39 of 48 people found the following review helpful
A Long-Overdue Contribution to the Historical Jesus Studies Field,
Amazon Verified Purchase(What's this?)
This review is from: Proving History: Bayes's Theorem and the Quest for the Historical Jesus (Hardcover)When I first became interested in studying the historical Jesus for myself, I decided to get a quick overview of modern scholarship by reading The Historical Jesus: Five Views. What I quickly came to realize after reading only a few chapters is that the study of the historical Jesus is a mess. Not only is there almost no consensus on who the historical Jesus was but there seemed to be very little hope of resolving these differences in any sort of objective or scientific manner. While present-day Jesus scholars have attempted to develop criteria to evaluate what bits of data actually go back to a historical Jesus, as Carrier notes, "the concept of Jesus we're supposed to believe existed is actually getting more confused the more people study it (p.12)." Enter Richard Carrier's book.
Carrier does not set forth a view of the historical Jesus in this volume. Rather, his goal is to "present a new method that solves the problem... so progress can finally be made in the field of Jesus studies (p.15)". His new method is Bayes's Theorem (BT). One need not be an expert in mathematics or even statistics to follow along in the book: a basic understanding of multiplication, division, and fractions will suffice. However, even if your eyes tend to glaze over once Carrier begins to plug in some numbers in the formula, he still adequately conveys conceptually the arguments he is defending.
The main arguments that an amateur reader like myself can take away from Carrier's work are the following:
1) Contrary to what some (most recently, Bart Ehrman) say, history IS a science. "The fact that historical theories rest on far weaker evidence relative to scientific theories, and as a result achieve a far lower degree of certainty, is a difference only in degree, not in kind (p 48)." Thus, when evaluating historical claims and evaluating when the evidence should cause us to believe the claim, history, like science, is Bayesian.
2) Precision is not necessary to apply Bayes's Theorem. "Rules of thumb" will work just fine and be accurate enough for all historical inquiry.
3) Rather than increasing the amount of disagreement among historians due to quibbling over probabilities, BT will actually expose historians' biases and force them to argue for their premises.
4) All current historical methods in Jesus studies (arguments from evidence, arguments to the best explanation, etc..) reduce to Bayes's Theorem. In other words, whether or not something really is the "best explanation" can only be determined by running the probabilities through BT.
5) Current historical Jesus criteria have failed to solve what Carrier calls the "Threshold Problem". In other words, do any of the historical Jesus criteria (dissimilarity, embarrassment, multiple attestation, etc...) in and of themselves tell the historical whether the claim is to be believed? Only by applying BT, argues Carrier.
I highly recommend this book. You need not agree everything Richard Carrier has ever written to recognize that this work is a great contribution to the field of Jesus studies. If it does nothing more than force current scholars in the field on all sides of the debate to abandon the unwarranted certitude many employ to their conclusions and put all arguments through the same, objective test then this book will have served its purpose.
Sort: Oldest first | Newest first
Showing 1-1 of 1 posts in this discussion
Initial post: Sep 2, 2013 10:04:08 PM PDT
Last edited by the author on Sep 16, 2013 11:58:03 PM PDT
F. Ramos says:
Hey C. Murphy,
Pretty good review of the book's main points. Though you mention you are an amateur reader, you did a pretty good summary. Good job.
Well I hear you on the Jesus studies. Historians have been going nuts on Jesus to the point that he is everything to everyone. Carrier's suggestion of using Bayes theorem would simply make things more "transparent", but the mess that you see in Jesus studies would simply transfer over to a numerical mess by doing so. Bayes theorem is pretty old and has more efficient utility in the natural sciences since natural phenomenon has more ontological consistency than does social phenomenon. It really a simple concept and the math is very easy to work with. I think Carrier does a good job in explaining it overall.
But the problems I see in using it in history are numerous. In history there are way too many variables to consider that are not numerical or quantifiable and are very much subjective/arbitrary to put into a numerical value for all the 4 probability variables involved in the theorem, as he has it in the book (I simplified it with less notation) : P = [h X e] / [(h X e) + (~h X ~e)] with all values between 0 = 0% probability and 1 = 100% probability. The "canon of probabilities" numerical values he mentions for using in the theorem may be useful as a start, though, since he has different levels of certainty between 0 and 1 . Though he provides many hypothetical examples through the book, he applies Bayes theorem only on one historical case study - the eclipse example from 30 AD. Here one can see the arbitrariness of assigning values in his calculations. The background is 'everything we know about science, Christianity, human nature, and 1st century Palestine'. He assigns .01 for h (the eclipse did happen - extremely improbable) and then he assigns .01 for e (evidences or documents making the claim of an eclipse - extremely improbable) and then he assigns .99 for ~h (eclipse story is made up - extremely probable) finally he assigns 1 for ~e (evidences for eclipse will be lacking - virtually certain). His result is [.01 X .01] / [(.01 X .01) + (.99 X 1)] = .0001 = .01% - that the probability the eclipse happened is pretty much impossible.
He assigns very low values for h and e and conversely high values for ~h and ~e pretty much based on his assumptions in his "background" which of course is biased by his naturalistic/atheistic worldview, which he mentions in the beginning of the book. Why did he choose these values as opposed to any others, he does not really explain. Overall, miracle claims would all be of low probability in the first place under his background assumptions. I am sure others, of different worldviews, who asses the same event with a different background data would put in other values. The theorem won't really help in standardizing results since the opinions of scholars from one extreme to another would fluctuate and also when more evidence or discoveries are made, they would also change the variables to unpredictable values as some scholars will downplay (reduce their probability values by quite a bit arbitrarily) and others will exalt (increase their probability values by quite a bit arbitrarily) to make the probabilities look more friendly to their position or argument. Obviously probabilities from Bayes theorem are affected by the assumptions you make. For instance if you believe the Bible is reliable then you will likely choose values that generate higher probabilities of belief (plus the evidence gets greater as more sources are taken into account), while someone who thinks the whole Bible is corrupt, will choose values that generate lower probabilities of belief (most of the evidence available will be dismissed for some reason). This is a given. I think William Lane Craig used a probability calculation from another source on a debate against Bart Ehrman once to show that miracles did occur (resurrection of Jesus)!
In the end, the resulting probabilities ALWAYS reflect a historian's personal subjective certainty, NEVER the certainty of the historical matters themselves nor the solutions proposed to historical problems. Since subjectivity plays a major role in historical interpretation (i.e. UFOs are in AREA 51 or military research lab; Holocaust happened or Holocaust did not occur; etc), it will make many results from Bayes' objective theorem, pretty much subjective. Your output in any equation is as good as your input.
Throughout the book Carrier does use many unexplained values in Bayes theorem. It would have been nice if he had systematically justified many or all of them in detail to see his assumptions for choosing one value over any other. For all practical purposes, many of his values seem "ok", but they also look like "guesstimates" to me.
What would have been good is if he had provided multiple case studies from multiple cultures (more than just Jesus). Most of his examples are hypothetical problems to illustrate the usefulness of the theorem, but the theorem is not applied to multiple actual historical events in detail to see if it really produces viable and useful results which can be confirmed by independent evidence. It lacks tested results and confirmation from actual historical inquiries.
Historical data on any one particular question only marks one sample too. In order to be able to use arguments of probability more effectively, especially Bayes Theorem, one MUST have an array of multiple samples (including control samples) that are functionally similar or have a key common characteristic(s) to do a comparative analysis. In order to do this there has to be objective and clear uniformity of characteristics (culture, events, beliefs, etc) among the samples to see if Bayes theorem does produce consistent and reliable information on historical inquires. In order to show that Jesus did not exist or exist on needs to performs a comparative study assessing the sources, manuscript traditions of those sources, and historical records of other historical figures (i.e. Socrates, Pythagoras, Hippocrates, the original Buddha (Siddhartha Gautama), Zoroaster, Confucius, Homer, Epicurus, Democritus, Leucippus, Spartacus, Cicero, Archimedes, Euclid, Zeno, Aesop, Muhammad, and other historical figures) to calibrate any conclusions on Jesus. It makes no sense to only look at one sample - Jesus - and leave others untouched by critical eyes. Carrier and other mythicists would need to evaluate if Jesus' historical record is a special case or the norm in the ancient world.
Historical research has way too many variables and are very much arbitrary in a major sense since no historical arguments speak for themselves and the background data ALWAYS changes as more historical information from new discoveries (other manuscripts, other testimonies, etc.) and artifacts emerge. Even in archaeology where actual direct empirical evidence is perpetually used, Bayes Theorem would not work or provide any reliable insights (except maybe relevance) since the background data is ALWAYS incomplete and not crystal clear, as it is in historical studies too. Discoveries sometimes affect historical interpretation and unexpected findings often change our images greatly. Since this is the case, any numerical result that comes from using Bayes Theorem in non-precise fields like history is always suspect and unverifiable if it is indeed correct or not. How the heck would we know that a probability value is indeed the end of the matter, let alone able to help settle any matter in any significant way? Using Bayes theorem may only help make historians assessments more transparent, but then again historians already are very transparent by their own conclusions either way. Its not hard to figure out their assumptions and views.
The historical record is choppy, fragmentary and incredibly incomplete. It is also not systematic or linear or congruent either. It would be nice if it was, but that is the empirical reality. People in past wrote how they wanted and what they wanted. They sure were not writing for us in the future in our historical structures and our historical criteria nor our expectations in mind. So this, on top of all compounding of biases an selective scholarship that historians generally engage in selectively citing each other makes using Bayes theorem not really promising to somehow save historical studies in general.
Historians tend to make all sorts of fallacies when assessing history too (Historians' Fallacies : Toward a Logic of Historical Thought) which render more subjectivity to these matters on top of the other variables involved. Carrier relies heavily on some modern historians' reified and imagined reconstructions of Jesus (which are all arbitrary and clear reifications of Jesus) and not the direct evidence from the available sources of the historical record to support his views. Carrier's comments on Ch.2 and Ch. 5 of the book show the fine levels of subjectivity that modern historians work in and he also claims that his axioms are philosophically dependent (this is a major problem in historical studies since each individual historian has diverse assumptions and justifications for those that are incongruent with other historians assumptions and their justifications). The Ch.2 highlights the idealistic mentality historians should have and the axioms are nice principles, but usually these do not work that way in practice because of the limits of historical evidence for different cultures, people, places, and events in the past require lots of assuming and filling in gaps that are not explicitly in the historical record.
The ideal principles he mentions are not how most historians including Carrier work through, nor do they follow these by the letter because the historical record limits their usage and exceptions are common. One of the major problems is that modern historians, including Carrier, have and do compound biases and selective sources which they will use or ignore or dismiss on any historical question. No evidence speaks for itself, so interpretation, inferences, and speculation of things they do not have direct evidence for, end up being a massive part of historical reconstruction - inevitably.
‹ Previous 1 Next ›