The Bayes’s Theorem approach to determining probabilities divides the evidence into categories: background evidence, which figures into the “prior probability,” and the evidence being evaluated. Mathematically, there is no difference in the final outcome that results from making a different decision about which category to place a particular piece of evidence. Given this, Carrier determines his “prior probability” by determining the best “reference class” to place Jesus into and determining the percentage of known figures in that reference class in each of the two main categories: historical and mythical. Carrier then places all other evidence into the second category. So, a natural place to start our discussion is with this reference class that Carrier applies to Jesus, which is covered in Chapter 6 of his book.
Carrier’s opening example in this chapter illustrates the concept of this reference class quite strikingly, so I am going to quote it directly (p. 235):
In 1945 Betty Crocker was rated in a national survey as the second most admired woman in America, and to this day a street is named after her in Golden Valley, Minnesota, where she still lives. Her father was William Crocker, a successful corporate executive in the food industry, and she started her career answering letters on cooking questions for her father’s company, then acquired her own national radio show where she delivered cooking advice for twenty-four years. Later she had her own television show, while making appearances on other TV shows and in TV commercials to promote her products. I’ve seen actual video tapes of her cooking and speaking, and her picture still adorns various General Mills baking products. She has also published several cookbooks, and now has her own website. All that is 100 percent true. And yet she doesn’t exist. She was never born, never lived, never spoke, never appeared on TV, and never wrote a word. Others simply wrote or appeared in her name. Welcome to the world of the mythical corporate mascot.
Corporate mascots like Tony the Tiger are clearly invented; however there are plenty of others where it’s not initially obvious whether they were originally fictional, like Betty Crocker, or based on a real person, like Colonel Sanders. Faced with a new, unfamiliar corporate mascot, one way to start the evaluation of whether the mascot is based on a real person or is completely fictional, before looking at any of the direct evidence related to this particular mascot, would be to see what fraction of known mascots are based on a real person. If three-quarters of all human corporate mascots are purely fictional and one-quarter are based on a real person, then that would give us a prior probability of 75% that a new human corporate mascot was fictional, and a prior probability of 25% that it was based on a real person. Any specific data about the new mascot would then figure into the full Bayesian probability calculation, and could shift the final probabilities either way from this starting point.
In trying to apply this approach to Jesus Christ, we must therefore decide what the appropriate reference class to consider is. Carrier argues that the best reference class is what he calls the Rank-Raglan hero-type, after the two scholars who discovered and described it, Otto Rank and Lord Raglan. Rank and Raglan identified twenty-two features that were distinctive of this particular type of hero (pp. 229-230, discussed as element 48 in the background context):
- The hero’s mother is a virgin.
- His father is a king or the heir of a king.
- The circumstances of his conception are unusual.
- He is reputed to be the son of a god.
- An attempt is made to kill him when he is a baby.
- To escape which he is spirited away from those trying to kill him.
- He is reared in a foreign country by one or more foster parents.
- We are told nothing of his childhood.
- On reaching manhood he returns to his future kingdom.
- He is crowned, hailed or becomes king.
- He reigns uneventfully (i.e., without wars or national catastrophes).
- He prescribes laws.
- He then loses favor with the gods or his subjects.
- He is driven from the throne or city.
- He meets with a mysterious death.
- He dies atop a hill or high place.
- His children, if any, do not succeed him.
- His body turns up missing.
- Yet he still has one or more holy sepulchers (in fact or fiction).
- Before taking a throne or a wife, he battles and defeats a great adversary (such as a king, giant, dragon, or wild beast).
- His parents are related to each other.
- He marries a queen or princess related to his predecessor.
There are at least fifteen known real or mythic people, including Jesus, that clearly meet more than half of these criteria (i.e., twelve or more of the elements). These fifteen are (in order of descending number of criteria met, that number in parentheses):
- Oedipus (21)
- Moses (20)
- Jesus (20)
- Theseus (19)
- Dionysus (19)
- Romulus (18)
- Perseus (17)
- Hercules (17)
- Zeus (15)
- Belleraphon (14)
- Jason (14)
- Osiris (14)
- Pelops (13)
- Asclepius (12)
- Joseph [i.e., the son of Jacob] (12)
With the exceptions of Moses, Jesus, and Joseph, all of these characters have been recognized as mythical over the last century of scholarship. Tracing their origins leads in all twelve of these cases to either (a) deities that were translated into men and placed into history (e.g., Hercules, Dionysus, and Osiris) or (b) heroic characters in absurdist supernatural dramas that have no plausible claim to historicity (e.g., Jason, Perseus, and Bellerophon). Jesus is the case we are exploring. Moses and Joseph similarly are generally considered to be mythical by nearly all secular experts in biblical antiquities and a large fraction of religious experts (Jewish and Christian).
However, in order to argue a fortiori, we should grant historicity the greatest possible benefit-of-the-doubt in interpreting the evidence. Therefore we’ll go against the scholarly consensus and the evidence to grant that Moses and Joseph count as historical people. Further, to avoid the argument of special pleading, let’s further grant against the evidence that at least two others on the list are historical. This means that within the category of Rank-Raglan heroes, four out of fourteen (excluding Jesus, the case in question), or 31.25%, are historical. Carrier then bumps this up even further to an even 1 in 3, or 33%. That, then, will be our upper bound on the prior probability that Jesus Christ was historical.
Carrier then calculates the prior probability by several alternate means, and in each case arrives at the same generous prior probability of 33% historical. Some of these alternate means are:
- Historical people with a convenient name: “Jesus is an English derivation from the Greek spelling of the Hebrew name Joshua (Yeshua), which means ‘Yahweh saves.’ Christ is from the Greek christos, meaning ‘anointed,’ which in Hebrew is māšǐaḥ, ‘messiah.’” (p. 240)
- Chance of a historical Jesus being remade into a Rank-Raglan hero.
I won’t go through the calculations and reasoning here, for, like Carrier, I find the more straightforward Rank-Raglan classification to be the most compelling.
It is important to keep in mind that this 33% prior probability is not the probability that Jesus was historical. It is the starting probability if we only consider the class that the claim exists within, before looking at any other evidence.
Suppose, contrary to fact, that Caesar Augustus had a Rank-Raglan score of 20. Because we also have a ridiculous amount of evidence supporting his existence, each item of which is highly improbable unless Caesar Augustus really existed, even starting with a prior probability of 33% based on Rank-Raglan categorization would be completely overcome by this other evidence. This also can be shown quantitatively from Bayes’s Theorem, though I will spare you the mathematics here.
But if that is the a fortiori prior probability, what is the more reasonable prior probability where we don’t make make as generous assumptions in favor of historicity? Well, we take away the four assumed historical figures in the list of Rank-Raglan heroes, and assume that Jesus is the only truly historical example. Doing this gives us a 6.25% chance of historicity for the prior probability.
So what we are left with is, simply based on Rank-Raglan classification, and looking at no further evidence yet, the probability of Jesus having been a historical figure is somewhere between 6.25% and 33%. The rest of these essays will examine how Carrier reasons the rest of this evidence affects these two limits.