A Bayesian approach to informal argument fallacies


by Scientia Salon

This paper, which appeared in the journal Synthese in 2006, touches on a sacred cow of internet discourse, especially within atheist and skeptical communities: the idea of informal logical fallacies. I have a paper on the same topic currently in press, together with my co-authors Maarten Boudry and Fabio Paglieri, so we will return to the issue once our paper will be out (Maarten is writing a precis for Scientia Salon, and you’ll see that we advocate an even more radical approach to informal fallacies than the present authors). Meanwhile, below is a taste of what Ulrike Hahn and Mike Oaksford wrote. You can find the full paper here (free).

We examine in detail three classic reasoning fallacies, that is, sup- posedly “incorrect” forms of argument. These are the so-called argumentam ad ignorantiam, the circular argument or petitio principii, and the slippery slope argu- ment. In each case, the argument type is shown to match structurally arguments which are widely accepted. This suggests that it is not the form of the argu- ments as such that is problematic but rather something about the content of those examples with which they are typically justified. This leads to a Bayesian reanalysis of these classic argument forms and a reformulation of the conditions under which they do or do not constitute legitimate forms of argumentation.


Categories: notable

Tags: , ,

85 replies

  1. brodix: “… logic is generally quite nebulous and the truth often subjective. So “seizing the altar” and “herding the believers” is about power and politics, not truth and logic.”

    Isn’t that what phallogocentrism* holds? 🙂

    * http://en.wikipedia.org/wiki/Phallogocentrism


  2. I was not entirely happy with the article.

    First, I am unconvinced by the effort to identify fallacious structures with valid forms:

    “Ghosts exist because no one has proved that they do not.”

    “Drug A is not toxic because no toxic effects were observed” (following appropriate tests).

    I’m sorry, am I missing some professional logician’s curlicues here? Because I don’t see how a conclusion derived from an empirical test could ever be a kind of ‘argument from ignorance’ that presupposes that if a test is not possible, an asserted entity must be presumed to exist. The two structures can be related in only the emptiest way.

    This leads to what I think are indiscriminate claims, such as:

    “‘We should have sympathy for the wretched situation of this person P, therefore we ought to accept the conclusions of the argument that P maintains’. (Walton 1998)
    This argument again poses difficulties because, though it appears fallacious, sound examples exist, e.g., appeals for donations for research to cure a crippling childhood illness (Walton 1998). Here it seems justified to arouse pity in the potential donor.”

    No, Walton’s example is actually an enthymeme, and it might or might not be fallacious, once we unpack the hidden premises the author believes the audience assumes without direction.

    I’ve deep suspicion that this is yet another effort to utilize logic to control rhetoric. Logicians, philosophers, scientists, other rational (and rationalist) thinkers have been trying to bring rhetoric to heel since Socrates’ confrontation with Protagoras. The effort has contributed to the development of civilization, and of civil society. However, rhetoric cannot be easily contained by logic, because it has in its arsenal appeals to the non-rational. That means that the fallacies a rhetorician deploys successfully can not be dis-empowered by reshaping them as stronger arguments. At best logic can be used to reveal their fallacious nature. But the better strategy is to rebut them with stronger rhetoric appealing to the same emotions they do – in the process, reconstructing previously held beliefs along more reasonable lines. (Think of the great addresses by Darrow, Lincoln, King – master rhetoricians who helped change opinions and make history in the process.)

    Psychology can tell us how rhetoric works, but it can never give a complete taxonomy of this, for the simple reason that successful rhetoric is historically contingent. There are some base motivations a clever rhetorician can always appeal to, but he/she needs to ‘keep it fresh’ for the current generation. Failures to realize this have filled foot-notes in histories of politicians failing at the polls.

    I quote a brief essay of mine to make the broader point: “Because the notion that language developed as simply a means of communication is probably false. Its function seems more likely always to have been modification of behavior, either our own or that of others. Rhetoric is not the surplus of grammar and logic, it lies at their very origin.” (https://nosignofit.wordpress.com/2015/01/15/the-truth-of-rhetoric/)

    Liked by 2 people

  3. I know historicity of Jesus is somewhat off-topic (although perhaps not as much as energy flows), but regarding that discussion:

    A frequent problem appears to be that one side says, it seems more probable than not that there was a failed, small-scale doomsday prophet at the root of the cult, a Harold Camping of his time with perhaps only a few dozen followers before his death, and so he would not have registered in the history books. And then the Carrier et al. side says, ha, if such a person had existed then Roman and Jewish historians should have mentioned all the wonders, like zombies walking around after the crucifixion. Yes, I have had discussions like those, e.g. over at WEIT. Many mythicists are straw-manning everybody else as believing in a supernatural son of god with thousands of followers although that belief seems to be a minority position even among the more educated of the European Christians today.

    For what it is worth – and it isn’t much because I am not a historian – I think that the argument from silence does not work well against the hypothesis of a rather minor, mundane doomsday preacher. I also have yet to see a convincing explanation for the gospels, including the apocrypha, which treat Jesus as a human walking around Galilee and suchlike and collect sayings that somebody must have said at some point. (I would compare that to the Shakespeare authorship discussion: If the person with the Christian name William Shakespeare didn’t write the plays, somebody else still wrote them, and that person is then the author known as ‘Shakespeare’ in every way that matters. Likewise, whoever came up with all those teachings summarised in the earliest saying-gospels was, for all practical purposes, ‘Jesus’, even if their real name was different.) Bizarrely, the best shot of mythicists seems to be to consider the gospels fantasy novels at a time when there was no tradition and concept of fantasy novels.

    On the other hand, if Paul really does not seem to have known that Jesus was supposed to be a person who did anything described in the gospels then that is a very strong argument in favour of mythicism, but as far as I can tell it depends a bit on how to interpret and translate his writing. Also, he basically took over and transformed a growing cult while often being at loggerheads with other members. He may simply not have cared as much about the life of the founder as they did, focusing instead on making it palatable to philosophy-educated Greeks, something where cursing of fig trees and railing against money-lenders would just get in the way. I mean, if you were trying to spread a new religion among educated but slightly silly Americans, would you dwell on the partly embarrassing details of the life and slips of the tongue of the Indian guru who founded it or wouldn’t you rather stress the spiritual and self-help side of the teachings?


  4. ejwinner: “function [of language] seems more likely always to have been modification of behavior, either our own or that of others”

    That is the focus of the Conferences on Computational Models of Argument*: “non-monotonic and defeasible reasoning, deliberation about action, and agent communication scenarios such as negotiation”
    * http://comma.csc.liv.ac.uk/


  5. Coel, it matters not whether the 0.0008 is a low end, or a precise number in general. Per Aravis, that’s not how you do history — or any other of the humanities. Bayesian probabilities or anything else, you simply cannot be that precise with history. And, you know that.

    Let’s put it this way. Carrier has a Ph.D. in ancient history. Whether I phrased as just 0.008 or per you:

    “The probability that Jesus existed is somewhere between 1 in 12,500 [the 0.008%] and 1 in 3. In other words, less than 33% and most likely nearer to zero. We should conclude that Jesus probably did not exist”

    But, instead, said that about, Anaximander, Pythagoras, or another of the pro-Socratics, or about Homer, he would laugh in my face, and so would you. I know Aravis or Massimo would.

    But, because it’s about Jesus, Jesus denialism, and Gnu Atheism, such utter rot, to use a good old British term, is acceptable, eh?

    Well, no, it’s not.

    The rest of your opinion is just that — an opinion. And, it may become more “mainstream” among Gnu Atheists. That doesn’t make Carrier any more accurate than Dr. Andrew Wakefield.

    The “argument from silence” is not done sensibly by them. Again, if I used the argument from silence on classical history the way Carrier does on Jesus, again, you and he would laugh at me. But, because it’s about Jesus, Jesus denialism, and Gnu Atheism, such utter rot, to use a good old British term, is acceptable, eh?

    Well, no, it’s not.

    As for the rest of your comments, again, you’re not a Biblical scholar, and neither is Carrier, and you continue to prove that with vague comments about “Paul’s letters” that I know are wrong just as easily as an Ehrman knows are wrong.

    And, also per Aravis, my undergrad degree was in classical languages and history, so, yes, I know you don’t do history that way. (As I told Massimo in an email, the first writing I ever read on free will was in an independent study on Augustine, which included his tractate on free will.)


    To complete the snark, I await Ted Cruz or somebody even worse among US “birthers” using Bayesian probabilities the way Carrier does to “prove”:

    The probability that Barack Obama was born in the United States is somewhere between 1 in 12,500 [the 0.008%] and 1 in 3. In other words, less than 33% and most likely nearer to zero. We should conclude that Obama probably was not born in the United States, but was born in Kenya.

    Yep, lies, damned lies and misuse of Bayesian probabilities.


  6. SocraticGadfly,

    I don’t find Carrier’s style particularly likeable, I lean towards historicism myself, and I am pretty fed up with the constant all around Bayes-worship these days, but I think you are being a bit unreasonable here. “That’s not how you do history” is not really an argument when Carrier’s position is that that is how you should do history. Funnily enough I have seen the same logic before: my PhD supervisor once submitted a paper suggesting a new method to do X, and it was trashed by one of the peer reviewers not because they could find a flaw in the new method but because… he didn’t use the old-established method to do X instead. Way to miss the point.

    Further, I second what Coel wrote about the precision. In calibrated phylogenies in my area we always see things like “these two lineages diverged 8.2-13.7 million years ago” or “the Bayesian Posterior Probability for that clade is 0.96”, and nobody complains that you can’t know it to the degree of precision that is, say, 8.2 million years – because everybody understands that +/- 2.25 million years and only 0.96% certainty are the opposite of a claim to unrealistic precision, just like 0.08 to 33%.

    Also, in what sense is Carrier not a Biblical scholar? He is said to have got a PhD in ancient history and writes about little else but Biblical scholarship and possible misinterpretations of old Aramaic words. Does it only count as Biblical scholarship if one is a believer?


  7. Well, all I can say is it’s just like Jesus to find a way to get a mention regardless of the topic. 🙂

    Liked by 2 people

  8. Alex And Carrier’s argument is wrong. Ergo, it’s not how you do history, and it’s not how you should do history, either. Per Aravis, it’s more scientism.

    And, you then go on to make our point. Tracing separation of lineages in biology is far and away from what Carrier’s doing. And, I’ll put that rhetorical question to you, too: If Carrier were doing this with Homer or Anaximander, do you think he would have even the small in-group credibility he does now because he’s doing it with Jesus?

    And, while he may comment on misunderstanding of old Aramaic words, I see no information that he has any knowledge of Aramaic or Hebrew on his quite extensive CV, which speaks only about the Greco-Roman world in general. I would think that, if he actually knew Aramaic, as long as his CV is, he’d explicitly mention it.

    HAH! Argument from silence. But this one IS in a moderate amount. Anyway, read away. Does any nonbiased observer, looking at this, think he knows Aramaic? http://www.richardcarrier.info/cv.pdf

    Beyond that, I even did a Google search: “Does Richard Carrier know Aramaic?” And I can’t get any hits that will confirm that he does.

    Assuming he does not, the fact that he would still think to comment on misunderstandings of old Aramaic words “goes to character,” your honor. And, that’s putting it politely.

    But, places where he calls a Targum an “Aramaic translation of the Old Testament” show he’s no biblical scholar. http://religionatthemargins.com/2012/04/the-death-of-richard-carriers-dying-messiah/

    Fuller quote: “A Targum is an Aramaic translation (or paraphrase or interpretation) of the OT. So really, this is akin to a textual variant for this passage.” http://richardcarrier.blogspot.com/2011/10/dying-messiah.html

    Targums, as actual scholars know, were far more than that. They were commentaries, exegesises and more.

    And, click that first link. It’s clear that not only does he not know Aramaic, but that he just doesn’t know the bible that well, especially the Tanakh or Christian Old Testament, especially when he’s engaged in quote-mining and gets caught.

    Beyond that, Alex, this?

    He … writes about little else but Biblical scholarship and possible misinterpretations of old Aramaic words.

    I’m not even sure what logical fallacy that should be named, but it’s definitely a fallacy.

    There are people who write about nothing other than how the Earl of Oxford wrote Shakespeare. Do you call these people “Shakespearean scholars”?

    And, no, I never said one had to be a believer to be a Bible scholar. One of the best today, Bart Ehrman, is an agnostic.

    To extend another analogy to US politics, Gnu Atheists defending the scholarship of Richard Carrier is like Democratic muckety-mucks defending the transparency of Hillary Clinton.

    Anyway, while I was primarily answering Alex, that should also be information to other people don’t automatically think Richard Carrier is the greatest thing since sliced bread, but are curious about him.


  9. Re: Richard Carrier and his absurd application of probability theory to antiquities, I thought this post says it very well.


    The opening salvo is priceless, and spot-on:

    “To begin with, it’s illustrative to note who uses Bayes Theorem to analyse history and who does not. In the first category we have William Lane Craig, the conservative Christian apologist, who uses Bayes Theorem to “prove” that Jesus actually did rise from the dead. And we also have Richard Carrier, the anti-Christian activist, who uses Bayes Theorem to “prove” that Jesus didn’t exist at all. Right away, a curious observer would find themselves wondering how, if this Theorem is the wonderful instrument of historical objectivity both Craig and Carrier claim it to be, two people can apply it and come to two completely contradictory historical conclusions. After all, if Jesus didn’t exist, he didn’t do anything at all, let alone something as remarkable as rise from the dead. So both Carrier and Craig can’t both be right. Yet they both use Bayes Theorem to “prove” historical things. Something does not make sense here.

    Then if we turn to who doesn’t use Bayes Theorem to analyse history we find this category includes … pretty much every single historian on the planet. Again, this should strike the objective observer as distinctly odd. After all, if Bayes Theorem can genuinely be applied to determine the truth or otherwise of a historical event or proposition, it’s exceedingly strange that thousands of historians all over the world are not applying this remarkable tool all the time. Richard Carrier maintains that this is because every historian on earth, except him, is too ignorant and mathematically illiterate to understand the wonders of this remarkable tool and only he has been clever enough to realise that it can be applied to history. Given that Thomas Bayes ‘ theorem was first published in 1763, our objective observer would be forgiven for finding it remarkable that no-one noticed that it could be used in this way until Richard Carrier, an unemployed blogger, came along.”

    The rest of the take-down is even better. Very much worth a read.

    Oh, and by the way, I have no idea whether Jesus existed and being Jewish, I don’t care. (I might even get a little malevelont pleasure, if it turns out he didn’t.) What I do know, however, is that it can’t be proven one way or another, using mathematics.

    Liked by 5 people

  10. I prefer to make comments that I have prepared in advance and worked on; but I feel I must interject in the whole Carrier/Jesus question, exactly because I’m a non-theist (as are Socratic and Aravis, we should remember), who yet has no interest in the topic.

    Whether Jesus as a living human once existed or not is *completely irrelevant* to any question concerning his possible divinity.

    Atheists do not need this point to make their case; and, it must be said, theists will not be impressed with it no matter how well-grounded it might be.

    So the question doesn’t have to do with Jesus. Carrier, or the a/theism debate. The question really has to do with, can Bayesian probability theory be applied to the study of history?

    On that I remain agnostic; it could be. But according to some, Carrier hasn’t made that case. Not having read Carrier (and not being very interested in reading him), I leave my judgment suspended.

    What we could use here, then, is a stronger, less controversial consideration of that topic (probability and the study of history), which could be very interesting, if well handled. IMHO.


  11. brodix: “… So “seizing the altar” and “herding the believers” is about power and politics, not truth and logic.”

    Agree 100%. But, does it make any difference for whether the “dishonesty fallacy (DF)” being a logic one or not? The one which is worse than the DF is by viewing something wrong as Gospel. The GR (general relativity) is one example.

    If GR Is Wrong, in What Way Is It Wrong? It is not wrong as a theoretical physics-framework, as many of its consequences are empirically verified. It is wrong as a PART of the final-physics. Here are its BIG shortcomings.

    BS1, cannot be assimilated to QM (quantum mechanics):
    It is not too difficult to show that this universe arose with only one PATHWAY. Thus, there are three possibilities.

    P1, GR and QM are two expressions of the same thing.

    P2, one is master, the other squire.

    P3, one is right, the other wrong.

    BS2, GR plays zero role in the construction of quark/lepton. And this was discounted as the gravitation INTERACTION being too weak to play a role. Yes, this explanation works for the Newtonian gravity which sees it as an interaction. But, no, this must not be the case for GR. The GR gravity is not an interaction but is a geometrical structure of the space-time sheet. If quark/lepton are confined in/on the space-time sheet {similar to a mouse roaming under a bed sheet}, they must be described by (related to) that space-time sheet. The fact of BS2 shows that something (SM model or GR) is wrong, and GR could be at the losing end this time.

    BS3, GR plays zero role in the calculations (derivations) of Nature constants. In order to do these calculations, the precise answers for {what time is / what space is} must be known. And, GR’s description of Space-time is useless. Only at this point, I know that GR is totally wrong to play any PART in the final-physics.

    We would still like to know where went wrong. The accuracy for the equivalence principle is now at one part in 10 trillion, not yet enough to silence the challenge from the M-string theory. But, no, any discrepancy (if any) on the equivalence principle will not be the source for the wrongness of GR, which is at conceptual and fundamental level.

    Newton viewed gravity as the ‘interaction’ between two massive objects (apple and Earth), and its calculation is based on {sweeping out the same “area” from their center-of-mass}. This is a total Geometry description; that is, timeless {leads to instantaneity and simultaneity (IaS)}.

    GR views the gravity as the geometry of spacetime (timed, thus totally causal), and this is wrong as the Nature does encompass the non-causality. See, https://scientiasalon.wordpress.com/2015/01/16/reductionism-emergence-and-burden-of-proof-part-ii/comment-page-1/#comment-11037 .

    As the IaS is the core essence of Nature, the Newton view is a better one, but still not the correct one {unable to derive the Nature constants}.

    Liked by 1 person

  12. If I counted correctly, this is my last comment, so from now on I will not be able to pursue this further.

    What ejwinner just wrote is exactly correct: I do not particularly care about the existence of an actual doomsday preacher at the root of Christianity either, and the real question is whether Bayesian reasoning can be applied to history.

    And this is just why Socratic’s last two comments (and Aravis’) leave me a bit exasperated. Leaving aside the “all human reasoning is Bayesian” hyperbole that one often hears, it should still be obvious that we use informal Bayesian reasoning every single day, and that numerous areas of scholarship and science use it formally every day. Therefore I cannot think of any possible reason why one would believe that it cannot be used in history, as one of several available tools. The outright rejection displayed here seems to be based on some mixture of distrust of the actual percentages used and the fear that once one starts to use less intuitive methods scientism will destroy all that is good and decent. The former is easily solved by doing a sensitivity analysis, ie. putting in very different values and observing if the results change. The latter is just silly.

    What is particularly odd is that the study of history is actually extremely similar to the study of evolutionary and biogeographic history. One uses buried artefacts and ancient documents, the other uses buried fossils and DNA sequence data, in both cases to infer if and when and why certain events happened in the unobservable past. You know, the past luminaries of my field could have refrained from developing phylogenetic analyses (Bayesian and otherwise), models of character evolution and tests of neutrality. We could also still be sitting there and do systematics and evolutionary biology as Haeckel did, looking ponderously at a specimen and saying, “I think it goes there on the tree of life because the jawbone is shaped similar to that of the weasel”.

    And of course there are people who would prefer just that, who are offended when somebody examines one of their hunches with what they call a “mechanistic” analysis producing actual p, Bootstrap, Jackknife, Bremer support or BPP values. They’d rather go back to the time when people would rely blindly on their personal status in the community and their intuition and not ask to loudly if there is any way of hypothesis-testing so that we know when they have fallen prey to their biases. Luckily, they are a tiny and disappearing minority, otherwise biological systematics would still be a popularity contest crossed with a shouting match instead of a science.

    Now I am not saying that you could run every sub-field of the humanities like that. But history? The parallels should be obvious. And of course linguists, for example, are already using some of the same methods.

    Liked by 1 person

  13. Hi Coel ,
    I initially found carrier attractive but I found this response from Bart Ehrman pretty damning….

    Liked by 1 person

  14. Hi Socratic,

    … you simply cannot be that precise with history.

    Carrier’s conclusion is that the probability is somewhere between one in 12000 and one in 3. That is a range of about four orders of magnitude. How is that a claim to unwarranted precision?

    The rest of your comment is really just sneering at mythicists rather than anything substantive. That, of course, is entirely the norm, which is why it’s going to be fun watching as mythicism goes mainstream.

    Hi Aravis,

    … this is precisely how *not* to do history.

    All evaluations of likelihood by historians are essentially Bayesian, based on general knowledge about how things were (“priors”) and evaluation of specific pieces of evidence.

    Applying Bayesian formalism is merely a formalisation of the argument. Colloquially stated arguments often have implicit assumptions in them, while Bayesian formalism makes them explicit. The OP is an example of formalising colloquially stated arguments and evaluating them [to make a nod to the OP and not get too off-topic :-)].

    As for the review you said was “priceless, and spot-on”:

    “… how, if this Theorem is the wonderful instrument of historical objectivity both Craig and Carrier claim it to be, two people can apply it and come to two completely contradictory historical conclusions.”

    Bayesian formalism is merely a method for making implicit assumptions and claims explicit. That’s all. Craig and Carrier arrive at different answers because they make different assumptions and claims. The advantage of Bayesian formalism is that the assumptions are then explicit, and thus one can straightforwardly evaluate them.

    Really, the piece you quoted is not “spot-on”, it’s a complete misunderstanding of why Bayesian formalism is appropriate and useful.

    Now, I might agree that in many areas of historical research Bayesian formalism might not, in practice, be much use. That’s because, on topics such as Homer or Alexander, the historians are likely already making pretty sound assessments of evidence.

    But the entire field of Jesus-studies has so long been steeped in apologetic tradition that the entire mainstream is way out of kilter with the actual evidence. Thus, formalising the analysis and making implicit assumptions explicit is a good way of demonstrating the discrepancy.

    Hi Alex SL,

    The standard mythicist case is that the earthly character of Jesus was made up by the author “Mark” (though we have little idea who he was), writing in the post-AD71 apocalypse as a theological commentary on the destruction of the Temple by the Romans (note how he explicitly links the destruction of the Temple to Jesus’s death).

    The divine character of Jesus was already there, in the Old Testament. To see where Paul and “Mark” got such ideas from, just follow the literally hundreds of quotes and references to the OT in Paul’s letters, the gospels and the book of Hebrews, etc.

    The idea of humans making up characters and speech is hardly outlandish — just walk in to any good book shop and note the thousands of examples of such lining the shelves.

    Further, the canonical Bible writings are only half the total, there are just as many other early-Christian writings (see here), different versions of the gospels, that even the Christians accept are just made up as storification of theological themes related to Jesus!

    So the mainstream Christian scholars say, yes, those early stories were just made up, but these ones are based on real historical events . . . and we know that because … well, back around 300 AD, long after anyone would actually know, we, um, decided it.


  15. Massimo,

    There’s another strange aspect of the article. On p. 710 the authors show in a diagram that P(-C/-e) and P(C/e) converge to 1 if you apply the Bayesian formula repeatedly.

    A few back-of-the-enveloppe calculations suggest very strongly that this is a general mathematical property of this type of formula, as long as P(C), the sensitivity, the specificity etc. are between 0 and 1. The magnitude of the priors P(C) and P(-C) has no influence on this result, except that convergence takes a bit longer.

    This is odd. Say I confront John with the statement “God exists, because there’s no evidence that he doesn’t exist.” If I repeat this statement often enough, and John adjusts his prior belief each time according to the Bayesian formula, his belief that god exists will increase until it’s (almost) 1. The odd thing is this: this will happen regardless of the values of the sensitivity and the specificity. Regardless of the “quality” of the evidence in other words.

    On the other hand, if I confront John often enough with the statement “God doesn’t exist, because there’s no evidence that he does”, the same thing will happen. John will become convinced that god doesn’t exist, no matter his original prior or the quality of the evidence.

    Why? It’s a general mathematical property of the formula that repeated application will converge to the value 1.

    Perhaps my back-of-the-enveloppe calculations are wrong, but this seems to suggest that there’s no such thing as a fallacy, a bad argument or, indeed, a good argument. There are only arguments that aren’t repeated often enough.

    This may well be true – it’s certainly a prior conviction hold by many – but you have to wonder what insight Bayesian analysis then gives in the quality of arguments and in fallacies.


  16. History and especially antiquities is a narrative discipline, not a statistical one. The objects that we use to try and piece together “what happened” have to be interpreted and do not provide neutral “facts” that can be plugged into formulae. This is especially true of ancient texts which, if one is familiar with the subject, were not intended to be “factual” in the manner understood by modern historians. Ancient historians’ accounts have a multiplicity of purposes beyond recounting facts and those dimensions of the texts are not demarcated in any way from their factual content, such that one can separate out the facts and plug them into statistical formulas. This is why the best we can do with ancient history, oftentimes, is write a kind of historical fiction, an excellent example of which is Robert Graves’s “I, Claudius,” which gives a fictionalized description of the intimate goings-on in the Imperial court and whose sources include Suetonius, Tacitus, and Plutarch. When you read the actual sources themselves, you will see how little real information they actually provide, and that information is a mess of facts and propaganda. The Bible, of course, is even worse, as a source, and pretty much provides the only information we have on Jesus (the stuff in Josephus being widely believed to be later Christian ad-ons.) The idea that one could use this sort of text to provide a basis for a statistical analysis as to how likely it is that Jesus existed is such utter nonsense that it is hard to believe anyone buys it.

    So, yes, the critique I posted is spot-on, Coel, and once again, we disagree. What a suprise.

    Liked by 3 people

  17. Oops!

    > this will happen regardless of the values of the sensitivity and the specificity. Regardless of the “quality” of the evidence in other words.

    Another back-of-the-enveloppe calculation shows this is not entirely true. There are bounds on sensitivity an specificity, or more correctly on the relation between the two.
    But the quality of the evidence can be pretty bad while still having a convergence to 1.


  18. “Evidence”, in law, history, and much of science, is all about establishing in what “universe” (in the sense given in Logical Treatises) the logos of the debate is going to live.

    Informal Bayesian analysis is used all the way to do so. It is informal, because it depends blatantly upon subjective elements (so does all and any logos).

    It can be fraught: some used it to “prove” the existence of Jesus, or its opposite.

    A good way to understand the root of a flawed reasoning is to understand the logic that exert psychological pressure to produce that lie. There was a need for a Jesus character, so plenty of Jesus characters were produced, by the general logic in attendance.

    What was that logic?

    Jewish faith was Judeo-centric. It had a great strength: an undivided god. Many religions recognized a god of the gods, but having no god but god was simpler, and less subject to contradictions, while being more sympathetic to a state led by just one “Prince” (Princeps).

    A message more oriented towards all people, not just Jews, and normal human ethology, that is, with more love than Rome experienced, fit the species better.

    Hence a full century before the alleged Jesus, there was another, just like him in his philosophical message, but this one gentleman was fully historically documented, in Alexandria.

    The logic wanted a Jesus, so Saint Paul produced it (with several caveats in his writings which basically recognized he made Jesus up, and those caveats were produced by me, long ago, and Carrier, more recently).

    When Laplace furthered “Bayesian” analysis, he was interested by some games of chance.
    When philosophers produce truth, they do not blindly parrot gnu logic. Gnus are herd animals, travelling by the millions. Gnu Christians have stampeded all over civilization for 17 centuries.

    How does new philosophy produce new truth? By pondering why gnus do what they do.

    Why did Saint Paul want Jesus to be? Why was the “Jesus” message welcomed by the empire? Emperors and bishops who governed the empire in 400 CE, had interest to eliminate the logics those questions called for.

    New truth is produced by introducing new facts, which break the universe the old logic rested on.
    The best way to do that, is through a meta-logic making the old logic a special case (as General Relativity did to Classical Gravitation).

    Arguably, Jesus was just the meta-logic towards a more human society, which the Roman Empire was sorely in need of.

    Having a reason for Jesus the myth, makes the historical Jesus less likely.

    What sort of reasoning is this? Having a reason for an hypothesis can make axioms that led to this hypothesis superfluous. This is not properly speaking what came to be called “Bayesian” (a recent term) analysis. But it is related.

    When Laplace presented his book on Celestial Mechanics to Napoleon, the tyrant retorted: ”I do not see God in your book.” Laplace retorted: “I did not need this hypothesis.”


  19. Hi David Ottlinger,

    I found this response from Bart Ehrman pretty damning …

    I don’t think that either Ehrman or Carrier did themselves any favours in that exchange (and see Carrier’s reply before concluding that it is “damning”). They are both far too ready to sneer at each other, and the discussion quickly delved into inconsequential details.

    But, to pick one major and important issue, from that blog post:

    “That Jesus lived recently is … the view of all of the Gospel Sources, V Q, M, L …”

    That is typical Ehrman writing, treating Q, M and L as “gospel sources”. He has also said, referring to Q, M, L etc, that:

    “we have numerous, independent accounts of [Jesus’s] life in the sources lying behind the Gospels … sources that originated in Jesus’s native tongue Aramaic and that can be dated to within just a year or two of his life”.

    But, no he does not “have” those sources, they are entirely conjectural!

    The argument for Q, M and L goes like this: the anonymous authors known as “Mark”, “Matthew”, “Luke” and “John” were writing around AD75 to AD110 or so. But, where did they get their info from? None of them even claim to have spoken to an eye-witness (not even Luke, if you read carefully).

    The parsimonious hypothesis is this: Mark made it up; Matthew took “Mark” and embellished it; “Luke” took both “Mark and “Matthew” and embellished further, then “John” based his account on all three.

    There is no good evidence against that idea. Really, there isn’t! But, to the mainstream it is literally unthinkable, because if “Mark” just made it up then, well, … no, one can’t possibly even think that!

    So, there just *had* to be sources. So we’ll hypothesize that Mark used a source that we’ll call “Q”. But Matthew and Luke can’t have just made stuff up, can they?, so they had to have had independent sources, didn’t they? So we’ll call those M and L.

    And, look, it’d make Jesus much more real if we say that these sources were *independent*, and we’ll say they were in Aramaic, because that makes them more authentic, and we’ll say they dated from “within just a year or two of his life” for the same reason.

    That is the “evidence” for Q, M and L. The alternative and parsimonious explanation is unthinkable, so, they invent three *entirely* *hypothetical* “sources” in order to avoid thinking it!

    But no-one actually “has” these “sources”, they are purely conjectural. All we have are Mark, Matthew, Luke, etc.

    So along comes Ehrman and says:

    “we have numerous, independent accounts of his life in the sources lying behind the Gospels …

    Sounds good, doesn’t it? But, no we do not have those sources! They are merely *conjectured* in order to boost the case that the gospel accounts relate to a real, historical Jesus! They come solely from an a priori rejection of “they made it up”.

    Ehrman is indulging in circular reasoning of breathtaking proportions! He then says that since we “have” those sources, and since they are “independent”, and since they are “in Aramaic” and since they date “to within just a year or two of his life” (none of which is known) then they show that Jesus really did exist!

    And yet Ehrman (an agnostic) is supposed to be one of the more sensible scholars in the field. The Christians are much worse again. There is something deeply rotten about this entire field.

    Hi Aravis,

    Historical *sciences* (cosmology, geology, evolution, etc) are also “narrative” fields. Yes, it is indeed the case that many ancient texts were not intended to be factual. That’s obvious but changes nothing. Alex SL‘s account here is a lot more “spot on” than the piece you quoted.

    Hi Patrick,

    Iterating the formula by repeatedly adding new evidence will indeed eventually (and quite properly) overwhelm the prior. But repeating old evidence is not new evidence.

    Liked by 1 person

  20. @ ej winner: “I prefer to make comments that I have prepared in advance and worked on….”

    Undoubtedly. I prefer to read people who actually treat the discussion of the use of Bayes’ idea in a non-tendentious fashion, not, like some on the thread, trying to shoehorn the humanities and the social sciences into the hard sciences because of a hegemony-complex (what Massimo himself has otherwise aptly identified as scientists’ “monistic” tendencies- ultimately derived from an Enlightenment hangover). Aravis’ link to the Quora answer is perfectly in order. Further, it also mentions a Bayesian analysis that ends up humorously doubting Carrier’s existence:


    @ Alex : “One uses buried artefacts and ancient documents, the other uses buried fossils and DNA sequence data, in both cases to infer if and when and why certain events happened in the unobservable past.”

    Neither fossils nor DNA need to take into account human intentions behind their creation. Historians do, and they must sift the data at each stage, making ultimately subjective calls. Reading an historical document is nothing like analyzing a fossil. One must ask all sorts of questions about how it was created and why, for what purpose. More often than not, that purpose has nothing to do with leaving us a record of that past civilization. None of these sorts of questions need even be broached by the scientist- he just has to figure out how the evidence fits or not with the reconstruction of the natural world that he’s working with. He does not have to ask whether the trilobite was really half propaganda. Further, there is the problem of source texts. For example, one of our key sources for the Germanic people in the Classical period is Tacitus’ Germania. That text we have only because of a single surviving copy (Codex Hersfeldensis) acquired from a monastery near Fulda, Germany and then analyzed by the Renaissance Humanist Piccolomini (later Pope Julius II). Is it (logically) possible that some monk in Medieval Germany was good enough with Latin to completely make it up? Yes. And how does one assess the likelihood of that, multiplied over scores of other documents/artifacts, etc? Carrier’s a joke not only because of his misuse of Bayes, but also because per his technique, why should I believe that Caesar crossed the Rubicon or that Boadicea was a real person? Why shouldn’t I think that Carrier’s mathematization is a stalking horse for something else, since he doesn’t disavow his own doctoral field, much of which could be analyzed the same way?

    As Aravis has made clear, history is the best story we have of the past (historians like Hayden White have known this for decades). Pace von Ranke, it is not a science, nor can it ever be, given what it is that is being studied and the limitations of access to the subject itself.

    Liked by 1 person

  21. All right, let’s try to clear up a few issues.

    1. There is probabilistic reasoning, which can operate intuitively. There are probability theories. One of these theories is the Bayesian. These cannot be collapsed into each other.
    2. Do historians use probalistic reasoning in the constructions of their narrative? Yes, this is obvious. Can they learn something in this process from Probability theory? Probably. Can this be reduced to Bayesian theory? Possibly. But that needs argument.
    3, Are the stories of human history and cosmology both narrative in structure? Yes. Do they share the same narrative structure? Obviously not. Can we adduce the narrative structure from the one to the other? If they do not share the same narrative structure, then the answer is clearly ‘no.’


    I am really tired of your attempts to hijack this webzine to your scientisimist/new atheism agenda. Nothing you have discussed on this thread has had anything to do with the original posted article. You have your own blog, why not post your theories there? You can provide links, Massimo has not restricted that.

    Perhaps the readers here would rather discuss the originally posted article. I think it worse than hubris to assume that you have the solution to all our problems, to the point of carrying the discussion off into your private obsessions. ‘Have you at last no sense of decency?’

    I think you have something to share here, whatever our disagreements. But I have had enough of your arrogant assertions that somehow your point of view will ‘become mainstream’ one day; who cares? If you cannot reply cautiously and charitably to criticism, perhaps you should wrap it up until you can.

    (And shame on those who have played Coel’s game; you don’t respond to Patrice Ayme or Tienzen Gong, because of the difficulty of their language; yet you think you are scoring points with Coel? Think about it.)

    The whole debate over Carrier has been a waste of time, as far as I can tell. And it has nothing to do with the original essay linked here.

    Next question?

    Liked by 1 person

  22. Coel, I’m sorry, but you simply haven’t answered any of my biblical criticism critique of Carrier. I didn’t really expect you to do so, though. And, your answer to David Ottlinger is even worse, showing that you simply don’t have that much of a grasp of this field.

    The “sneering” that you mentioned is otherwise not mine. You, in an earlier post, talked about “apologetics,” for example, as though that were something always engaged in by “the other side.” Well, no, I see enough of it by people not named me, David, etc.

    Also, you apparently opted not to take the challenge I put forth to Alex about showing me that Carrier actually knows Aramaic. (So did Alex, in his last comment.) As far as I can tell, the five languages that he reportedly knows are English, French, German, Latin and (classical) Greek, which means he doesn’t know Hebrew as well as not knowing Aramaic.

    Jarnauga has a good follow-up to the first part of my comment to Alex, speaking of, and also per why we should accept this selectively tendentious application of Bayes’ theorum to Jesus, but not Homer, pre-Socratic philosophers, or, per him, Boadicea.

    To go further, talking about manuscripts, tracing the lineage of a manuscript family is of course far different than tracing a biological lineage. So, Alex, if you choose to be exasperated, sorry, I can’t further help you.

    That said, this whole thread has been useful for showing how Bayesian probability can be misused in the humanities — beyond a scientism-type grab, or “hegemony complex,” as Jarnauga calls it, it’s also that the humanities are narrative, as Aravis notes, and driven by human foibles.

    And, for Aravis’ one brief note on his one comment? He did exist, he was Jewish, and Philo, Jewish Gnostic texts from various locales, the Dead Sea Scrolls, and even Pauline and other writings, etc., indicate that late pre-Rabbinic Judaism was quite diverse and dynamic indeed. Let’s not forget that the Mandeans, a today-tiny world religion, come from the lineage of Judaism via the thread of John the Baptizer, too.


  23. EJWinner wrote:

    The question really has to do with, can Bayesian probability theory be applied to the study of history?


    What we could use here, then, is a stronger, less controversial consideration of that topic (probability and the study of history), which could be very interesting, if well handled.


    I agree with you entirely. And I think that this is precisely the issue that I spoke to, in my last post, which I think was substantive and only incidentally about Carrier. My main point was with respect to ancient history and the nature of the sources that one finds in antiquity. (I would argue, incidentally, that a very similar point can be made with regard to non-ancient history and non-ancient sources.) I gave a sketch of the reasons, which Jarnauga expanded upon: the sources to which we commonly appeal in the practice of history, require interpretation in a way that that nature does not, and this renders them largely unsuitable as a basis upon which to engage in formal probabilistic analysis.

    How would one even get such a probabilistic analysis started, without simply inventing all the relevant values? Suppose, for example, that we want to know whether the drama surrounding Tiberius’ divorce from his wife Vipsania and subsequent marriage to Augustus’ daughter, Julia, really occurred. What are our sources? Well, it turns out there’s exactly one — Suetonius’ “About the Lives of the Caesers.” (Tacitus who writes quite a bit about Tiberius, never mentions it.) The trouble is, there is no reason to think Suetonius’ book is a reliable source. It’s substance is that of a soap opera, Suetonius himself was heavily biased, given what we know of him — and how do we know any of *that*, by the way? — and as a more general matter, we know that the practice of history in antiquity was not as it is today; that historians did not view their job as being to report facts, in as neutral and objective a manner as possible.

    Given this situation, the suggestion that one could somehow calculate the probabilities — mathematically — as to the likelihood that the drama surrounding Tiberius, Vipsania, and Julia actually occurred is worse than a joke. It is silliniess of the highest order. Depressing silliness.


  24. After reviewing material on the subject and re-reading the article, I realize that some of the problems surfacing in the article appear to be endemic to the topic. Hahn and Oaksford are basically proposing that Bayesian theory can help determine why arguments appearing to have similar structures can be more or less convincing given differing content and differing background beliefs.

    The unstated assumption of the article is that all informal argumentation derives from, and returns to, rational thought processes. That’s what triggered my originally noted suspicion that this was another in a long line of efforts to ‘control rhetoric;’ because that assumption happens to be false.

    For instance: “Electrons exist, because we can see 3 cm tracks in a cloud chamber, and 3 cm tracks in cloud chambers are signatures of electrons,” is a proposition derived from a community engaged in constructing and testing theories of explanation; the term ‘electron’ is a conventionally accepted artifice – that is, the community agreed that entities that behave in this way in these tests would be called ‘electrons.’ The proposition not only has a rational structure, but is derived from a rational process of behavior and communication.

    “God exists, because the bible says so, and the bible is the word of God,” on the other hand, is not an argument at all, it is a profession of belief. It only links into argument when directed at an audience that is likely to agree to it; when directed to those unlikely to agree to it, it is really intended to close down the conversation. The belief itself is not rational; while I would not say that one could not reason one’s way to some such belief, the evidence weighs in the balance that it derives from inculcation, emotions, or interpretation of personal experience.

    So the proposition concerning electrons is clearly not begging any questions, while the profession of belief cannot be an inference to the best explanation.

    Note that the effort to interpret the profession of belief as a product of rational thought has caused a confusing comparison between it and the proposition about electrons. While obviously Bayesian theory may help us understand the process of thought and communication that led to the proposition about electrons, it seems less likely to tell us anything interesting about the profession of faith.

    What this example suggests is that discussions of informal fallacies will probably not be normalized through recourse to Bayesian theory, because sentences seeming to share a basic form may be revealed, once properly contextualized, to be hopelessly disjunct. That in turn suggests that the application of any probability theory to informal fallacies must be tested on a case-by-case basis. Which leaves the subject of informal fallacies still plagued with a problem of systematic categorization, as it ever was.


    I apologize for the contentiousness of my last post. But I would have liked greater discussion concerning the problems involved with fallacies in informal arguments. I just reached a point of exasperation; sorry.


  25. In the end: I think the subject in the OP can addressed either via a bayesian or a paraconsistent mechanism.

    Bayesian Logic Programming (BLOG)
    Paraconsistent Logic Programming (P-Datalog)

    (But I don’t know What Would “Jesus” Compute!)


  26. Aravis,

    I agree with you, that re-narration of specific events cannot be crunched down into numbers. As I tried noting in my somewhat poorly worded and over heated post: Of course historians use an intuitive probabilistic reasoning; indeed they have to. What much well-written history amounts to is, ‘given these documents and artifacts what probably happened is this.’ However, that doesn’t yet trigger any formal probabilistic analysis derived from probability theory (although knowing some probability theory may enhances one’s intuition in such reasoning). (However, it should be noted that Bayes is not the only probability theory in town – a point that got lost in the discussion.)

    But another question is whether a statistical probability theory can be applied to any historical study at all, and the answer is yes; but only in addressing sociological questions concerning populations. For instance, if we want to know whether Medieval serfs were relatively happy or unhappy with their social station, probability theory could address statistics derived from what evidence we have from traces of their expressions on the subject.

    So my point is, that I felt the discussion was tumbling down one slippery slope after another, and important distinctions were getting lost: probabilistic reasoning is not probability theory; historical events are not historical trends.

    And of course not all narratives have the same structure or modality, except in the emptiest ‘A through B gets to C’ sense. Evolution is a story; but “Origin of Species” doesn’t look anything like “Decline and Fall of the Roman Empire.”

    I’d like to take the opportunity of this last comment to make an important point about the presumed existence of certain ancient personages of value to people. Christopher Hitchens liked to say that he was as little convinced that Socrates actually existed as he was that Jesus existed, but that the issue was moot. That’s because the Socrates that Hitchens admired (above Jesus) is the Socrates of Plato’s dialogues. That Socrates doesn’t have to have been the ‘real’ Socrates that may (or may not) have actually existed in Athens at the time. I know Buddhists who will say much the same of Siddharta Gotama (and I am one of these). And I’m sure something similar can be professed by adherents to other philosophies, as well as religions, having their origins in the murky past. That’s because such choices are – sometimes – derived from the content of what was said, not out of deference to the person who may (or may not) have said it. So the whole Carrier discussion was traveling down a blind alley to a dead end, from my perspective.

    (However, I did read the linked material, and in the process, I admit, I did learn something more about Bayesian theory and its use – or misuse – in argumentation.)


  27. I think the the author’s reliance on Walton is misplaced. They acknowledge that Copi characterizes the argument from ignorance differently. I think Copi has the better understanding.

    For example let’s say I conclude that there are not 4 horses in my brother’s garage. The reason I conclude that there are not 4 horses in my brother’s garage is because I am standing in his garage, can see the entirety of it, and see no evidence of any horses. This is not arguing from ignorance. I am arguing from the evidence of my senses.

    Now the authors would plug this into Bayes formula and find that I am justified. But I think this is a poor way to understand the fallacy and Copi has the better view. Once we have a situation where we can say if P were true we would have evidence of P, but we don’t, we are no longer arguing from ignorance. According to Walton we are. The problem with Walton’s formulation is, at that point we are no longer appealing to ignorance but to the evidence we have.

    I think the better view is to just agree with Copi. This then properly directs our attention to the reasons for our belief so that they can be analyzed.

    Bayes theorem is great, but once you are analyzing whether the evidence properly yields certain probabilities you are no longer arguing fallaciously. You are then properly discussing the issue.

    Of course Bayes Theorem doesn’t help address where the conflict really lies. The conflict really lies in what you put in Bayes Theorem. Garbage in garbage out.

    BTW: Why do you cut off comments on your blog after a certain time? Philosophical issues are not like current events, and can properly be discussed for centuries not just a few weeks. Please consider us less aggressive bloggers, and leave the comment section open.


  28. Joe, comments have a limited time window because we actually read them all, and we have limited resources, I can’t allow this to be a forever going job. Also, it focuses people.


  29. Folks will be relieved that this is my fifth and last:

    Hi Aravis,

    The trouble is, there is no reason to think Suetonius’ book is a reliable source. It’s substance is that of a soap opera, Suetonius himself was heavily biased, …

    There you are, entirely properly, evaluating the likelihood of reliability, using the English language.

    Using mathematical language instead of English is not in itself a claim to precision and certainty. Mathematical language can just as well codify uncertainty. What Bayesian formalism does do is reveal hidden assumptions and fallacious reasoning, and that is why Carrier uses it.

    There is huge irony in your comments since you should be far more critical of Carrier’s opponents, who are the ones claiming certainty about Jesus.

    We have *zero* accounts by *anyone* who claims to have met a historical Jesus; and, further, we have zero accounts by anyone who claims to have met someone who had met the historical Jesus. We also have no third-party accounts that are anything like contemporaneous, and of course no archeological evidence.

    So, in the manner of your comments about Suetonious, what would a historian say about the reliability of the decades-later, anonymously written accounts full of supernatural events and clear theological agenda, that are the only source for the personage of Jesus?

    The irony is that it is the mythicists who are willing to consider a wider range of motives for the gospels, and it is the “mainstream” who treat them as fact-reporting documentaries and claim certainty over Jesus’s existence.

    Hi ejwinner,

    No problem! If no-one had replied to my early aside about Carrier I wouldn’t have said more about it; but discussing Jesus mythicism doesn’t prevent anyone discussing the OP.

    I largely agree with your latest comment, but note that all that Bayes’s theorem actually does is combine likelihood estimates non-fallaciously. That’s all! I’m puzzled why anyone would object to that.

    Hi jarnauga111,

    Is it (logically) possible that some monk … completely make it up? Yes. And how does one assess the likelihood of that …

    On the evidence, of course, using what you know about human nature and history. (And, yes, often the evidence will be insufficient to reach any conclusion.)

    … why should I believe that Caesar crossed the Rubicon or that Boadicea was a real person?

    You should believe such things to the extent that the evidence supports them! What’s your alternative, that you’re just constructing stories without caring whether they are true?

    People often reply: “Well, if Jesus wasn’t real, maybe other accepted historical people were also not real?”, as though that is some sort of refutation of the idea of judging the historicity of Jesus on the actual evidence.

    Hi Socratic,

    Also, you apparently opted not to take the challenge … showing me that Carrier actually knows Aramaic.

    Why discuss what is a minor side issue (especially since none of the NT was written in Aramaic)? If the case for Jesus’s historical existence is so clear and strong that those doubting it are fairly labelled “denialists”, then why not just present the evidence?

    Liked by 1 person

  30. I do think this part of the article is important:

    “The burden of proof is a concept that has been imported into general argumentation theory from law, where it is used to allocate the duties regarding the provision of evidence (along with the risk of failing to fulfill these duties) between opposing parties. We have argued in detail elsewhere (Hahn & Oaksford, in press) that argumentation theory has somewhat overextended the concept. There is no doubt that the burden of proof is important in any argumentative context involving decision about an action (in law, e.g., to acquit or not to acquit). Here, the need for a decision imposes a threshold on degrees of conviction such that one must decide whether or not one is convinced enough to proceed (see Hahn & Oaksford, in press, for a detailed analysis). Without the need for a decision, however, it is not clear where a threshold on degrees of conviction would come from. Degrees of belief can vary continuously from absolutely not convinced to entirely convinced, consequently some boundary degree of belief that is deemed “convinced enough” must be determined; without such a threshold or boundary that might or might no be met, the idea of the burden of proof is simply undefined.3 Thus the key question is who gets to set this boundary, at what level of conviction, and why. Outside law and other decision-theoretic contexts, the answer to these questions is unclear. In particular, determination of threshold cannot just be up to the whim of the opponent—–that is, a threshold corresponding to whatever it actually takes to convince him or her—–given the emphasis in argumentation on a rational, reasonable critic.”

    People often talk about “the burden of proof” as if we all know what this means. As a lawyer who has to deal with many different burdens of proof I have come to believe that talking about “the burden of proof” in a philosophical sense is hopelessly vague and leads only to poor thinking.

    I would add to the list of problems that, “the burden of proof” in philosophy – unlike law – does not clearly indicated to whom one needs to prove something. Also -unlike the law- it does not spell out the consequences of failure to prove it to this person or group? Are we to say unless we can prove something to some person we are unreasonable to continue to believe it? What if the person is stubborn or biased? Which brings up the next problem. This notion of “the burden of proof” also fails to recognize the subjective aspect of proving something and how much weight evidence is to be given.

    I did a blog post on this topic here:

    I intend to do another more in depth blog on the topic shortly. But I am interested in any thoughts on this part of the paper.


  31. ejwinner: “I am really tired of your attempts to hijack this webzine to your scientisimist/new atheism agenda.

    No one can truly hijack a discussion. Coel is just very skillful on playing the game: making himself as a punching bag (by making some easily refutable points) and thus drawing a pack of hunger wolves with the following mentality and strategy.

    S1, I disagree with you by quoting these great authorities, that is, I am well-read.

    S2, I even disagree with those great authorities in some points, that is, I know more than them.

    Now, everyone can get the CONCLUSION from the two points above.

    The OP has only two points.
    P1, the issues of logic fallacies.

    P2, applying Bayes theorem to analyze those logic fallacies.

    When an issue is 100% certain, there is no need for using Bayes theorem (BT). When a logic fallacy needs the help of BT, it is not a genuine LOGIC anymore. For history, it arises by having some narratives (ancient or current). If there is no conflict among narratives, there is no issue. If there are conflict stories, the WEIGHT of each story can be analyzed, and BT might play a minimal role at this point. These reasons are why I did not get into this BT discussion.

    I have said that logic fallacy is totally benign, can hurt no one. Yet, there are some fallacies are deadly to the advancement of humanity.

    I have asked readers (including …) many times to discuss the true issues discussed at (https://scientiasalon.wordpress.com/2014/06/05/the-multiverse-as-a-scientific-concept-part-ii/comment-page-1/#comment-3158 ), as they can be understood by the every 8th grader. The results are:

    R1, dishonesty: no physicist will talk about it.

    R2, cowardice: no one about 8th grader will talk about it.

    Yet, Nature sits there silently, blocking all detour attempts, that is, no one can go over, go under or go around it. See (http://prebabel.blogspot.com/2013/10/silent-truth-blocking-all-detour.html ).

    The two fallacies above can never be corrected with REASON but will definitely be rid of by Nature’s silence, sooner or later.

    I have nothing against the Bayes theorem. But, when one is begging the help from BT, he is definitely ignorant about the issue. I have used GR (general relativity) as an example.

    GR plays zero role in the construction of quark/lepton while the GR gravity is all about the space-time sheet. I have showed,

    String 1 = (V, A, A 1) = {1st, red, 2/3 e, ½ ħ} = red up quark.

    String 2 = (-A, V, V 1) = {1st, red, -1/3 e, ½ ħ} = red down quark.

    These are just LANGUAGE which describe quarks. Yet, these are also the {mouse/bed sheet} description of quarks. The (V, V, V) is the space-time sheet while the ‘A’ is the mouse. Yes, this quark-language is all about the mouse (mice) under a space-time sheet. But, GR failed to do this. No, we need no BT to reach this conclusion.


  32. My last comment too.

    Is it possible that we are conflating absence of evidence with evidence of absence only to arrive at an argument of incredulity?

    Coel, when you write, “all that Bayes’s theorem actually does is combine likelihood estimates non-fallaciously. That’s all! I’m puzzled why anyone would object to that,” I think your being puzzled is something of a puzzle to those who honestly question the efficacy of the OP’s argument of a Bayesian approach to analyzing informal logical fallacies, along with what might be perceived by some as disingenuous opportunism on your part to introduce a pet polemic of yours into the thread.

    While it may be true, as you say, that “If no-one had replied to [your] early aside about Carrier [you] wouldn’t have said more about it,” it may be just as true that you could have taken this action yourself by openly acknowledging it as the distraction it turned out to be much earlier in the thread.

    At any rate, I think we can all look forward to discussing this topic further, as SciSal indicates above, when the joint paper by Massimo, Maarten, and Fabio comes out.

    Liked by 1 person

  33. Herewith a few short comments that can safely be made.

    1. The different perspectives from different participants, while rather fixed and repeated in different contexts, are very interesting in and of themselves. I learn a lot from different perspectives, and it is made easier and clearer through repetition.

    2. The chances that Coel and Aravis are going to become of one mind are nil, by any analysis, Bayesian or otherwise. I have informally noticed changes in emphasis from both over time; they probably would not be here, unless they thought they could learn something or, at least, sharpen their perspectives. Their contributions are excellent, as are many others’.

    3.. Changing a frame of mind is not that easy. It would require a change in physiology and possibly a change in neuro-anatomy. Experiences while growing up actually change the gross and microscopic physical structure of the brain. These processes continue into adulthood but plasticity of the brain decreases with age. (One way to stay young is to keep challenging the brain?)

    4. SciSal has positioned itself at the interface of philosophy and science and tensions are to be expected. Philosophers might be more personally vulnerable in this situation because they are personally invested. Their work reflects who they are as individuals. Scientists are usually not that close to their subject, in fact, a dispassionate approach is the professional standard.

    5. Have no fear of political hegemony. As a scientist I would say that such a goal would be unscientific. Those scientismists that have attacked philosophy were expressing their personal (misguided) opinion. Are there philosophers that attack science?


  34. Robin
    First I would point out that “proofs” are much more subjective than we often acknowledge. As you say an argument can be perfectly sound (in a technical sense) but still not be a “proof” to someone. That is because even though the premises are in fact true, the person you are discussing the issue with might not *believe* the premises are true. They may be true but the fact that the person you are discussing a topic with doesn’t recognize their truth might mean your perfectly sound argument falls on deaf ears. And yes even reasonable people don’t always recognize the truth. No one is perfect.

    You ask: why number sentences and then put a therefore at the end. The answer is because that helps us identify where the disagreement lies. Usually people who are decent at philosophy will not present arguments that are invalid. (and here I mean in the technical sense – if the premises are true the conclusion must also be true if the argument is valid)

    So if we take the quote from Sam Harris and number the premises this will help us know if we disagree with them and to some extent determine if the argument is in fact valid.

    P1: Morality and values depend on the existence of conscious minds—and specifically on the fact that such minds can experience various forms of well-being and suffering in this universe.
    P2: Conscious minds and their states are natural phenomena, fully constrained by the laws of the universe (whatever these turn out to be in the end).
    C: Therefore, questions of morality and values must have right and wrong answers that fall within the purview of science (in principle, if not in practice).

    Now unfortunately what Sam Harris wrote is not valid. That is the conclusion does not logically follow from the premises. So this argument is a non-sequitur and is not logical. Can it be corrected? Possibly.

    It seems that at least one problem with this argument is that he makes a hidden assumption.

    He assumes this:
    A1) if morality depends on something that is a natural phenomenon (consciousness) that is constrained by the laws of the universe, then all questions of morality must also be constrained by the laws of the universe.

    Well there are few more assumptions as he is a bit fast and loose with his wording. If he was a philosopher proficient at critical reasoning he would tighten this up. And if he did we would see the hidden assumptions even more clearly.

    The hidden assumption seems to be false. We can use an analogy to spell the problem with this assumption. My life depends on the continued beating of my heart. My heart continues to beat entirely due to the natural laws that involve my heart’s muscle tissue continuing to work. Yet all the questions about my life are not also contained in the laws that involve my hearts muscle tissue continuing to work.

    By numbering each premise and putting a therefore at the end for your conclusion you make what premises you claim clear. But you also help people identify if you are actually being logical. Here Sam Harris is being illogical. Even if we agree for the sake of argument that his two premises are true, his conclusion still does not follow from those premises.


  35. Coel, The issue of Carrier’s knowledge of Aramaic, not whether any of the NT was written in Aramaic, was the issue at hand. I would call this a “good”use of the Overton Window, and again, not surprising.

    The reality, as properly framed, is that Aramaic certainly lies behind the background of the NT, including those Aramaic-language Targums that Carrier oh-so-willfully misinterprets.

    Then, there’s this:

    The irony is that it is the mythicists who are willing to consider a wider range of motives for the gospels, and it is the “mainstream” who treat them as fact-reporting documentaries and claim certainty over Jesus’s existence.

    Otherwise, modern critical scholars certainly look at a quite wide range of reasons for authorship of various books. The idea that they’re narrow-minded is yet another straw man.

    Specifically, it’s a Gnu Atheist straw man. It first assumes that every non-denialist Bible scholar is a Christian, which isn’t true, and it assumes that every Christian is a fundamentalist. Your idea of “mainstream” might be accepted at some fundamentalist bible college, but certainly nowhere else.

    I do thank you, though, for sharing the full degree of your delusion about what you think biblical scholarship is versus what it actually is.

    There’s nothing else to say, other than, should this topic come up again elsewhere, I’ll start from this point, in remembering your seeming delusions about actual scholarship.

    I ask a rhetorical question: Do you actually believe this, or is this just some sort of literary trope?

    Otherwise, along with Thomas, I too await Massimo’s piece.


%d bloggers like this: