Why do we trust some historical sources but not others? How am I justified in believing with high confidence that Alexander the Great existed, but doubting the existence of someone like Romulus or Moses? Do I have good grounds for being more skeptical of the Gospels than other ancient sources? As a follow up to my previous blog, I thought it would be worthwhile to lay out some historiographical criteria by which ancient historians evaluate historical claims and pass judgments about their probability of occurring. Frequently, I am asked by apologists why I doubt the reliability of the Gospels. Occasionally, I am accused of being credulous of other ancient sources, but unfairly doubting biblical ones. Below I provide seven historiographical criteria that aim to put these issues in perspective and provide solid reasons for being skeptical of the New Testament accounts. The seven criteria I provide are by no means fully exhaustive; so feel free to propose other dimensions that help us evaluate historical claims and work out how to answer apologists on these issues.
1. Distance from the Event to to the Record
Historians generally look for contemporaneous sources to events rather than later accounts. The operating principle is simple: the further an event is from the one recording it, the more likely it is that clear understanding of the event has been distorted by a lack of memory, later embellishment, and ultimately the transformation of the past into legend. Take the history of regal Rome, for example. Archeological findings, such as the Lapis Niger, have confirmed that Rome was once ruled under kings (c. 753 – 509 BCE). However, Fabius Pictor (our earliest known historical source to write about these events) did not record the regal period until around 200 BCE. Because there had been a gap of around 300-600 years between the event and the one recording it, clear knowledge of this period had drifted out of social memory and had gradually come to be replaced by fanciful legends (e.g. Romulus being exposed on the Tiber and nursing from a she wolf).
The same pattern happens in the Bible. Biblical scholars generally agree that the Pentateuch was composed in stages between c. 1000 – 500 BCE. Since the events described in the Pentateuch (such as the Adam and Eve down to the Egyptian Exodus) would have spanned from thousands to at least hundreds of years before the earliest recording, the same problem occurs where clear recollection of the past had drifted out of memory, had become embellished, and had eventually been replaced by fanciful legends (e.g. Moses being sent down the Nile in a story very similar to Romulus). Because of the distance of time from the event to the record, the vast majority of historians have doubted the historicity of both Romulus and Moses, even if these legends have some faint memory in a distant past (even then, Romulus is actually much more likely to have existed, since there is at least evidence of Roman kings but no evidence of a Jewish Exodus in Egypt).
In the case of the Gospels, they are generally agreed to have been written 40 to 60 years after Jesus’ death. Many historians argue that this is too short of a time for Jesus to have developed as a purely legendary figure like Moses. However, as biblical scholar Perkins points out, the Gospels do not provide “eyewitness or contemporary accounts of Jesus’ life and teachings” (Oxford Annotated Bible, 1744). Thus, there is still a substantial gap of time from the event to its recording.
Furthermore, legendary developments can easily occur within such a time span. Here is a well-researched article written by Kris Komarnitsky analyzing the rapidity of myth growth rates and the Gospels, showing how the historical core could easily have been lost due to later hearsay, rumors, and embellishments. See also my essay “When Do Contemporary or Early Sources Matter in Ancient History?.”
2. First-Hand vs. Second-Hand Accounts
Next when evaluating a particular source that records an event, historians question the author’s relationship to the event. Did the author actually witness the event? Or did the author learn of the event second-hand? Apologists like to argue that the Gospels provide first-hand, personal accounts, clinging to the authorial traditions given by the church in the 2nd century. However, these traditions, as all skeptics should know, the large majority of mainstream NT scholars doubt these authorial traditions on multiple grounds. To name just a few: The Gospels are completely anonymous and do not claim to be witnesses, nor do the authors even imply such through the text. Jesus’ inner circle of followers spoke Aramaic and were probably illiterate, which does not align with the literate Greek authors who wrote the Gospels. Furthermore, the Gospels of Matthew and Luke are heavily dependent on Mark’s material, which suggests that they were not independent sources writing about personal experiences, but were instead later authors revising an earlier tradition. Finally, when we consider the claims of the 2nd century church fathers in context, there were many spurious authorial traditions floating around at the time because Christians were battling over which scriptures to accept as canonical. During this process, outside critics would attribute or the authors themselves would claim various scriptures to be written by the early apostles (e.g. the Gospel of Peter) in order to establish grounds for canonical status. Some of these were eventually accepted by the church, others were rejected, but that does not change the fact that many spurious attributions were created, which very likely included the attributions of Matthew, Mark, Luke, and John.
For the above reasons (in addition to many others), leading scholars like J.R.C. Cousland (OAB 1746), Richard Horsley (OAB 1791), Marion Soards (OAB 1829), and Jerome Neyrey (OAB 1876) reject the canonical authorial traditions, and instead maintain that the Gospels were written by anonymous authors who provide second-hand accounts.
For a further analysis of why most New Testament scholars doubt the traditional authors and “eyewitness” status of the Gospels, see here.
3. Oral vs. Textual Sources
Even if an author is writing about an event second-hand, the reliability of his or her source material can still enable him or her to furnish an accurate account. For example, the historical biographer Plutarch (living in the 2nd century CE) wrote a biography of Alexander the Great (living in the 4th century BCE). Plutarch was a second-hand source writing about Alexander at a distance of nearly half a millennium. How could Plutarch possibly furnish a reliable account of Alexander’s life? The answer is simple: Plutarch followed contemporaneous written sources, such as Alexander’s letters, the speeches of Demosthenes, and the lost works of Alexander’s biographers, when he constructed his own account. Therefore, even though Plutarch’s version was second-hand and very far from the event, we have good reason to trust it, since Plutarch made use of written material. Written material provides a good source, because, once something is written on paper on stone, the original material can be preserved so that it does not get changed over time.
Oral sources, however, are much less reliable. When people pass on stories by word of mouth, every time the story is told is a new rendition, and changes can very easily and very quickly creep into the original. Anyone who has played the elementary school game “Telephone” has experienced this. Unlike Plutarch, who quotes Alexander’s letters and other contemporary sources, the Gospel authors are virtually silent on where they obtained their material (even though we can tell that they copied from each other and thus probably had little outside material). Mark, who provides the earlier source, seems to be patching together various oral traditions and Christian doctrines. One exception outside of Mark that scholars have identified based on some shared passages in Matthew and Luke is the probability of an early Q Gospel (short for the German Quelle “source”). However, this seems to be nothing more than a collection of saying attributed to Jesus, which would not provide early documentation of Jesus’ miracles. Thus, what remains, as far as original source material in the Gospels, are oral traditions and doctrines that were passed down for about half a century. Given how those repeating the stories had a religious, rather than a historical, interest in the subject matter, it is even more conceivable that embellishments inevitably crept in during this process. The result is that we have a substantial gap of telephone between Jesus and the earliest narratives about him.
For a further analysis of how we have far more reliable historical sources for Alexander the Great than Jesus, due in part to Plutarch’s and Arrian’s access to contemporary written documents, see here.
4. Genre of Literature
So far I have presented my view that the Gospels are not contemporary sources, are second-hand, and are largely derived from oral tradition. Given these difficulties, it is questionable whether a later author could provide an accurate narrative, even if he or she were trying to. But a further problem arises when we consider whether the authors even intended to provide researched and historical accounts. This is why it is important to remember that the Gospels in genre are hagiographies, or what some scholars call “laudatory biographies.” As Pheme Perkins elaborates, “neither the evangelists nor their first readers engaged in historical analysis. Their aim was to confirm Christian faith” (OAB 1744). Given the fact that the authors’ goal was to encourage faith in Christianity, and that their readers were not concerned with evaluating the text critically (at least not in a historic sense), it is highly conceivable that the authors took license in embellishing the events and adjusting the oral material to suite their narrative aims. This would not be uncharacteristic for the genre of a laudatory biography, because the real goal of the genre is to provide praise of the subject, rather than historical information. Therefore, when a modern historian evaluates a laudatory biography to find historical information, he or she must not, by definition, take the material at face value.
For a further analysis of the Gospels’ literary style and how they are not historical in their purpose, genre, or character, see here.
5. Authorial Bias
A laudatory biography is biased toward praising its subject and does not have to provide researched information to achieve this. Given the Gospels’ bias toward promoting Jesus as the Messiah, it is not surprising that they fabricated their narratives so that Jesus fulfilled certain messianic prophecies. For example, both Matthew and Luke provide genealogies of Jesus attempting to show that he had descended from King David to fulfill the prophecy in Jeremiah 23:5. Beyond the fact that the two genealogies contradict each other, that they do not provide enough names to fill in the time gap between David and Jesus, and, most importantly, that they do not specify where they obtained their source (I mean, did Joseph have a detailed and faithfully recorded pedigree in his wood shack?), the authors furthermore have an invested bias in linking Jesus with David. Therefore, I see no reason to doubt that the authors of Matthew and Luke merely invented their genealogies. An apologist could object that I am being too skeptical, but forging genealogies was actually a common practice in antiquity. For example, Suetonius records that the Roman emperor Galba put on display an elaborate pedigree that traced back his ancestry to the Cretan king Minos (Gal. 2.1), so that he could establish his noble heritage. I see no reason to trust the Gospels’ account of Jesus descending from an ancient Israeli king than Galba’s claim to have descended from an ancient Cretan king. Both have biased motives that would lead them to make this assertion.
6. Authorial License
Beyond the bare-bone details of the Gospels, the authors further shaped a cohesive and story-like narrative. Thus, even if some of their anecdotes came from reliable sources, they still had to flesh out the material in between them. During this process, an author could easily fill-in and invent certain material to suite their narrative aims. This, once more, is not a pattern unique to the Gospels. The Greek philosopher Plato, for example, wrote dialogues featuring his former teacher Socrates. Socrates was a real person and Plato was actually a first-hand witness of him. However, in order to flesh out the dialogues and speeches of Socrates, Plato had to inevitably invent much of the material. Accordingly, Platonic scholars have long battled over how much of the “real” Socrates vs. “Plato’s” Socrates is found in the dialogues. The same pattern occurs in the Gospel of John. John has long been considered the most vivid Gospel that provides more intimate details than the other three. However, I believe that this vividness reveals that John is simply took more creative license than the others and probably invented material in order to achieve its narrative cohesion. Therefore, the authorial license in John, as well as the other Gospels, is a further good reason not to take the Gospel’s material at face value
7. Plausibility vs. Probability
Finally, and I think most importantly, before I can believe an event probably occurred, I first have to consider the intrinsic likelihood that an event would happen. In Bayesian terms, this is referred to as the prior probability. Strictly speaking, one might argue that all things are possible except logical contradictions (e.g. a square circle), but certain events have a much higher prior probability (e.g. the sun rising tomorrow) than others (e.g. me flapping my arms and flying tomorrow). Following from this principle, the more that an event is rare or outside the realm of normal experience, the greater the evidence it needs in order to overcome its low prior. Put simply: extraordinary claims require extraordinary evidence.
Applying the principle of uniformitarianism (that things operate in the present as they did in the past), we can assess the prior probability that things would occur a certain way in the past based on present observation. So, if an author claims that the rate of gravity on earth occurred at 1 meter per second squared (as opposed to 9.8 meters), I would be automatically skeptical of the claim, since this would be highly outside of the realm of normal experience. Many of Jesus’ miracles include things that likewise contradict observable patterns in nature (e.g. walking on water, and rising from the dead after being crucified and pierced in the side with a spear). The extraordinary rareness of such events (ones that would have never occurred before and contradict everything we know about physics and biology) entails an extraordinarily low prior probability that they would occur, and thus places an extraordinary onus of evidence that would be needed to verify them. Personally, I do not believe any ancient historical account can match this burden. Ancient history is simply not a precise enough epistemological method, and such, I would only believe such claims when verified through the hard sciences, which are our most precise methods.
Occasionally, apologists argue that skeptics have to provide counter histories claiming that the events did not happen, citing how no Pagans denied Jesus’ miracles until centuries later (even though most of them had never even heard of such obscure rumors). However, I think that this completely distorts the burden of evidence. If X is extraordinarily unlikely, I do not need to provide a counter history stating that X did not occur, especially when only weak evidence is given for it in the first place. I will give you a non-biblical example to demonstrate this. The Roman historian Tacitus records, among various other supernatural prodigies, which occurred in the year 69 CE during the brief reign of Otho, which lasted from January to April, that a cow in Etruria had actually spoken in human language (Hist. 1.86). Tacitus’ claim of a cow speaking is very precise, as I can trace it to a specific year, a span of only a few months, and a specific location (moreover Tacitus is actually a historical source).
I doubt this claim, not because I have a contrary history stating otherwise, but because I know that it is extraordinarily unlikely that cows would be able to speak in human language today! And by the principle of uniformitarianism, I can infer that this was likewise extraordinarily unlikely 2,000 years ago. Can I supply a contrary history from another contemporary author stating, during the year 69 CE, between the months of January to April, in Etruria, that the cow did, in fact, not speak? None that I know of. Does this mean that, by default, I should accept Tacitus’ claim about the speaking cow? Of course not! Extraordinary claims require extraordinary evidence. If cows cannot speak today, it is perfectly reasonable to infer that they could not speak in Tacitus’ time. Applying the same principle of uniformitarianism, I can infer that, since people cannot walk on water today, people could likewise probably not do so in Jesus’ time. My “bias” in disregarding Jesus’ miracle is no different than my “bias” in disregarding Tacitus’ speaking cow.
You may say that this is unfair, since I am assigning a priori probabilities that make Jesus’ miracles appear extraordinarily improbable. How then could you then possibly convince me of the miracles? Simple: provide a contemporary example (or some tangible evidence) demonstrating that X actually does have a higher prior probability and is not next to impossible. Once more, I will provide a non-biblical example. Herodotus (3.151-3) records that during a Persian siege of Babylon, a mule gave birth to another mule. When I first read this, I thought that mules are completely sterile, so naturally I doubted Herodotus’ claim. However, I later found this scientific publication, which discovered that a female mule can, in fact, sometimes give birth:
By the standard of uniformitarianism, I must now concede that Herodotus’ claim was at least plausible (I’m still not sure if Herodotus had a reliable source for this detail). Therefore, in accordance with the same standard of evidence, if someone were able to walk on water today and scientists were able to observe that he or she was genuinely creating enough friction on the surface of the water to support his or her own weight, then I would concede that it was plausible that Jesus could likewise have done so 2,000 years ago. I make this challenge to Christian’s often: If Jesus can do anything (Matthew 19:26), then bring him here today and have him demonstrate under scientific observation that he is, in fact, performing miracles. Then I would believe them. But ancient, anonymous, and dubious laudatory biographies are simply not enough.
Many Christians think that this is a bizarre or unfair request, but I think that it is perfectly reasonable. After all, god is supposed to be able to do anything, so why is he always hiding from us, especially if he wants us to know him? The impotence of apologists to provide modern proof of miracles or their deity is the very reason that they have to rely on the authority of ancient scriptures in the first place. So their god ultimately becomes a god of history, or a god of the gaps of science, completely invisible in the world we see today.
For a further evaluation of what I would consider to be sufficient evidence to reasonably believe in the resurrection of Jesus, see here. In the essay I lay out the multiple epistemic problems surrounding the resurrection claim and also explain how reasonable doubt toward the resurrection is based on evidential, not solely presuppositional grounds.