 46:10)

zillipede said...

"is there anything you believe, full stop, 100%?"

It's hard to pin down the exact odds of the things I believe; I believe with near-100% certainty that I have two legs and that the Earth is roughly spheroid. But I can't really disprove that I'm a brain in a vat that's fed false sense impressions, so technically I'll have to admit that I could have no legs and that the Earth might be a cube or non-existent.

Going in a more abstract direction, I feel pretty darn certain that there can be true statements, as the converse leads to a paradox. Still, that seems to hinge on me having 100% confidence that my sense of logic is sound...

cautiouslycurious said...

Crude,
“Alright. So your subjective experiences as experiences is something you cannot be in error about, according to you? "I have the experience of looking at a monitor right now" - sure, you could be hallucinating, but you're you have 100% belief that you're having that experience?”

Right, even if I’m hallucinating, even if I’m just a brain in a vat, even if I’m in the Matrix, I’m still experiencing a ‘monitor’ in front of me. The experience is taken at face value, however, any interpretation of that experience is subject to error.

“How about the labels? You say you accept them - does that mean you don't think you could be wrong about a label?”

Right. I’m talking about labels that you would use yourself. When I define X as Y or label X as Z (e.g. I’m going to label that scruffy little thing ‘Fido’, I believe my name is “___”), I’m assigning that meaning and since that is part of sense experience, it is more or less undefeatable. However, labels can be part of empirical claims such as Crude defines X as Y, or X is commonly understood to be defined as Y and they can be subject to error.

“Is that just another way of saying what I said? I mean, my phrasing was "No, I don't believe I'm looking at a monitor. But I believe I have this very high chance that I'm looking at a monitor."? If there's a difference between the two, I'd like to hear what.”

The difference is in the actions. Suppose you have two people, one is an atheist and one is a theist. The atheist is your run of the mill atheist, and the theist goes to Church weekly, tithes, and prays each night. However, they both think that the probability of God exists is unlikely, but their differing beliefs lead to radically different actions.

“Do you 'believe it to be water' with some relevant confidence interval?”

I suppose you could say that, but I think it misses an important distinction. When I play poker, I don’t stay in a hand because I believe I will win that particular hand, I stay in because I think it is profitable to stay in. Its two completely different calculations going on that differentiate those two different beliefs. I suppose you could say that I believe that I will win a certain hand with %20 confidence, meaning that I expect to lose. However, I find it semantically clumsy to say that I am acting as if I’m going to win (which is somewhat already shaky) while expecting to lose at the same time. I would want to differentiate between someone staying in because they have calculated the pot odds and found it favorable versus someone who would stay in regardless because they believe they will win (which does not require a pot odds calculation) so I would avoid such phrasing.

Dan Gillson said...

Before bed last night, I was rereading through Peirce's essay "How To Make Our Ideas Clear", and I thought I'd share with you all a quote from it which I find to be pertinent to the topic at hand:

"And what, then, is belief? It is the demi-cadence which closes a musical phrase in the symphony of our intellectual life. We have seen that it has just three properties: First, it is something that we are aware of; second, it apeases the irritation of doubt; and third, it involves the establishment in our nature of a rule of action, or, say for short, a habit. As it appeases the irritation of doubt, which is the motive for thinking, thought relaxes, and comes to rest for a moment when belief is reached. But, since belief is a rule for action, the application of which involves further doubt and further thought, at the same time that it is a stopping place, it is also a new starting place for thought ... belief is only a stadium of mental action, an effect upon our nature due to thought, which will influence future thinking."

Crude said...

zillipede,

It's hard to pin down the exact odds of the things I believe; I believe with near-100% certainty that I have two legs and that the Earth is roughly spheroid. But I can't really disprove that I'm a brain in a vat that's fed false sense impressions, so technically I'll have to admit that I could have no legs and that the Earth might be a cube or non-existent.

Alright. So it doesn't sound like there's anything you believe with 100% certainty? (Maybe excepting, say, CC's example of experience?)

Going in a more abstract direction, I feel pretty darn certain that there can be true statements, as the converse leads to a paradox. Still, that seems to hinge on me having 100% confidence that my sense of logic is sound...

I'd say it seems similar to me too.

CC,

Right, even if I’m hallucinating, even if I’m just a brain in a vat, even if I’m in the Matrix, I’m still experiencing a ‘monitor’ in front of me. The experience is taken at face value, however, any interpretation of that experience is subject to error.

Okay, so you grant the cogito ergo sum thing with 100% certainty. Anything else?

I suppose you could say that I believe that I will win a certain hand with %20 confidence, meaning that I expect to lose.

Okay. What's the % confidence of your evaluation of your 20% confidence?