It is currently Fri Apr 20, 2018 4:52 am

All times are UTC - 7 hours [ DST ]




Post new topic Reply to topic  [ 192 posts ]  Go to page Previous  1 ... 4, 5, 6, 7, 8, 9, 10  Next
Author Message
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Sun Jan 07, 2018 8:15 am 
God
User avatar

Joined: Thu Feb 15, 2007 3:24 pm
Posts: 3818
Location: La Mancha
Johannes wrote:
I knew Hector. I worked for Hector. Hector was a friend of mine. And Dr Carrier....

Image

_________________
It’s relatively easy to agree that only Homo sapiens can speak about things that don’t really exist, and believe six impossible things before breakfast. You could never convince a monkey to give you a banana by promising him limitless bananas after death in monkey heaven.

-Yuval Noah Harari


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Sun Jan 07, 2018 11:46 am 
God
User avatar

Joined: Wed Feb 03, 2010 11:17 pm
Posts: 6969
Thanks for the detailed reply, Analytics. It was informative, independent of this discussion which I love.

I snipped out the majority of the post to focus on a couple of quick items I still had some question about -

Analytics wrote:
honorentheos wrote:
...what differentiates Carrier proposing an alternative universe from his making bad faith arguments?


The difference is that these are two independent topics. Whether he makes “bad faith arguments” has to do with his skill and honesty as a researcher. In contrast, the Bayesian “alternate universe” we are talking about is an intrinsic part of the question; To address the question “Did Jesus Exist?” using Bayesian reasoning, you need two universes—one where he exists and one where he doesn’t.

...

That is what I mean by saying those two issues are independent. Whether Carrier is honestly carrying out Bayesian reasoning in good faith is a different question than whether he should be taking a Bayesian approach at all.

I agree with both Nate Silver and Sean Carroll on the issue—the Bayesian approach to dealing with questions of uncertainty and improving knowledge is the correct approach. If somebody wants to effectively rebut Carrier, she should do it from within a Bayesian approach. The huge irony here is that the use of Bayes’ Theorem might be the only thing that Carrier got right.

Sean Carroll wrote:
Bayes’s Theorem is one of those insights that can change the way we go through life. Each of us comes equipped with a rich variety of beliefs, for or against all sorts of propositions. Bayes teaches us (1) never to assign perfect certainty to any such belief; (2) always to be prepared to update our credences when new evidence comes along; and (3) how exactly such evidence alters the credences we assign. It’s a road map for coming closer and closer to the truth.

I especially appreciated the thoughts Carroll's quote add to the discussion. But it left me wondering if part of what we are discussing with regards to the opening post falls between the cracks of the three points Carroll makes in the statement I quote above? Those being:

Bayes teaches us:

(1) never to assign perfect certainty to any such belief;
(2) always to be prepared to update our credences when new evidence comes along; and
(3) how exactly such evidence alters the credences we assign.


If we include Carroll's earlier comment that Bayesian approaches, "remind us that we assign prior credences, and update them appropriately, to every factual proposition that may or may not be true about the world", where do we also insert the human propensity to not actually update our prior credences but instead find new ways to defend them? For example, from this article in The New Yorker last February -
https://www.newyorker.com/magazine/2017 ... -our-minds

Even after the evidence “for their beliefs has been totally refuted, people fail to make appropriate revisions in those beliefs,” the researchers noted. In this case, the failure was “particularly impressive,” since two data points would never have been enough information to generalize from.

It seems we're still obligated to find ways to insert objectivity - against our nature when it conforms to our prior beliefs - into evaluating the priors in order to judge the results rather than just observing how cohesive a particular proposed universe may be. I'm guessing I'm missing something still?

_________________
The world is always full of the sound of waves..but who knows the heart of the sea, a hundred feet down? Who knows it's depth?
~ Eiji Yoshikawa


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Sun Jan 07, 2018 3:24 pm 
God
User avatar

Joined: Mon Feb 25, 2013 4:32 pm
Posts: 1071
Location: Parowan, Utah
Analytics wrote:
I agree with both Nate Silver and Sean Carroll on the issue—the Bayesian approach to dealing with questions of uncertainty and improving knowledge is the correct approach. If somebody wants to effectively rebut Carrier, she should do it from within a Bayesian approach. The huge irony here is that the use of Bayes’ Theorem might be the only thing that Carrier got right.


I hate to keep disagreeing with you, Analytics, but I find it hard to agree on this too because it is a general statement that fails to account for the contexts that actually make anything about the past knowable and known at all. And it reframes the issues incorrectly: this isn't a question of uncertainty in a vacuum but about how well either of two hypothesis fits with what we already know.

For so much of our discussion, I have felt that the Carrier skeptics and detractors have done a tremendous job—except for humble insignificant me—demolishing every claim of his that has been introduced here, using the methods of ancient historians. It's as if there are two disputants in front of an unopened fridge: A claims there are eggs inside the fridge, and B thinks not. B suggests that Bayes's theorem is the best way to find the answer, but A's solution is to open the fridge and point out that, in fact, there are eggs right there. B responds that A's glance into the fridge doesn't address the subtleties of B's original argument. That might be true but it is irrelevant, because A has established that there eggs in the fridge, which is what the original question was.

On the question of Jesus, we have a wide variety of disciplines and a substantial body of evidence to deal with: we can, in effect, just look in the fridge. If we didn't have any knowledge of Mediterranean antiquity in any aspect, or if within that knowledge we had no real evidence about Jesus (say, if all we had was the one line in Suetonius), then I could see how probability theorems could be very useful and I would say they might be the best way to go. But as it is, of the 60,000,000 or so people in the Roman empire in the first century BCE, there are probably less than 100 whose existence is better documented than Jesus's, if even that. Most that we do know about are more like the one line Suetonius. And we have vast bodies of evidence against which to way the evidence regarding Jesus that we have. That project has been going on for 200 years.

Two things then make me hesitant about claim you are making for the necessity of Bayesian approach:

1) any analysis can't be done in a vacuum; there should be some kind of baseline. That is to say, you can't quantify the evidence for Jesus's existence in the same way you would do so for Mark Twain's; it has been done in the same way we would do so with a contemporary of Jesus whose existence is established by a similar body of evidence. It has to account for context.

2) the results of any approach should not conflict with or call into question more immediate approaches that tackle the evidence head on when those approaches are otherwise unproblematic; e.g. if the result of Bayesian analysis implies that we need to rethink our understanding of a basic Greek verb that both synchronic and diachronic corpus-based studies have shown is straightforward (and that is just one example from hundreds, if not thousands), then the manner in which the analysis has been applied is flawed or the analysis itself is.

The implication of 2) of course is that there is really no point at all for using Bayes because it can only tell us what we already know. Indeed, it should do that in a certain sense: it should reinforce what we do know, since presumably its inputs have been quantified in such a way that "what we know" has been assumed to be true.

The point I have been trying to make over and over (and apparently without success) is that any hypothesis about Jesus has to fit what we do know (coheres, to use Stak's term) before we start weighing its probability against any other hypothesis—and that second hypothesis should also cohere with what we do know. Both must fit what we know.The problem with mythicist Jesus is that it does not cohere with what we do know. It therefore does not deserve consideration.

We know that Mark Twain, to analogize again, was born in the 19th century, and any hypothesis that contradicts that fact isn't even worth considering.

Why then do we need to rebut Carrier from within the framework of Bayesian probability? We can already demonstrate that Carrier's theory is nonsense. The only reason I can think would be to satisfy the demand of his supporters for whom the only threshold that matters is ostensibly a statistical one—emphasis on ostensibly.

_________________
"As to any slivers of light or any particles of darkness of the past, we forget about them."

—B. Redd McConkie


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Sun Jan 07, 2018 10:48 pm 
Sunbeam

Joined: Tue Jan 10, 2017 1:07 am
Posts: 49
Symmachus wrote:

For so much of our discussion, I have felt that the Carrier skeptics and detractors have done a tremendous job—except for humble insignificant me—demolishing every claim of his that has been introduced here, using the methods of ancient historians. It's as if there are two disputants in front of an unopened fridge: A claims there are eggs inside the fridge, and B thinks not. B suggests that Bayes's theorem is the best way to find the answer, but A's solution is to open the fridge and point out that, in fact, there are eggs right there. B responds that A's glance into the fridge doesn't address the subtleties of B's original argument. That might be true but it is irrelevant, because A has established that there eggs in the fridge, which is what the original question was.

On the question of Jesus, we have a wide variety of disciplines and a substantial body of evidence to deal with: we can, in effect, just look in the fridge. If we didn't have any knowledge of Mediterranean antiquity in any aspect, or if within that knowledge we had no real evidence about Jesus (say, if all we had was the one line in Suetonius), then I could see how probability theorems could be very useful and I would say they might be the best way to go. But as it is, of the 60,000,000 or so people in the Roman empire in the first century BCE, there are probably less than 100 whose existence is better documented than Jesus's, if even that. Most that we do know about are more like the one line Suetonius. And we have vast bodies of evidence against which to way the evidence regarding Jesus that we have. That project has been going on for 200 years.


Symmachus, I think your refrigerator analogy is actually a good example of the value of Bayesian analysis, which is just a fancy way of discussing how a given hypothesis "fits" the data. A key component of Bayesian analysis is to actually go and find data, both pro and con, so the person following the Bayesian formula would advocate opening the refrigerator and gathering data about the existence of the eggs.

Your analogy posits the Bayesian (person B) as merely performing an intellectual exercise where they imagine all possible permutations. Critically, that is not the process a Bayesian would follow. They would instead gather all the relevant data and determine which model/hypothesis best fits the data.

I agree with Analytics that Carrier was right to utilize Bayesian analysis, but Carrier critically fails within the process when, seeing that his model does not fit the data, he manufactures data in an effort to bolster his model. In doing so, he departs from Bayesian analysis and enters into apology.


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 12:05 am 
God
User avatar

Joined: Fri Oct 05, 2012 5:37 pm
Posts: 6590
Location: On walkabout
Symmachus wrote:
Analytics wrote:
I agree with both Nate Silver and Sean Carroll on the issue—the Bayesian approach to dealing with questions of uncertainty and improving knowledge is the correct approach. If somebody wants to effectively rebut Carrier, she should do it from within a Bayesian approach. The huge irony here is that the use of Bayes’ Theorem might be the only thing that Carrier got right.


I hate to keep disagreeing with you, Analytics, but I find it hard to agree on this too because it is a general statement that fails to account for the contexts that actually make anything about the past knowable and known at all. And it reframes the issues incorrectly: this isn't a question of uncertainty in a vacuum but about how well either of two hypothesis fits with what we already know.

For so much of our discussion, I have felt that the Carrier skeptics and detractors have done a tremendous job—except for humble insignificant me—demolishing every claim of his that has been introduced here, using the methods of ancient historians. It's as if there are two disputants in front of an unopened fridge: A claims there are eggs inside the fridge, and B thinks not. B suggests that Bayes's theorem is the best way to find the answer, but A's solution is to open the fridge and point out that, in fact, there are eggs right there. B responds that A's glance into the fridge doesn't address the subtleties of B's original argument. That might be true but it is irrelevant, because A has established that there eggs in the fridge, which is what the original question was.

On the question of Jesus, we have a wide variety of disciplines and a substantial body of evidence to deal with: we can, in effect, just look in the fridge. If we didn't have any knowledge of Mediterranean antiquity in any aspect, or if within that knowledge we had no real evidence about Jesus (say, if all we had was the one line in Suetonius), then I could see how probability theorems could be very useful and I would say they might be the best way to go. But as it is, of the 60,000,000 or so people in the Roman empire in the first century BCE, there are probably less than 100 whose existence is better documented than Jesus's, if even that. Most that we do know about are more like the one line Suetonius. And we have vast bodies of evidence against which to way the evidence regarding Jesus that we have. That project has been going on for 200 years.

Two things then make me hesitant about claim you are making for the necessity of Bayesian approach:

1) any analysis can't be done in a vacuum; there should be some kind of baseline. That is to say, you can't quantify the evidence for Jesus's existence in the same way you would do so for Mark Twain's; it has been done in the same way we would do so with a contemporary of Jesus whose existence is established by a similar body of evidence. It has to account for context.

2) the results of any approach should not conflict with or call into question more immediate approaches that tackle the evidence head on when those approaches are otherwise unproblematic; e.g. if the result of Bayesian analysis implies that we need to rethink our understanding of a basic Greek verb that both synchronic and diachronic corpus-based studies have shown is straightforward (and that is just one example from hundreds, if not thousands), then the manner in which the analysis has been applied is flawed or the analysis itself is.

The implication of 2) of course is that there is really no point at all for using Bayes because it can only tell us what we already know. Indeed, it should do that in a certain sense: it should reinforce what we do know, since presumably its inputs have been quantified in such a way that "what we know" has been assumed to be true.

The point I have been trying to make over and over (and apparently without success) is that any hypothesis about Jesus has to fit what we do know (coheres, to use Stak's term) before we start weighing its probability against any other hypothesis—and that second hypothesis should also cohere with what we do know. Both must fit what we know.The problem with mythicist Jesus is that it does not cohere with what we do know. It therefore does not deserve consideration.

We know that Mark Twain, to analogize again, was born in the 19th century, and any hypothesis that contradicts that fact isn't even worth considering.

Why then do we need to rebut Carrier from within the framework of Bayesian probability? We can already demonstrate that Carrier's theory is nonsense. The only reason I can think would be to satisfy the demand of his supporters for whom the only threshold that matters is ostensibly a statistical one—emphasis on ostensibly.


I'm not a whiz at Bayesian analysis, but your description and criticism of it doesn't jibe with my understanding of what it does. The purpose of the analysis, as I understand it, is to assess the relevance of piece of evidence in context with everything else we know -- not in some kind of vacuum. In fact, my understanding of one of the strength's of the methodology is that it forces one to examine the evidence in context. I think the description of comparing two hypothetical universes that has been used in the thread may be giving a sort of false impression of what the analysis does. I think the counterargument of you and others could be framed in Bayesian terms, but the point of contention would be, in my opinion, not the methodology but Carrier's application of it.

_________________
​“The ideal subject of totalitarian rule is not the convinced Nazi or the dedicated communist, but people for whom the distinction between fact and fiction, true and false, no longer exists.”

― Hannah Arendt, The Origins of Totalitarianism, 1951


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 8:15 am 
God
User avatar

Joined: Thu Feb 15, 2007 3:24 pm
Posts: 3818
Location: La Mancha
honorentheos wrote:
Thanks for the detailed reply, Analytics. It was informative, independent of this discussion which I love.

I snipped out the majority of the post to focus on a couple of quick items I still had some question about...


Hi honorentheos,

Excellent question. I don't think you are missing anything--you are seeing what the actual issue is.

The big underlying issue that I think you are getting at is how we are all saddled with a huge basket of cognitive biases, and we can't objectively evaluate what is going on in the world unless we can somehow get past this. Otherwise, whatever model we use for aggregating evidence, Bayesian or otherwise, will be dominated by GIGO (garbage-in-garbage-out).

Doing some probability problems with Bayes's Theorem can help you get insights into how our beliefs ought to be weighed and updated in an ideal universe where the probabilities associated with the evidence are given. In that sense, "seeing the light" and becoming a Bayesian disciple can "change the way we go through life," in the words of Carroll. What I'm getting at is that really understanding the genius behind Bayes's Theorem can help you get a little bit past a couple of the biases that plague human thought.

But it isn't a magic bullet. The big issues with cognitive bias still drive everything we put into the formula. To get valid results, we need to think very carefully about the evidence, be humble about it, and somehow think logically and objectively despite our biases. If you make an earnest effort to do from a Bayesian perspective, in theory you should get closer and closer to the truth.

_________________
It’s relatively easy to agree that only Homo sapiens can speak about things that don’t really exist, and believe six impossible things before breakfast. You could never convince a monkey to give you a banana by promising him limitless bananas after death in monkey heaven.

-Yuval Noah Harari


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 9:54 am 
God
User avatar

Joined: Mon Feb 25, 2013 4:32 pm
Posts: 1071
Location: Parowan, Utah
Res Ipsa wrote:
I'm not a whiz at Bayesian analysis, but your description and criticism of it doesn't jibe with my understanding of what it does. The purpose of the analysis, as I understand it, is to assess the relevance of piece of evidence in context with everything else we know -- not in some kind of vacuum. In fact, my understanding of one of the strength's of the methodology is that it forces one to examine the evidence in context. I think the description of comparing two hypothetical universes that has been used in the thread may be giving a sort of false impression of what the analysis does.


That description does sound different—and better—from my characterization, which I must emphasize is informed entirely by how I understand what is presented here. I may not be getting the full picture, and I may not grasp fully what picture I am getting; I have 0 training in statistics so I can only ask forgiveness for my ignorance. The idea of the two universes, however, is exactly what I found sort of ridiculous, because we don't have two universes for this question: there is just the one that has meticulously reconstructed by thousands of scholars. If inventing one means calling into question a lot of that work just under the false claim that they are equally probable pre-Bayesian analysis, I don't really see the point of it, especially when the few examples of Carrier I've seen adduced just show how inept he is. The historical picture we have is like a delicate ecosystem where one small alteration has the potential to affect the whole thing in drastic ways; to me and I'm sure to scholars who are familiar with how that historical ecosystem is constituted, the mythicist position is chaotically disruptive, and for no good reason beyond itself. In my fridge analogy, it would be like telling me that my eyes are mistaken, and that indeed I need to rethink what eggs really are, what a fridge is, and indeed what it means to see. Carrier's method, as I see it here and from snippets on his blog, is patently absurd. I'm happy to learn that that is reflection of his misapplication of Bayesian analysis and not a flaw in the analysis itself.

In general, I could imagine statistical analysis to be more useful if it were deciding between interpretations that competent scholars from across the different methodological zones of the field already agree are valid—if it doesn't pointlessly destroy the ecosystem. Mythicism doesn't fit the picture, but there are still different interpretations of the evidence within the larger general framework that has been reconstructed. For example, was Jesus a social revolutionary, a political insurgent, an apocalyptic prophet, or a miracle-worker? The usual way to escape this, for those who aren't committed to one or the other (and most today I think would say apocalyptic prophet) is to argue for some combination without specifying much beyond that. In other words, it's just a rhetorical trick to avoid tackling the issue more concretely. I can see that maybe a Bayesian approach would be useful in trying to settle this and other questions of this sort, and I can see that some of the evidence would lend itself well to statistical analysis. For instance, the word used for "bandit" to describe the two malefactors crucified on either side of Jesus does have instances in the corpus where there is a political connotation, but usually it is more vanilla than that. In any case, it's countable. Whether one interprets that word politically or not is, in my view, largely a matter of taste, and so it is not really ever settled; perhaps a Bayesian analysis could add some certainty to one position or the other. But whatever a Bayesian approach might reveal, it shouldn't force us to rethink whether we really know Greek or not.

_________________
"As to any slivers of light or any particles of darkness of the past, we forget about them."

—B. Redd McConkie


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 10:22 am 
God
User avatar

Joined: Thu Feb 15, 2007 3:24 pm
Posts: 3818
Location: La Mancha
Hi Symmachus,

Hoping to be clear, when I quoted the last line of the Iliad, my point was that the debate was over and you won.

However, I still agree with Nate Silver and Sean Carroll about the universal applicability of Bayesian thinking when dealing with probabilities. I sense that you feel I’m an arrogant outsider attacking your entire field when I say that, but that isn’t my intention. All I’m doing is agreeing with Sean Carroll that Bayes gives us a "universal scheme for thinking about our degrees of beliefs," and at this point I’m not willing to carve out ancient history as an exception to this. Doing so would be patronizing.

While the "do we have eggs?" question is trivial, I maintain that the most rigorous way to address it is in fact with a Bayesian approach.

Interestingly, when Nate Silver and Sean Carroll are emphasizing the universal applicability of Bayesian thinking, they both use the question, "is there a God?" as an example of something applicable. That question emphasizes the need to look at two hypothetical universes. According to Richard Dawkins, "no evidence for God's existence has yet appeared." Yet according to Alma, "all things denote there is a God." To make a good-faith effort at understanding what the evidence actually says when it can be interpreted so diametrically, we need to view the evidence from the perspectives of the paradigms we are actually attempting to evaluate. While you might sincerely believe that "all things denote there is a God," if you want to honestly address the question, you need to visualize a universe without a God and evaluate the evidence from that perspective as well. You might realize that the evidence says something different than what you thought it did.

Ironically, the way you suggest we address the topic of this thread instead of Bayesian analysis maps nicely onto valid Bayesian reasoning. If hypothesis B doesn’t cohere with what we already know, then we should give it a very low a priori probability, which not only will likely drive the ultimate results, but also serves as a rational justification for not giving the question serious attention in the first place.

So I suspect you may have been reading more into what I was saying than I intended. While taking a more explicit Bayesian approach might refine your actual underlying argument, it will also help clarify your point to the people you are trying to convince. What Carrier’s argument lacks in persuasiveness it makes up for in being systematic. If somebody wanted to show him how he is wrong, it could systematically be done, numbered point by numbered point, within an explicit framework which is, according to Sean Carroll, "the basis of all science and other forms of empirical reasoning."

_________________
It’s relatively easy to agree that only Homo sapiens can speak about things that don’t really exist, and believe six impossible things before breakfast. You could never convince a monkey to give you a banana by promising him limitless bananas after death in monkey heaven.

-Yuval Noah Harari


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 10:37 am 
God
User avatar

Joined: Thu Feb 15, 2007 3:24 pm
Posts: 3818
Location: La Mancha
A Bayesian Analysis of Eggs

I remember purchasing eggs last week, and I don't think anybody in the household has used them all yet. I'm 90% confident we have eggs. But being a fanatical Bayesian, I take the following approach to verify:

There are two hypothetical universes that are exactly equivalent in every way except one: in one universe, there are in fact eggs in the fridge. In the other, there are no eggs.

I open the door and see eggs. In a universe where there actually are eggs, seeing them happens 95% of the time (5% of the time the eggs are really there, but I’m absent-minded and the fridge can be a mess so I miss them). In a universe without eggs, I see them 0.1% of the time (once, my little brother used all the eggs but for some mind-boggling reason, put the carton back in the fridge filled with empty shells. Thus, it is possible to see something that isn’t really there).

Consequently, I plug these numbers into Bayes’s formula, and conclude that I am 99.999% certain that I do in fact have eggs.

_________________
It’s relatively easy to agree that only Homo sapiens can speak about things that don’t really exist, and believe six impossible things before breakfast. You could never convince a monkey to give you a banana by promising him limitless bananas after death in monkey heaven.

-Yuval Noah Harari


Last edited by Analytics on Mon Jan 08, 2018 12:17 pm, edited 1 time in total.

Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 11:55 am 
Anti-Mormon

Joined: Mon Apr 19, 2010 9:32 pm
Posts: 5123
Location: In the Politburo
Symmachus wrote:
Why then do we need to rebut Carrier from within the framework of Bayesian probability? We can already demonstrate that Carrier's theory is nonsense. The only reason I can think would be to satisfy the demand of his supporters for whom the only threshold that matters is ostensibly a statistical one—emphasis on ostensibly.


I think what is frustrating is that people will invoke "Bayesian Analysis" like it is tantamount to something like Form Criticism and don't understand just how opaque they are being. It seems like advocates seem to switch between an epistemology, a decision theory, and a confirmation theory without even acknowledging the problems with those equivocations. Simply employing a version of subjective probability function so you can play around with degrees of belief isn't anymore Bayesian than if you used classical syllogisms and insisted we all should be Aristotelians.


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 12:33 pm 
God
User avatar

Joined: Thu Feb 15, 2007 3:24 pm
Posts: 3818
Location: La Mancha
Bayesian Eggs, Part 2: A More Interesting Example

Again, I'm 90% sure I have eggs. I check the fridge, and I don't see any. That is the evidence.

Universe 1: There really are eggs. 5% of the time that we have eggs I don't see them. So in this universe, there is a 5% chance that I would see the evidence I do.

Universe 2: There are no eggs. When there are no eggs, 99.9% of the time I don't see eggs.

Plugging all of that into Bayes' Formula tells me that there is actually a 31% chance that we do have eggs. That might be considered surprising--I checked and didn't see eggs. That should settle it, right? I didn't see eggs, therefore no eggs.

However, Bayes' Formula gives us some insight. I need to consider how often I am wrong when I don't see eggs, and how confident I was that we had eggs to before I checked. Plugging all of that into the formula tells me that there is a fair chance we really do have eggs, and that I better check behind the leftover chili to see if they are there hidden.

_________________
It’s relatively easy to agree that only Homo sapiens can speak about things that don’t really exist, and believe six impossible things before breakfast. You could never convince a monkey to give you a banana by promising him limitless bananas after death in monkey heaven.

-Yuval Noah Harari


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 12:44 pm 
God
User avatar

Joined: Mon Apr 25, 2011 12:19 pm
Posts: 2986
Analytics wrote:
Consequently, I plug these numbers into Bayes’s formula, and conclude that I am 99.999% certain that I do in fact have eggs.


In my experience, a quantitative researcher is never happy unless a statistics chapter is included in someone else's qualitative findings.


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 12:47 pm 
God
User avatar

Joined: Fri Oct 05, 2012 5:37 pm
Posts: 6590
Location: On walkabout
Analytics, your second example is a good illustration of how Bayesian analysis forces one to consider evidence in context. The evidence is looking and not finding eggs. The context is how often in the past you missed eggs when you were looking for them. The analysis forces one to think about false positives — something we naturally don’t seem to do.

_________________
​“The ideal subject of totalitarian rule is not the convinced Nazi or the dedicated communist, but people for whom the distinction between fact and fiction, true and false, no longer exists.”

― Hannah Arendt, The Origins of Totalitarianism, 1951


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 12:51 pm 
God
User avatar

Joined: Fri Oct 05, 2012 5:37 pm
Posts: 6590
Location: On walkabout
By the way, I don’t think we need to think about two universes to apply the method. All we have to ask in the second egg example is “how often do I miss the eggs when I look for them?”

_________________
​“The ideal subject of totalitarian rule is not the convinced Nazi or the dedicated communist, but people for whom the distinction between fact and fiction, true and false, no longer exists.”

― Hannah Arendt, The Origins of Totalitarianism, 1951


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 1:41 pm 
God
User avatar

Joined: Thu Feb 15, 2007 3:24 pm
Posts: 3818
Location: La Mancha
Morley wrote:
Analytics wrote:
Consequently, I plug these numbers into Bayes’s formula, and conclude that I am 99.999% certain that I do in fact have eggs.


In my experience, a quantitative researcher is never happy unless a statistics chapter is included in someone else's qualitative findings.


I held that belief with a 74.2% confidence level until this thread. But after examining the evidence in two distinct hypothetical universes, I have updated my belief to 79.3%.

_________________
It’s relatively easy to agree that only Homo sapiens can speak about things that don’t really exist, and believe six impossible things before breakfast. You could never convince a monkey to give you a banana by promising him limitless bananas after death in monkey heaven.

-Yuval Noah Harari


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 1:54 pm 
God
User avatar

Joined: Thu Feb 15, 2007 3:24 pm
Posts: 3818
Location: La Mancha
Res Ipsa wrote:
By the way, I don’t think we need to think about two universes to apply the method. All we have to ask in the second egg example is “how often do I miss the eggs when I look for them?”

That's a fair point. My verbiage of "two universes" has obviously caused more confusion that enlightenment.

_________________
It’s relatively easy to agree that only Homo sapiens can speak about things that don’t really exist, and believe six impossible things before breakfast. You could never convince a monkey to give you a banana by promising him limitless bananas after death in monkey heaven.

-Yuval Noah Harari


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 2:29 pm 
God
User avatar

Joined: Fri Oct 05, 2012 5:37 pm
Posts: 6590
Location: On walkabout
Analytics wrote:
Res Ipsa wrote:
By the way, I don’t think we need to think about two universes to apply the method. All we have to ask in the second egg example is “how often do I miss the eggs when I look for them?”

That's a fair point. My verbiage of "two universes" has obviously caused more confusion that enlightenment.


In this case, I think it did. Take the evidence “the brother of the Lord.” Bayesian analysis wouldn’t tell us whether that phrase is more consistent with hypothetical Jesus and not-Jesus universes. It would look at other uses of the original Greek phrase to answer the question “when this word is used, what are the odds that it is intended to mean a biological brother?”
That tells us the weight we should put on that piece of evidence.

_________________
​“The ideal subject of totalitarian rule is not the convinced Nazi or the dedicated communist, but people for whom the distinction between fact and fiction, true and false, no longer exists.”

― Hannah Arendt, The Origins of Totalitarianism, 1951


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 4:14 pm 
God
User avatar

Joined: Mon Feb 25, 2013 4:32 pm
Posts: 1071
Location: Parowan, Utah
Analytics wrote:

Hoping to be clear, when I quoted the last line of the Iliad, my point was that the debate was over and you won.


Ha! I didn't know we were competing. :smile: Besides, I've never won anything in my life (although I once got an "I Voted" sticker after I voted).

Quote:
However, I still agree with Nate Silver and Sean Carroll about the universal applicability of Bayesian thinking when dealing with probabilities. I sense that you feel I’m an arrogant outsider attacking your entire field when I say that, but that isn’t my intention. All I’m doing is agreeing with Sean Carroll that Bayes gives us a "universal scheme for thinking about our degrees of beliefs," and at this point I’m not willing to carve out ancient history as an exception to this. Doing so would be patronizing.


It's not really my field (and see my response to Stak below), but I think most of the confusion is on my end. I don't really understand Bayesian analysis, but I do know when the philology and the history are bad.

Quote:
While the "do we have eggs?" question is trivial, I maintain that the most rigorous way to address it is in fact with a Bayesian approach.


Again, my lack of understanding is no inhibition: why can't we just open the fridge? It seems like a lot of work with the one approach and very little with the other. I guess it depends on what the "it" is that we are adressing: the fact of the eggs (or not) or our epistemological confidence in asserting the fact—whatever that is—of the eggs (or not)?

Quote:
Ironically, the way you suggest we address the topic of this thread instead of Bayesian analysis maps nicely onto valid Bayesian reasoning. If hypothesis B doesn’t cohere with what we already know, then we should give it a very low a priori probability, which not only will likely drive the ultimate results, but also serves as a rational justification for not giving the question serious attention in the first place.

So I suspect you may have been reading more into what I was saying than I intended. While taking a more explicit Bayesian approach might refine your actual underlying argument, it will also help clarify your point to the people you are trying to convince. What Carrier’s argument lacks in persuasiveness it makes up for in being systematic. If somebody wanted to show him how he is wrong, it could systematically be done, numbered point by numbered point, within an explicit framework which is, according to Sean Carroll, "the basis of all science and other forms of empirical reasoning."


On the point of systematicity, there is probably something aesthetically pleasing about that that renders one open to the possibility of persuasion that otherwise wouldn't exist. I feel the same way about Brian Stubbs's work: unlike Nibley, it attempts to establish systematicity. It is much more rigorous and "scientific" from a historical-linguistic perspective. Unfortunately, for the system to be real, it would require a massive renovation of a few sub-branches of linguistics, to say nothing of archaeology, and then ultimately even small little things like physics and biology—which is of course something that classical Mormon apologists understand on a case-by-case basis and do all the time, and which is one of the preeminent reasons they make themselves ridiculous.

Quote:
I think what is frustrating is that people will invoke "Bayesian Analysis" like it is tantamount to something like Form Criticism and don't understand just how opaque they are being. It seems like advocates seem to switch between an epistemology, a decision theory, and a confirmation theory without even acknowledging the problems with those equivocations. Simply employing a version of subjective probability function so you can play around with degrees of belief isn't anymore Bayesian than if you used classical syllogisms and insisted we all should be Aristotelians.


That encapsulates a lot of my frustration, and with a great deal more compression and precision than I can muster. Because the discussion about Bayesian analysis is operating at a much higher order than I what I'm talking about and able to talk about (universal topics like epistemology vs. mundane things like linguistics and cultural history), there is an imbalance: whatever I or Kish or Johannes or others have to say about Greek doesn't have any implications for Bayesian analysis, but the conclusions on this question drawn by Carrier via that analysis will require us to bracket things we know to be the case in the service of a higher epistemological claim. It's this balance between the universal and the mundane, not an interest in protecting the boundaries between intellectual disciplines that make equally far-reaching claims, that motivates the frustration in some of my responses.

_________________
"As to any slivers of light or any particles of darkness of the past, we forget about them."

—B. Redd McConkie


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 4:41 pm 
God
User avatar

Joined: Fri Oct 05, 2012 5:37 pm
Posts: 6590
Location: On walkabout
Symmachus wrote:

Again, my lack of understanding is no inhibition: why can't we just open the fridge? It seems like a lot of work with the one approach and very little with the other. I guess it depends on what the "it" is that we are adressing: the fact of the eggs (or not) or our epistemological confidence in asserting the fact—whatever that is—of the eggs (or not)?


Ah, I think that indicates some of the disconnect. Bayesian analysis helps us decide how important evidence is. It says nothing about gathering additional evidence. Take Analystics' first eggs example. If he were at the store trying to decide whether to buy eggs, the refrigerator door isn't there to be opened. So, he could work through this reasoning and conclude that he shouldn't buy eggs.

Take his second example, again at the store. He remembers that he opened the refrigerator that morning and didn't see eggs. Given the rate of false positives, he might conclude that he should buy eggs, but maybe only a dozen instead of two.

If Analytics were standing in front of the refrigerator door, nothing in Bayesian analysis would discourage him from opening it to look for eggs.

_________________
​“The ideal subject of totalitarian rule is not the convinced Nazi or the dedicated communist, but people for whom the distinction between fact and fiction, true and false, no longer exists.”

― Hannah Arendt, The Origins of Totalitarianism, 1951


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 4:43 pm 
God
User avatar

Joined: Thu Feb 15, 2007 3:24 pm
Posts: 3818
Location: La Mancha
Symmachus wrote:
Analytics wrote:

Hoping to be clear, when I quoted the last line of the Iliad, my point was that the debate was over and you won.


Ha! I didn't know we were competing. :smile: Besides, I've never won anything in my life (although I once got an "I Voted" sticker after I voted).

Well, you've managed to teach me something and make me change my mind. People have been given trophies for less.

Symmachus wrote:
Quote:
While the "do we have eggs?" question is trivial, I maintain that the most rigorous way to address it is in fact with a Bayesian approach.


Again, my lack of understanding is no inhibition: why can't we just open the fridge? It seems like a lot of work with the one approach and very little with the other. I guess it depends on what the "it" is that we are adressing: the fact of the eggs (or not) or our epistemological confidence in asserting the fact—whatever that is—of the eggs (or not)?


I swear I've had this exact same conversation with the folks who are now on the D&D board, and they said almost exactly the same thing. "Why go through all that convoluted Bayesian analysis when I can just ask God if it's true? What could be stronger proof than that?"

The point is that depending upon the question, it is possible to get stuck in a wrong paradigm where the evidence as seen from within reinforces the misunderstanding. A Bayesian methodology gives you a logically cohesive approach to evaluate two competing paradigms. Beyond that, it provides a general structure to weigh and interpret new evidence and properly update your views to reflect it.

_________________
It’s relatively easy to agree that only Homo sapiens can speak about things that don’t really exist, and believe six impossible things before breakfast. You could never convince a monkey to give you a banana by promising him limitless bananas after death in monkey heaven.

-Yuval Noah Harari


Top
 Profile  
Reply with quote  
 Post subject: Re: Craig Evans v Richard Carrier Did Jesus Exist SUPERB
PostPosted: Mon Jan 08, 2018 11:01 pm 
Anti-Mormon

Joined: Mon Apr 19, 2010 9:32 pm
Posts: 5123
Location: In the Politburo
Symmachus wrote:
Because the discussion about Bayesian analysis is operating at a much higher order than I what I'm talking about and able to talk about ...


To be blunt, Carrier's relationship to Bayes is that of a first year Greek student's to Aeschylus. He doesn't have the training or the practice and the result is tremendously inadequate for the task he sets out to do.

Symmachus wrote:
Because the discussion about Bayesian analysis is operating at a much higher order than I what I'm talking about and able to talk about (universal topics like epistemology vs. mundane things like linguistics and cultural history)...


It doesn't have to be, Text Criticism isn't really my bag but from what I'm given to understand about the "Coherence Based Genelogical Method" coming out of Germany, looks far more like Carrier's discussion should.

Symmachus wrote:
...there is an imbalance: whatever I or Kish or Johannes or others have to say about Greek doesn't have any implications for Bayesian analysis, but the conclusions on this question drawn by Carrier via that analysis will require us to bracket things we know to be the case in the service of a higher epistemological claim. It's this balance between the universal and the mundane, not an interest in protecting the boundaries between intellectual disciplines that make equally far-reaching claims, that motivates the frustration in some of my responses.


I think this is the result of people acting like this:

Image

...is some sort of universal formula that you can bring to any context or problem and it'll serve you well. Problem is that the above (rather recent) formula was trying/is capturing inverse inferences; predicting discreet events in contrived chance conditions in terms of probability. Come the 20th century, the good Reverend's definitions, propositions, lemmas, and assorted scholium get repurposed by philosophers and scientists trying to sketch out what is the exact nature of scientific theories.

See, everything Carrier uses was developed in response to growing dissatisfaction to the idea that scientific theories were just expressions of axiomatic theories we could formulate in mathematical logic, complete with "theoretical vocabularies" that had to correspond to "observational vocabularies" with a certain amount of "cognitive significance" that had to be met. This is why prior probability gets a recast as "background knowledge" and we get talk of "evidence" updating "degrees of belief" instead of straight conditionalization.

What I'm trying to drive home is that none of the development had ancient history and philology in mind, so if someone wants to do that they have to go back and make major additions to account for a totally new purpose. Instead of trying model a relationship between a given hypothesis and scientific observation via instruments, we have to work with ancient textual witnesses, testimonia, and the fragmented remains of material culture.


Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 192 posts ]  Go to page Previous  1 ... 4, 5, 6, 7, 8, 9, 10  Next

All times are UTC - 7 hours [ DST ]


Who is online

Users browsing this forum: Google [Bot] and 15 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Jump to:  
Revival Theme By Brandon Designs By B.Design-Studio © 2007-2008 Brandon
Revival Theme Based off SubLite By Echo © 2007-2008 Echo
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group