Primary Sources: An Overview

For evidence to count as ‘reliable primary evidence,’ Carrier considers two criteria:

  1. “it must be plausibly capable of being causally connected with the facts (persons, properties or events) whose existence is in question, and
  2. “it must be relevantly independent of all other primary evidence” (p. 254)

I will not here go through Carrier’s analysis of these two criteria except to point out an important fact concerning the second of them. “If someone wrote something about Peregrinus for which their sole source of information was Lucian’s Death of Peregrinus, then only the latter counts as evidence. A fact or quotation taken from that book can be evidence of that book’s existence (or of its content at a particular point in time), but it is not evidence of what happened to Peregrinus, as long as we have the earlier (which is in our case the primary) source, which is the text of Lucian’s book” (pp. 255-256). The likelihood of the later source containing information from the primary source is the same whether or not the primary source was accurate. Therefore, the later, derivative sources must be considered irrelevant to the comparison of hypotheses about what originally happened.

This conclusion is still valid even if we cannot establish that a source is independent. That is, a later work that cites the same information as the primary source, if we cannot demonstrate that the later work did not derive from the primary source, cannot be counted as corroboration (though note that the later evidence can be evidence of how widespread the knowledge of the original story was, but that is a different issue). This is a particularly vexing problem when it comes to evidence for or against the historicity of Jesus, because most such evidence cannot be established as independent of earlier sources.

Let us look at the types of evidence that are available that could bear on historicity vs. mythicism. First, there is no directly-relevant archaeological evidence. The Shroud of Turn is a medieval fake. The so-called James ossuary is inconclusive and also probably at least partially faked. The Talpiot Tomb, and other tombs alleged to have been Jesus’s, “almost certainly [have] no connection to Jesus Christ” (p. 257). Archaeological evidence for incidental historical details, such as the existence of Pontius Pilate, is equally expected under the minimal historicity and minimal mythicism hypotheses, and thus do not in any way help to distinguish between them. In short, there is nothing at all from archaeology that qualifies as reliable primary evidence. This leaves only one type of evidence: “texts (books, letters, etc.), which are reconstructed (by modern scholars) from a variety of manuscripts (which are copies of copies of copies of even earlier originals now lost)” (p. 258).

Carrier further limits the sources he will evaluate thusly: “I will consider only texts that are known to have been written (or probably written) before 120 CE (or that record information from an identifiable source before that date), as after that time we can’t reasonably expect there to have been any surviving witnesses to the original decade of the cult’s creation (in the 30s CE), due to the limits of life expectancy…; and also because after that time the quantity of bogus literature about Jesus and early Christianity exploded to an immense scale, making the task of sorting truth from fiction effectively impossible” (pp. 258-259).

What texts remain relevant after taking these considerations into account fall into three main categories:

  1. Things not written by Christians. This category of materials is quite small, and most of it cannot be established to be independent of earlier works written by Christians (such as the canonical Gospels). What material does remain in this category will be discussed briefly below and considered in detail in the “Extrabiblical Evidence” essay.
  2. Things written by Christians outside of the New Testament. This category is also quite small, as much of what exists is either not independent of the New Testament or is not plausibly connected to anything that is likely to be true about Jesus (g., the Infancy Gospels). The remaining material will also be discussed briefly below and will be considered in detail in the “Extrabiblical Evidence” essay.
  3. The New Testament. Except for the last section below, which surveys the extrabiblical evidence, the dating and authorship of the New Testament documents will be discussed below, and will be considered in detail in a series of essays broken up in the same subheadings I use below.

In my opinion, the most effective approach to working our way through the New Testament evidence is to go through it chronologically. That way we can trace the evolution of the religion avoiding assumptions resulting form later developments. Unfortunately, according to Carrier, “There is a great deal wrong with how a ‘consensus’ has been reached on the dates and authorship of all these Christian materials, and the conclusions usually cited as established tend to be far more questionable than most scholars let on. Nevertheless, as I found the task of trying to sort this out impossible, I will mostly rely on the majority consensus for no other reason than that I haven’t anything better to work with. I projected it would take a minimum of seven years of full-time research to adequately examine and analyze all the evidence and arguments regarding the dating and authorship of these materials, and that with no actual prospect of any clear resolution for any of it, since it is not simply a given that we actually know the answers to any debated question. I think there is a lot of work here that needs to be properly redone in NT studies, and I’m not alone in thinking that.” (p. 260) I will, below, go through Carrier’s evaluation of the consensus that he will work from for the rest of his book.

The Epistles

“Seven letters are commonly agreed to be authentically written by the apostle Paul in or around the 50s CE. Those are 1 and 2 Corinthians, Galatians, Romans, Philippians, 1 Thessalonians, and Philemon. With the possible exception of Philemon (which has no relevant content for our purposes), I think that assessment is probably correct—although we know that even these seven letters have been meddled with” (pp. 260-261). In a footnote, Carrier expands on the dating, stating that Paul wrote before the Jewish war in 66 CE, and probably before the Neronian persecution of 64 (if that was indeed a historical event). He also wrote well after Aretas assumed control of Damascus (between 37 and 40 CE), and most of his writings came fourteen to seventeen years after his conversion, all of which supports the dating of his letters to be in the 50s CE.

Of the remaining Pauline epistles, 2 Thessalonians, Ephesians, and Colossians are generally agreed to be forgeries, “although they could include pastiches or edits or redactions of Pauline letters, and are close enough in outlook and content to be of the Pauline ‘school’, perhaps from the latter half of the first century” (p. 262), and 1 and 2 Timothy and Titus are “even more certainly forgeries (definitely not from the Pauline ‘school’, and of significantly later date)” (p. 262). Since they are generally considered forgeries, they cannot provide evidence in favor of historicity, though they could provide evidence against it since any such evidence they contains couldn’t have been by design.

Of the non-Pauline Epistles, several are of uncertain authorship, date, and reliability (1, 2, and 3 John, Jude, James, and Hebrews; none of the authors identify themselves). In particular, the Johannines are late, and while the others could be earlier there isn’t any good evidence to suggest that they are. Of these, the best case could be made for Hebrews to have been written in or near Paul’s lifetime. For the rest, like the non-authentic Pauline letters, these cannot provide evidence in favor of historicity.

The remaining Epistles are 1 and 2 Peter. 2 Peter is clearly written by a different author than 1 Peter, and thus must be a forgery, and thus discarded. The scholarly consensus is that 1 Peter is likewise a forgery, though by a different author than 2 Peter. Carrier, however, considers it more likely than the consensus would suggest that 1 Peter is authentic, and thus should be included in the analysis with the authentic Pauline letters. The main reason the scholarly consensus considers 1 Peter to be a forgery, according to Carrier, is that “Peter was an illiterate fisherman, but that is information only the Gospels produce, and they have every reason to invent or exaggerate the humble origins of the cult’s founder (so as to make their appeal to the masses and their subsequent brilliance look all the more miraculous), whereas based on every precedent in history, prior probability heavily favors any religious leader and founder of the period being educated” (p. 263). However, Carrier doesn’t claim to have proven this thesis, and thus will consider 1 Peter to be uncertain.

Finally, there are letters from Jesus in Revelation, “which purports to record an extensive hallucination of the celestial Jesus by an unknown person named John, who composed this book like an extended epistle to various churches of his time” (p. 264). Carrier therefore treats Revelation as an Epistle as well. The scholarly community still debates its date and authorship, as well as the extent of redaction it has undergone. The most likely date for it is the early 90s CE due to its veiled commentary on the reign of Domitian. Given this late date, its dependence on the Gospels, and its character as an elaborate fiction, Carrier concludes that this book has little evidential value.

The Gospels

“The ‘usual’ consensus on the four canonical Gospels is that Mark was written around 70, Matthew around 80, Luke around 90, and John around 100 [CE]. Those are all arbitrary ballpark figures, which don’t really have much basis in fact. Of course, fundamentalists want all those dates to be earlier, while many well-informed experts are certain they are later, and I find the arguments of the latter more persuasive, if inconclusive. As to authorship, none of the Gospels was written by the person they were named after, or in fact by any known person…. The titles of the Gospels conspicuously assign them as ‘according to’ the names given…, which designation in Greek was not used to name the author of a work, but its source, the person from whom the information was received or learned, as in ‘according to [the tradition of] Mark’. And none of the Gospels says who these people are (nowhere is the titular Matthew or John ever called a disciple, e.g., or identified in any way)…. it must also be noted that all of the Gospels have undergone redactional activity after their original composition so really they’ve been meddled with by multiple authors” (pp. 264-265).

Mark pervasively assumes that the Jewish War has passed and that God has abandoned the temple cult. Mark also includes an apologetic for the failure of the apocalypse to have arrived, so it must have been written decades after Jesus predicted it. Paul’s authentic epistles show no knowledge of the stories in Mark. All in all, this and other evidence cited by Carrier dates Mark at least 70 CE, and possibly later. The main evidence that Mark isn’t significantly later is that the author “seems to have the Jewish War still in mind as a relatively recent event, but that’s largely a subjective judgement” (p. 266). Late tradition holds that the author was Peter’s secretary, but there is no real evidence to support that tradition, and given that Mark is advocating against Torah-observant Christianity, he would have been Peter’s opponent in that debate.

Matthew was clearly written after Mark, since he copies Mark, often verbatim. There is no real evidence otherwise for when Matthew was written, so Carrier decides to assume that it was written in the 80s or 90s CE to err as early as possible. Nothing is known of the author, except that he was not an eyewitness, “because he copies mark verbatim and just modifies and adds to him (and much of what he adds is ridiculous and literarily crafted, not eyewitness material…), which is not the behavior of a witness, but of a late literary redactor” (p. 267).

Luke also copied from Mark, and thus must come after. Carrier is “increasingly convinced that in fact Luke is also a redaction of Matthew and therefore postdates Matthew” (p. 267). There is also significant evidence that Luke got some of his material from the Jewish Antiquities of Josephus, which would date Luke after 93 CE. Numerous experts argue that Luke-Acts are later than 115 CE, but Carrier decides, as he did with Matthew, to err on the side of earlier dates, in the 90s. As with Matthew, nothing is known of the author of Luke. Tradition claims that he was a doctor and Paul’s companion, but there is no evidence to support this claim. If the author had had Paul as an authority, it would seem that the author “would have trumpeted those credentials” (p. 268), which he did not.

The scholarly consensus is clear that John wrote after Luke. Some evidence suggests that it was written as late as the 140s CE, though some argue that it was as early as the 100s. Carrier again decides to side with the earlier of these dates to give as much benefit-of-the-doubt as possible to the historicity hypothesis. Scholars agree that John had multiple authors.

Finally, scholars often cite an early document called ‘Q’, for Quelle, German for ‘source.’ This is, however, a hypothetical document, not a document that we have any direct evidence for. Carrier finds the arguments for ‘Q’ to be highly questionable. “[I]t looks far more likely to me that what we call ‘Q’ was nothing more than additions made to Mark by Matthew, which were then redacted into Luke” (p. 269). Since we have no direct evidence of this document, if it even existed, it is clear that appealing to it cannot be considered to be evidence in favor of historicity.


Acts can be thought of as the ‘sequel’ to the Gospel of Luke, chronicling the early development of the church, at least as Luke wanted to represent it, ranging from the 30s up to around 60 CE. Scholars debate the authorship and timeframe of Acts, but the overall consensus, with which Carrier agrees, is that the same author wrote both around the same time, but that there may have been significant later editing. For example, there are extant two versions of Luke-Acts, one which is 10-20% longer than the other (the shorter is the canonized one), and scholars disagree on which is the original. Because it must have been written after the Gospel of Luke, which puts Acts after the early 90s CE, and possibly several decades later.

Extrabiblical Evidence

Carrier mostly follows the scholarly consensus on dating and authorship for the extrabiblical documents summarized by Van Voorst in Jesus Outside the New Testament (Grand Rapids, MI: William B. Eerdmans, 2000).

A great deal of extrabiblical Christian literature is undatable, such as the Didache and 1 Clement. Carrier goes into detail about 1 Clement, which he says has been traditionally dated to around the year 95 CE, but which for various reasons Carrier finds implausible, favoring an earlier date in the 60s. Regardless of the conclusion about this particular document, however, 1 Clement is the earliest even modestly datable Christian text outside of the New Testament. And since these other texts cannot be established to be early, they also cannot be established to be independent. None of these other texts “contains anything either different from what’s found in the NT or that’s believable even if different (e.g. the Infancy Gospels or the Acts of Peter or the bizarrely improbable sayings attributed to Jesus in the Gospel of Thomas). One can speculate that some of it did [predate the Gospels], but that’s insufficient to build an argument from… ‘speculation in, speculation out’” (p. 273). Therefore, Carrier concludes that these documents cannot be used as evidence in favor of historicity, though in a manner similar to the forged Epistles they might contain evidence against. Hence, Carrier only looks at a few key examples of this in detail, including 1 Clement, Ignatius and Papias, and Hegesippus). Everything else dates later than 120 CE, and so is discarded.

Concerning the non-Christian extrabiblical evidence, most also cannot be established as independent and reliable. When we go through this evidence in detail in a later essay, we will explore Carrier’s argument that none of this evidence can be considered independent and reliable, although this is contrary to the usual view that some of it can be. In fact, Carrier will be arguing that the argument from silence (discussed in my essay on historical methodology) is the most compelling argument that can be drawn from the non-Christian extrabiblical evidence. The particular documents that Carrier will examine for positive evidence either way include:

  • The Babylonian Talmud: This is a very late source, completed in the fifth century CE. Carrier includes it in his analysis because it says things that are quite different from what is said by the New Testament, and therefore appears to be independent. Most scholars contend that because the document is so late, it doesn’t trace its origins back to any real first-century source; Carrier, on the other hand, thinks there is a case to be made that it is a response to a first-century non-canonical Gospel that is now lost, and thus might have some limited relevance.
  • The Jewish Antiquities of Josephus: Completed in or shortly after 93 CE, and which contains two references to Jesus.
  • A letter of Pliny the Younger: Written around 112 CE, this letter concerns Pliny’s interrogation of some local Christians in what is now northern Turkey.
  • The Annals of Tacitus: Concerning the Neronian persecution at Rome in 64 CE, this document was completed around 116 CE.
  • The Lives of the Caesars by Suetonius: Completed around 119 CE, this document includes a discussion of ‘Chrestus,’ who started a riot in Rome. Some take this figure to mean Christ.
  • Histories of Thallus: This document is undatable with certainty, but probably comes from the late second century.

All other authors, such as Celsus, Lucian, and Mara bar Serapion, all wrote significantly later, and thus are excluded. No other non-Christian texts both mention Jesus and predate these. That’s all there is.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s