The seventh criterion

(Book: On Guard, by William Lane Craig. Chapter 8: “Who Was Jesus?”)

William Lane Craig openly admits that he believes in assuming the New Testament is right until proven wrong. For the sake of appearances, though, he proposes a kind of academic neutrality that is at least nominally open to the possibility that individual events recounted in the NT might be true or false, regardless of the reliability of other reported events. Today we’re going to look at the six criteria that he uses to judge the authenticity of individual gospel events. We’ll also look at the seventh criterion, which he seems to avoid mentioning.

I should point out that these six criteria are not Craig’s own invention, but rather are intended to represent the kind of criteria historical scholars typically use to evaluate any historical claim. Craig may have adapted them somewhat to his own purposes, but in general most of them are not too bad. Let’s take Craig’s criteria in the order he presents them.

1. Historical fit: The incident fits in with known historical facts of the time and place.

For example, as Richard Carrier recently pointed out, Matthew and Luke each tie Jesus’ birth to a different date: Matthew has 4BC, prior to the death of Herod the Great, and Luke has 6AD, the year of the first Roman census. Christian apologists ignore the obvious explanation—that Mary was in labor for nine years—and instead try to find some way to make Herod co-governor with Quirinius in 4BC or some such. As Carrier points out in extensive detail, such explanations are not a historical fit, for a number of reasons. The Gospels contradict each other, so just deal.

2. Independent, early sources: The incident is related in multiple sources, which are near to the time when the incident is said to have occurred and which don’t rely on each other or on a common source.

This would be a good criterion, if Craig applied it correctly. Unfortunately, he considers the New Testament authors to be multiple independent sources, despite the fact that they are all passing on the common tradition as taught in the church. Just five pages earlier, he was arguing that the Jewish culture had perfected the art of passing on oral tradition intact, which means that the NT authors were not independent (at least after the initial formative period when the stories stabilized in more or less their present form). An independent source would have to be some non-Christian record, and it would need to say more than just “This is what the Christians tell us they believe.” Unfortunately for Craig, almost none of that has anything at all to do with the historicity of the New Testament.

3. Embarrassment: The incident is awkward or counterproductive for the early Christians.

This is another good criterion, and one that most Christian apologists get wrong. If Christians are going to preach a Messiah who is prophesied to be “the stone that the builders rejected,” then it’s not an embarrassment for Christians to tell stories about the “builders” (i.e. the scribes and chief priests) rejecting him. A genuine embarrassment would have to be something that contradicts the story, not something that reinforces it. For example, if the Christian God is supposed to be the judge of the living and the dead, and Jesus turns around and denies that God is the God of the dead, then that undercuts the gospel and makes salvation and eternal judgment a lie. As such, reports of Jesus saying this are more likely to be true than if he were reported as saying and doing things that only reinforce the gospel.

4. Dissimilarity: The incident is unlike earlier Jewish ideas and/or unlike later Christian ideas.

This one’s pretty iffy. Zola Levitt once remarked that Jews have a saying: “two Jews, three opinions.” Even among rabbis, there’s a range of opinion and a number of disagreements. Once you add in the gap between rabbinical beliefs on the one hand, and the mish-mosh of superstitions and heresies that lay believers absorb on the other, it’s pretty hard to objectively determine that certain disagreements are “dissimilar” whereas other disagreements are not. At best, dissimilarity shows only that the person relating the story had a bias in favor of the novel idea. Then again, there is some merit in observing cases where a recorded teaching contradicts subsequent Christian dogma. To the extent that a Biblical record disproves accepted Christian teachings, it’s less likely to be the product of mere orthodox propaganda. Somehow I don’t think Craig is going to present too much of that evidence, though.

5. Semitisms: Traces of Hebrew or Aramaic language (spoken by Jesus’ countrymen) appear in the story.

Linguistic evidence is another good criterion, and it’s one that highlights a singularly remarkable aspect of the Gospel. None of the early documents about Jesus’ life are in his native language. Not one. No manuscripts, and no mention of any early believer so much as hearing about the possible existence of any such document. [Note: no, I stand corrected, as Jayman reminds me in the comments. Eusebius claimed that Papias claimed to have heard of a Hebrew gospel of Matthew (though one might suspect that either Eusebius or Papias was mistaken, given the balance of the evidence).] Imagine if you were researching Mohammed, and discovered that prior to the late 900’s the only records of him—and the Qur’an itself—were all in French. Might raise some eyebrows, eh? I’m still skeptical of the conclusion that Jesus never existed, but if I were to be convinced, I’d find this bit of evidence particularly compelling.

6. Coherence: The incident fits in with facts already established about Jesus.

This one’s the ringer. Call it confirmation bias instead of coherence if you want a more accurate label. Remember, Craig has already admitted that he believes in assuming the NT is true until proven false. Criterion 6 takes standard dogma about Jesus, elevates it to the status of “established fact,” and then uses the stories as evidence confirming the authenticity of each other. Pretty bogus. Internal consistency is necessary in order for the accounts to be true, but it’s not sufficient. This one only works as a negative evidence, using contradictions to prove that at least one of the reports must be false.

So what about the seventh criterion, the one Craig avoids even mentioning? Let me add it to his list.

7. Real-world consistency: The incident is consistent with what we can actually observe in the real world.

Notice the difference between this and #6. Craig only uses the “coherence” criterion to measure the consistency of the stories with one another. That in itself is not a bad criterion, except that Craig is limiting the scope too much. For an alleged historical event to be acceptable as genuinely authentic, it needs to be consistent not only with the facts that Christians accept as “established,” but with the rest of real-world facts as well. The reason we don’t accept Achilles’ invulnerability is because in the real world, there is no river you can dip your baby into that will turn him invulnerable. The same goes for the Gospel as well.

This criterion is conspicuous by its absence, because Craig wants us to assume that miracles are possible by default. He does not want us evaluating miracle stories in the light of what we actually find in the real world. He wants us, in other words, to “keep an open mind”—where “open-minded” means neither more nor less than “gullible.” He wants us to believe what men say, just because they say it, regardless of the internal and external inconsistencies in what they tell us. He has spent this much of the chapter, including his five arguments for assuming that the entire New Testament is true, working towards that end.

What’s remarkable about these criteria is the way they so frequently fail to achieve his goal when applied fairly and objectively to the evidence. Craig therefore follows this discussion with a certain amount of pre-emptive damage control.

Notice a couple of things about these “criteria.” First, they’re all positive signs of historical credibility. Therefore, they can only be used to establish the historicity of some incident, not to deny it. If a story is not embarrassing or dissimiar or found in independent early sources, that obviously doesn’t mean that the incident isn’t historical.

Slick, eh? Craig is lying to us, but he makes it sound as if he’s telling the truth. Not all of the criteria can be used to deny the historical credibility of a report, but some can. For example, if the reported incident is not a historical fit—like the idea of Quirinius being governor in 4BC—then you can indeed use this evidence to disqualify the report as authentic history. And Criterion 6 can only be safely applied as a negative check. Craig, however, gives examples from the three criteria that do fit the conclusion he wants us to draw, and then applies it to the six criteria as a whole. Pretty sneaky. (And why on earth does he have scare quotes around “criteria”?)

The only way you could justifiably use the criteria to deny historical credibility would be by presupposing that the gospels are unreliable until they are proven to be reliable. We’re right back to the burden of proof issue again! If we adopt a position of neutrality in approaching the gospels, then the failure to prove an incident is historical just leaves you in a position of neutrality. You just don’t know whether it’s historical or not.

Agnosticism is the last refuge of the believer, and here we have Craig setting up a defensive position he can fall back to if the contest doesn’t go his way (e.g. if someone happens to notice the missing seventh criterion, and apply it). “We can never know!” is the rallying cry of the faithful, and if you say, “Yes, we can, and here’s how,”—well, I’m sure you all know how that turns out.

Second, the criteria don’t presuppose the general reliability of the gospels. The criteria apply to specific incidents, not to a whole book. So they can be used to detect historical nuggets of information in any source, even the apocryphal gospels or the Qur’an. That means that in order to defend the historical credibility of some event in the life of Jesus, say, His burial, you don’t need to defend the historical credibility of other events like His birth in Bethlehem, His feeding the five thousand, His triumphal entry into Jerusalem, and so on.

Aka the divide-and-conquer approach. If you have a good argument against X, the believer will say that it does not apply to Y and Z, and therefore Y and Z could still be true, and therefore X is probably true as well, because if Y and Z are true, then the gospel writer is reliable and should be taken as truthful by default. Remember, this was one of Craig’s arguments for why we should assume that the NT is always true: because the NT authors are reliable witnesses. The six criteria are just there so Craig can plausibly deny and/or isolate the historical problems with the NT, and hopefully score a debate point or two on allegedly historical grounds.

The rest of the chapter is going to go downhill from here. For all the time he spent setting up these “historical” criteria, he’s not really going to look at the New Testament from a neutral, scholarly stance. Instead, he’s going to set out to prove that Jesus claimed to be God Incarnate, with C. S. Lewis in the background warming up the old “Liar, Lord, or Lunatic” schtick. Serious historical issues aren’t even going to come up.

On the other hand, what he does say should give us the chance to have a little fun, so stay tuned.