[This is an older essay, and I plan to update some of the content in the near future. In the meantime, while I still agree with the general points of this essay, note that I am planning to edit and expand on them soon. -MWF 5/13/17]
A commonly misunderstood and complained about historical criterion among apologists is that of outside corroboration. “You do not demand that every single claim from every other ancient author outside of the New Testament be independently corroborated to trust their overall reliability!” This is both true and untrue. There are certain claims from Pagan and Christian authors alike that are dubious without outside corroboration and there are others that are perfectly believable, even if only one author reports them. Furthermore, apologists often dismiss any argument from silence out of hand as being fallacious, as if an argument from from silence is never valid. This response is itself fallacious, since arguments from silence can most certainly be valid under certain circumstances. For which historical claims should we demand outside corroboration in order to be trusted? Under what circumstances are arguments from silence valid?
There are at least three major circumstances under which outside corroboration should be sought. If no outside corroboration is found, then an argument from silence is a valid reason for doubting the historicity of the claim. All of these conditions vary by degree, where the more one of the conditions applies, the more outside corroboration is needed and the more an argument from silence carries force.
1. Multiple attestation would be expected, if the claim were true.
Imagine if I told you that a mile-long UFO flew over New York City yesterday in broad daylight before millions of people. Suppose then that you check the newspapers and discover that none have written about this incident, and thus I have no independent corroboration for my claim. Now suppose, when you tell me about this silence and say that you don’t believe me, because no one else had reported this incident, that I responded with, “Well, just because nobody else wrote about it doesn’t mean that it didn’t happen! Nobody wrote about you waking up this morning, and yet you would expect me to believe that you did if you told me so!” Very obviously my response would not be rational, but what are the key issues that justify skepticism about my claim?
Certain events are of such a nature to attract public attention and be reported by multiple witnesses. Others, of a more mundane or less public nature, very often go unreported, or are only briefly reported by a single witness. I would be very skeptical if only a single author reported Julius Caesar invading Italy in 49 BCE, but whether Caesar happened to shave the morning he was assassinated is a mundane detail that could very easily be true, and yet no authors would report. For claims about events that would be very prominent and attract a lot of attention, a historian should demand that multiple witnesses report the event. If there is only a single source to report a claim that should have attracted more attention, especially when that source already shows signs of being unreliable (to be discussed further below), then an argument from silence is a legitimate reason for doubting, or at least being suspicious of, such a claim.
First, it should be made clear that the Gospels of the New Testament are not independent sources. Matthew directly borrows as much as 80% of Mark’s verses, and Luke borrows 65%. In fact, as a Classics Ph.D. student, who regularly reads other Greek and Latin literature from antiquity, I do not know of any other set of ancient texts that show more interdependence than the Synoptic Gospels. While John does not follow the ipsissima verba of the earlier Gospels, the author(s) suggest awareness of at least the Gospel of Mark . This is very problematic for the historical reliability of the Gospels, since in many instances the authors merely borrow material from each other and thus do not independently attest to their shared claims.
The problem is exacerbated by the fact that the Gospels make many claims of a highly public nature that are not attested by any other ancient authors. For example, Mark 15:33, Matthew 27:45, and Luke 23:44 all claim that a three hour darkness at noon “covered the whole land” during Jesus’ crucifixion. This is indeed a very extraordinary claim and one of a very public nature. Matthew and Luke can be shown to derive this story from the Gospel of Mark, so they do not independently corroborate this event.
Given the public and incredible nature of this event, any historian worth her salt should seek outside corroboration. Do we have any for the three-hour midday darkness of the Gospels? In the Roman Empire there were hundreds of astrologers and natural scientists who would have immediately written about such an incredible event, and yet there is complete silence among contemporaries . If there was such a darkness, we would expect several contemporaries or near-contemporaries to record it, perhaps most notably Pliny the Elder in the second book of his Natural History, where he documents other astronomical abnormalities and eclipses. And yet there is complete and total silence. This would be extremely unlikely, if the darkness had actually occurred. Instead, the alternative hypothesis, given that we only have one source for this claim, namely the Gospel of Mark, is that the author merely reported an unsubstantiated rumor or even made up the entire episode out of whole cloth. It should also be noted that there would easily be an incentive to invent such a story, in order to draw parallels to Old Testament verses that describe the “Day of the Lord” as a day of darkness (e.g. Joel 2:1-2; Amos 5:18-20; Zephaniah 1:14-15).
In the case of Mark’s three-hour darkness, an argument from silence is more than sufficient for doubting this historicity of the claim, given that silence would not be expected if the claim were true.
The Gospels are littered with fantastical stories, many of a public nature, that curiously did not even make a single blip on the radar screens of Jesus’ contemporaries. This embarrassing silence, sufficient by itself for doubting many of the miracles of Jesus, was perhaps described best over a century ago in the words of Annie Besant in The Freethinker’s Textbook (pgs. 193-4):
“The most remarkable thing in the evidences afforded by profane history is their extreme paucity; the very existence of Jesus cannot be proved from contemporary documents. A child whose birth is heralded by a star which guides foreign sages to Judaea; a massacre of all the infants of a town within the Roman Empire by command of a subject king; a teacher who heals the leper, the blind, the deaf, the dumb, the lame, end who raises the moldering corpse, a King of the Jews entering Jerusalem in triumphal procession, without opposition from the Roman legions of Caesar; an accused ringleader of sedition arrested by his own countrymen; and handed over to the imperial governor; a rebel adjudged to death by Roman law; a three hours’ darkness over all the land; an earthquake breaking open graves and rending the temple veil; a number of ghosts wandering about Jerusalem; a crucified corpse rising again to life, and appearing to a crowd of above 500 people; a man risen from the dead ascending bodily into heaven without a concealment, and in the broad daylight, from a mountain near Jerusalem; all these marvelous events took place, we are told, and yet they have left no ripple on the current of contemporary history.”
While I do think that the later, non-contemporary sources we have are sufficient for establishing the relatively mundane claim of Jesus’ mortal existence, the paragraph above marshals a fatal argument from silence for doubting Jesus’ alleged miracles and the fantastical nature of his ministry described in the Gospels. For us to have no outside corroboration of these events, if they did actually take place, would indeed be a greater miracle than the events themselves! It would almost require a Cartesian demon to intervene to silence the Pagan contemporaries who would have corroborated such miracles in their midst and to suppress the evidence (very curious for a god who wanted us to know that he had intervened in history). Instead, the far more probable explanation is that they are merely later legends, which did not grace the papyri of contemporaries, since no such actual events occurred to be reported in the first place.
2. The claim is initially improbable or at tension with ordinary historical circumstances.
For events that are initially improbable, or of an extraordinary nature, outside corroboration can help overcome a low prior probability with greater expected evidence. This is a simple matter of Bayesian logic. If the prior probability of a claim is low, greater evidence will be needed to render its occurrence more probable. For certain well-established historical claims, if we did not have outside corroboration, we could no longer consider the claim to be as well-established or trustworthy as we do, given the low probability of such an event.
Let’s consider a non-biblical example:
A very interesting work to come down to us from antiquity is Lucian of Samosata’s Alexander the False Prophet (the introduction on the linked webpage even notes how outside corroborating material is crucial for trusting Lucian’s narrative). Alexander of Abonoteichus was a charismatic figure who managed to gather a following and start a new cult that worshiped a human-headed serpent named Glycon that was said to be the son of Apollo. Apparently, Alexander even had a large snake prepared for this charade that was trained to wear the puppet of a human head, in which there were pipes placed for people in concealment to channel their voice, so that it gave the impression that the snake was speaking.
Lucian reports many other incredible things about this Alexander, too numerous to elaborate on here, but one is that Alexander had even managed to convince the Roman emperor to mint coins of Glycon. This is very incredible indeed. Could an evangelizing, stage-show huckster like this really convince the Roman emperor to mint coins of his strange new snake deity? Lucian’s polemic might raise our suspicions, if it were not for the fact that we have actually independently corroborated his claim through a variety of archeological evidence. In fact, we have even found examples of the coins that were minted.
For events that are initially improbable, or of an extraordinary nature, outside corroboration can help overcome a low prior probability with greater expected evidence. That a charlatan like Alexander could have duped even the Roman emperor is initially improbable and we should demand greater evidence. It turns out, however, that in Lucian’s case we do have the greater evidence needed for trusting him.
The same does not often hold true for many of the stories in the Gospels. The Gospels make many claims with low prior probabilities that are then not substantiated by greater evidence. Furthermore, the Gospels also make claims that are at tension with what we know of the ordinary historical conditions of the time and region, which demands that we treat them with greater scrutiny.
I discussed one such example in an earlier post about the trial of Jesus and the release of the violent criminal Barabbas in the place of Jesus in the Gospels. We have no outside corroboration that Pilate would customarily release any prisoner that the crowd asked for during the Passover festival. There are outside reported instances where Roman governors would release certain prisoners on special occasions, but certainly no case where the Roman governor would regularly release any prisoner, including a violent revolutionary, simply at the capricious volition of the crowd during a holiday. Such a practice would be extremely improbable. It is furthermore not at harmony with what we know of Pontius Pilate outside of the Gospels, where he is described as a ruthless and merciless governor. The lack of outside corroboration for this custom immediately demands our suspicion. This suspicion is then further strengthened (which I discuss in the linked article) by the clear allegorical elements that would provide a basis for Mark to invent such an episode and custom (namely to mimic the Yom Kippur sacrifice, where one goat was released and another sacrificed to atone for sins). In the case of Barabbas, an argument from silence about Pilate’s custom is a valid initial reason for being skeptical of the episode, further strengthened by the evidence for it being an allegorical invention.
Another almost certainly fictitious story is in the Gospel of Matthew (2:13-20), when Herod the Great orders that all of the young infant males in Bethlehem be slaughtered, simply because three Magi had reported a rumor that the King of the Jews had been born there. In the episode, baby Jesus escapes from this slaughter when an angel appears to Joseph in a dream and tells him to flee into Egypt. This is a very extraordinary event, indeed, and if Herod had committed such an atrocity, I would expect other contemporaries to write about it. And yet Josephus, who provides a long catalog of Herod’s other crimes, is completely silent about it.
The event is furthermore initially improbable as an actual historical event. We do know that Herod was a violent and paranoid king, but to massacre an entire city’s infant males, merely at some vague rumor from a couple of Magi, is very unlikely even for Herod. We would need great evidence to trust such an unlikely claim, and yet we only have the Gospel of Matthew, which is very poor evidence. Instead, once more, an allegorical root of the story can be found. By having Herod slaughter the infants in Bethlehem, the author of Matthew could draw a parallel with Exodus (1:22-2:1-10), in which Pharaoh likewise has the male infants of the Israelites slaughtered. In the episode, baby Moses escapes Pharaoh’s massacre, just as baby Jesus escapes Herod’s. The parallel is strengthened by the author of Matthew explicitly having baby Jesus flee into Egypt. That such a coincidental parallel would take place is very unlikely and we would need greater evidence, such as outside corroboration, to trust it, but failing this criterion, we have strong reasons for being suspicious of this story based on a valid argument from silence.
3. A particular source repeatedly makes extraordinary claims that are not corroborated.
Now consider the lesson that we learn from the boy who repeatedly calls wolf. Each and every time his story is not substantiated by an outside witness he loses more and more credibility. He eventually is so untrustworthy that nobody even hesitates to doubt him when he shouts that a wolf is coming.
In like manner, the more a particular source fails the criterion of outside corroboration, the more the criterion increasingly becomes relevant and arguments from silence increase in their strength. If a source is otherwise independently corroborated for most of its extraordinary claims, but lacks such outside evidence for one claim in particular, we can grant it the benefit of the doubt. But when a source repeatedly makes extraordinary claims without corroboration, each new tall tale demands that we be more and more suspicious of the source and doubt most of these claims, unless we have very good reason not to.
Stephen Law sums up this problem nicely by what he calls the Contamination Principle. This principle in part entails that the more a text becomes contaminated by extraordinary claims that are uncorroborated, the more we should doubt them (and even mundane claims), without independent verification. In effect, a source eventually loses all credibility and can only be tug-boated along by more credible sources.
The Gospels are rife with extraordinary claims that Jesus’ contemporaries were completely silent about: a crazed king slaughtering of a whole town of infants, violent criminals being set free as part of a holiday festival, a midday darkness that covered the entire land right when a messianic figure was being crucified, and many other tall tales about this same miracle worker feeding whole crowds of people with only a few morsels, escaping his execution through a miraculous resurrection (curiously, with no subsequent manhunt or investigation in Acts), and then flying into space in broad daylight. And yet we hear none of this from even a single contemporary. From sources so contaminated by such unsubstantiated claims, we have a cumulative case for being skeptical of their content, particularly in the case of extraordinary events, but even for the more mundane ones. In the case of such texts, outside corroboration is especially needed, and historians are correct to demand it. Lacking such independent verification, arguments from silence bear particular weight and validity.
In short, apologists can complain about historians seeking outside corroboration for the Gospels, but the practice is both necessary and valid. The same would hold true for any other ancient text under the same circumstances. If the Gospels had outside corroboration for their claims about public events that would have drawn the attention of multiple witnesses, if the more improbable stories found within them were backed up by greater evidence, and if they had a greater track record for being independently corroborated when we would expect it, then perhaps we could treat them with greater credence. But alas, the opposite holds true for all three of the conditions I discuss above. Accordingly, arguments from silence are perfectly valid for critiquing the Gospels and provide a compelling reason for why we should doubt many of their claims.
 For a in-depth analysis of how John’s whole Passion narrative is written in response to the Gospel of Mark, see Louis Ruprecht’s This Tragic Gospel.
 For information about how Thallus does not record the darkness, see Richard Carrier’s “Thallus and the Darkness at Christ’s Death.” Furthermore, even if Thallus (whose date of writing is unknown, but may have been as late as the 2nd century CE) had claimed that the darkness was an eclipse, it would have been in response to Christian claims and propaganda. This would not make Thallus an outside or independent source, since his knowledge of the darkness would be dependent upon previous Christian claims. What is actually far more likely, however, is that Julius Africanus, trying to find an outside reference for the legendary darkness at Jesus’ death, falsely connected an irrelevant passage in Thallus about an eclipse and earthquake in Bithynia in 32 CE (also recorded by Phlegon) with the rumored darkness in Jerusalem in 30 or 33 CE. This merely means that Africanus made an error, or, worse, was completely unable to find an outside reference for the darkness, and thus had to misrepresent Thallus’ statement. Either way, Thallus does not by any stretch count as an independent reference for the darkness, which was almost certainly invented by the authors of the Gospels, to draw an allusion to OT verses (cf. Joel 2:1-2; Amos 5:18-20; Zephaniah 1:14-15).