CeleryKills

How Historians Decide Who Existed When Evidence Is Thin

We were sitting on a park bench, killing time the way you do when the weather is good and the conversation drifts. A friend of mine asked, half‑serious, half‑provocative, how you would actually demonstrate that someone like King Arthur or Robin Hood existed. Not whether the stories were meaningful, not whether they inspired people, but whether there was ever a flesh‑and‑blood person under the legend. The question lingered longer than either of us expected, because once you start pulling on that thread, you end up talking about method rather than heroes. What kind of evidence would you accept. What kind would you discount. And what do you do when the story is larger than the data.

That’s how the conversation slid, almost inevitably, to Jesus of Nazareth.

There are two statements that get offered as if they are worlds apart, but in practice sit uncomfortably close together. The first says: “A person existed who was later remembered and identified by followers as Jesus of Nazareth, and the tradition that this person was crucified under Roman authority is historically plausible.” The second says: “A belief emerged among early followers that a figure called Jesus of Nazareth existed and was crucified, but this belief may have developed without a corresponding historical individual.” On paper, one affirms a person, the other withholds that affirmation. In reality, they share far more ground than most people admit.

The first statement represents the mainstream historical consensus. Its basis is not theological enthusiasm, but a minimal explanatory hypothesis. Early Christian texts assume a founder. Non‑Christian writers like Tacitus report that Christians believed their founder was executed. The movement appears quickly and organizes itself around a named individual. For many historians, the existence of a marginal Galilean preacher who was executed under Roman authority explains these facts with fewer assumptions than wholesale invention (Meier, A Marginal Jew, 1991; Ehrman, Did Jesus Exist?, 2012). Crucifixion, in particular, is seen as a costly detail to fabricate, which gives it weight in historical reasoning.

One of the clearest pieces supporting the minimal hypothesis is the reference in Tacitus. Writing around 116 CE, Tacitus notes that “Christus” was executed under Pontius Pilate during the reign of Tiberius (Annals 15.44). Historians treat this cautiously, but not dismissively. Tacitus is hostile to Christians, has no incentive to flatter them, and situates the execution within a Roman administrative framework. Methodologically, this matters because it shows how historians weigh source hostility and contextual plausibility rather than proximity alone. Tacitus is not confirming miracles or theology; he is reporting what was understood as the origin point of a movement within Roman memory (Meier, A Marginal Jew, 1991).

Josephus provides a second, more complex case. While the so‑called Testimonium Flavianum is clearly interpolated, most scholars accept that a shorter, neutral core likely existed beneath later Christian additions. The passage referring to “James, the brother of Jesus who is called Christ” (Antiquities 20.9.1) is generally regarded as authentic. The methodological significance here is not that Josephus confirms Jesus’ life in detail, but that he references Jesus incidentally while discussing something else entirely. Historians give weight to such casual mentions, because they are less likely to be crafted apologetically and more likely to reflect assumed background knowledge (Grabbe, An Introduction to the Hebrew Bible, 2010).

Another often‑cited support comes not from external corroboration but from internal diversity within early Christian texts themselves. The earliest layers of tradition show disagreement, embarrassment, and attempts at reinterpretation rather than smooth myth construction. The crucifixion is a prime example. A messiah executed as a criminal is not an intuitive founding myth in a Second Temple Jewish context. Historians note that early texts struggle to explain the crucifixion’s meaning rather than celebrate it as a triumphant event. This fits a pattern where an unexpected historical outcome is reinterpreted theologically after the fact, rather than invented wholesale (Ehrman, Jesus Before the Gospels, 2016).

These instances support the historical method underlying the first statement by converging on a narrow conclusion through independent constraints. None of them establish certainty. None of them require theological assent. Instead, they reduce the number of assumptions needed to explain why a movement formed, why it fixated on an execution, and why that execution was remembered rather than erased. The hypothesis survives because it explains these features with fewer speculative steps than alternative models, not because the evidence is abundant or decisive.

That same method also explains why historians stop where they do. The lack of contemporary inscriptions, trial records, or material artifacts tied directly to Jesus sharply limits what can be responsibly claimed. The consensus does not expand to biography, teaching corpus, or miraculous events because the method does not allow it. The support for the hypothesis is cumulative and restrained, and its survival depends on that restraint rather than on any single decisive proof.

The weaknesses of this statement are also real. It rests heavily on movement‑generated texts and later reports of belief. There is no contemporary inscription, no administrative record, no material trace tied directly to Jesus. The phrase “historically plausible” does a lot of work, and it’s doing so precisely because certainty isn’t available. The hypothesis survives by being modest, not by being robust.

One weakness that consistently troubles historians is the temporal gap between the events being reconstructed and the sources describing them. Even the earliest Christian texts we have are written decades after the supposed execution. That gap matters methodologically, not because memory automatically fails, but because memory is shaped by community needs, theological development, and narrative consolidation. Historians of antiquity routinely downgrade confidence as that gap widens, especially when no contemporaneous external controls exist to stabilize the account (Ehrman, Jesus Before the Gospels, 2016). The hypothesis survives this gap, but only by narrowing what it claims to the bare minimum.

Another problem is source dependence. The early Christian texts are not independent witnesses in the way Roman administrative records or inscriptions would be. They are internally referential, shaped by shared traditions, shared scripture, and shared rhetorical goals. Even when multiple texts appear to attest the same event, historians must ask whether they represent multiple memories or multiple versions of the same tradition. This makes convergence weaker than it initially appears. The historical method compensates by lowering confidence rather than inflating it, which is why the conclusion remains phrased in terms of plausibility rather than assertion (Meier, A Marginal Jew, 1991).

There is also the issue of narrative saturation. Once a movement forms around a figure, later accounts tend to backfill biography, motive, and detail. This is a well‑documented phenomenon in ancient historiography and legend formation. The problem is not that invention occurred, but that invention and memory become difficult to disentangle once a narrative stabilizes. In the case of Jesus, the speed with which theological interpretation appears complicates attempts to recover earlier layers of tradition. Historians must constantly ask whether a detail explains the origin of the movement or is itself a product of the movement’s explanatory needs (Goodman, Rome and Jerusalem, 2007).

Finally, the absence of material corroboration does not disprove the hypothesis, but it sharply limits its scope. For non‑elite figures, lack of inscriptions or administrative records is expected. What is not expected is to treat that absence as neutral. It functions as a ceiling on confidence. Historians can infer existence because doing so requires fewer assumptions than total invention, but they cannot move beyond that inference without violating their own standards. This is why the statement remains deliberately tentative. It survives not because it is strong, but because it is constrained.

Taken together, these weaknesses explain why the first statement is phrased the way it is. It does not claim certainty. It does not claim detailed biography. It does not claim independent confirmation of events beyond execution. It persists as a hypothesis precisely because it resists expansion. The moment it overreaches, it collapses under the same scrutiny that keeps it alive.

Those same weaknesses are also what make the second statement viable as an alternative, which is where the mythicist position enters the conversation.

The second statement is closer to the mythicist position, most clearly articulated by Richard Carrier. Its foundation is methodological skepticism sharpened by Bayesian reasoning. Carrier argues that when you weigh prior probabilities, the silence of contemporary records, and the prevalence of mythmaking in antiquity, it may be more likely that early Christians developed a celestial or mythic savior and later historicized him (Carrier, On the Historicity of Jesus, 2014). In this view, crucifixion is not an event remembered, but a narrative constructed from scripture and revelation.

This position’s strength is that it takes bias seriously. It refuses to let cultural familiarity stand in for evidence. It asks why we are so quick to infer historicity in one case and not in others. It also forces historians to justify every step of their reasoning rather than rely on intuition shaped by centuries of Christian dominance.

Carrier’s distinctive move is not simply skepticism, but formalization. Instead of relying on intuitive judgments about what “seems likely,” he tries to translate historical reasoning into Bayesian terms. The core idea is simple even if the execution is controversial. You start with prior probabilities based on background knowledge of the ancient world. You then update those priors as evidence is introduced. In Carrier’s framework, the question is not “did Jesus exist,” but “given what we know about ancient religious movements, how probable is historicity versus mythic origin once all the evidence is accounted for” (Carrier, On the Historicity of Jesus, 2014).

One of Carrier’s strongest challenges is aimed directly at the historical habit of treating existence as the default. In Bayesian terms, historicity is not the null hypothesis. It is one hypothesis among others, and it must earn probability weight. Carrier assigns relatively low priors to the existence of obscure, non‑elite individuals unless there is direct attestation. He contrasts this with the frequency of savior deities, dying‑and‑rising figures, and revelatory cults in the Hellenistic and Jewish worlds. In his calculation, mythic formation is not exotic. It is statistically common. That background distribution matters, he argues, when evaluating how much weight to give later narrative sources.

Carrier also places heavy emphasis on what he calls “silence,” though this is often misunderstood. He does not treat silence as proof of non‑existence. He treats it as missing expected evidence. In Bayesian terms, if a historical individual existed and played the role later attributed to Jesus, Carrier argues we would expect at least some trace in contemporary records, correspondence, or disputes. The absence of such traces does not falsify historicity, but it lowers its posterior probability relative to mythic explanations. This directly challenges a common historical assumption that absence of evidence is neutral rather than evidentially negative (Carrier, Proving History, 2012).

For readers unfamiliar with Bayesian reasoning, the math itself is less exotic than it sounds. At its core, Bayes’ theorem asks a simple question: given what we already know about the world, how much should new evidence change our confidence in a hypothesis. Carrier applies this by assigning prior probabilities to competing explanations. One hypothesis is that a historical Jesus existed and was crucified. Another is that early Christians developed a mythic savior later placed into history. Carrier then asks what evidence we would expect under each scenario. For example, if a historical preacher existed and was executed, how likely is it that we would find independent contemporary references. If the figure were mythic, how likely is it that early texts would emphasize revelation and scripture rather than eyewitness memory. The absence of expected evidence does not set the probability to zero, but it lowers it. In simplified terms, if historicity starts with a prior probability of, say, one in three, and the available evidence fits mythic development slightly better than historical memory, the posterior probability shifts downward. This challenges a common historical habit of treating existence as the default and forces every assumption to carry numerical weight rather than rhetorical momentum. That pressure is exactly what Carrier intends his Bayesian framework to apply.

Crucifixion becomes the most contested node in this analysis. Mainstream historians treat it as a costly detail unlikely to be invented. Carrier pushes back by reframing the cost. He argues that scriptural exegesis, particularly of Isaiah and the Psalms, already provided a template for a suffering, executed savior. In that context, crucifixion is not an embarrassment reluctantly accepted, but a narrative solution to theological problems. In Bayesian terms, what others treat as improbable invention, Carrier treats as expected mythmaking given the scriptural environment. The disagreement here is not about data, but about how heavily to weight comparative religious patterns.

Where Carrier’s method is strongest is in forcing explicit accounting. Every assumption has to be named. Every intuition has to be justified numerically, even if the numbers themselves are approximate. That alone is disruptive to traditional historiography, which often relies on trained judgment rather than formal probability. Carrier’s critics argue that the numbers are too subjective to bear the weight he places on them, and that Bayesian formalism gives a false sense of precision. Carrier responds that all historical reasoning already uses informal Bayesian logic, and that making it explicit exposes hidden biases rather than creating new ones (Carrier, On the Historicity of Jesus, 2014).

This is where the tension with mainstream scholarship becomes clearest. Traditional historians accept uncertainty but resist quantification. Carrier accepts uncertainty but insists on quantification. The conflict is methodological rather than theological. Most historians agree that some form of Bayesian reasoning is unavoidable. They disagree that it can be operationalized with enough reliability to overturn long‑standing minimal hypotheses (Goodman, Rome and Jerusalem, 2007).

What matters for the comparison between the two statements is that Carrier’s approach does not simply deny the consensus. It reweights it. Under his model, the first statement’s modesty is not enough. Even a minimal historical individual, he argues, carries a higher assumption cost than many historians acknowledge. That is why his conclusion diverges, not because he rejects evidence outright, but because he assigns different prior probabilities to the same evidentiary landscape.

That methodological divergence is precisely what makes the second statement viable, even if most scholars ultimately reject it. It survives not as an alternative theology, but as a challenge to how historical plausibility itself is calculated.

Its weaknesses are just as clear. Denying existence requires explaining why hostile sources refer to a specific execution, why no ancient counter‑tradition claims invention, and why the movement’s earliest texts behave as if arguing about the meaning of a death rather than inventing one. Many historians find that mythicism, while possible, introduces its own stack of assumptions (Goodman, Rome and Jerusalem, 2007).

The appeal to the absence of an ancient counter‑tradition is often treated as decisive, but it’s less straightforward than it first appears. What counts as a counter‑tradition depends on what kind of invention one expects. There is no surviving ancient text that says “Jesus was made up,” but that standard would erase large swaths of ancient mythmaking from consideration. Movements rarely generate contemporaneous refutations that deny existence outright. More often, they generate competing interpretations, rival narratives, or internal fractures. Those do exist in the Jesus tradition, though they are usually framed theologically rather than ontologically.

Some scholars point to hypotheses like Caesar’s Messiah as an attempt to account for this absence by reframing the context entirely. While that specific thesis is widely rejected, it illustrates a broader methodological point: counter‑tradition does not have to appear as explicit denial to function as explanatory pressure. If narratives emerge that reshape, domesticate, or redirect the meaning of a figure in ways aligned with power structures, that itself can act as a form of narrative displacement rather than direct contradiction. The weakness of such hypotheses is not that they ask the question, but that they require speculative coordination on a scale the evidence does not sustain.

A more grounded version of this pressure comes from examining the internal diversity of early Christianity itself. Rather than a single, stable account of Jesus, we see at least three distinct narrative trajectories across the four canonical gospels, with differing chronologies, theologies, and emphases. Mark’s Jesus is stark and apocalyptic. Matthew’s is scripturally fulfilled. Luke’s is orderly and universal. John’s is cosmic and preexistent. These are not minor variations. They reflect different communities negotiating meaning, authority, and identity. That diversity can be read as evidence of memory being shaped by theology, or as evidence of a historical core being interpreted in divergent ways. The data alone does not force one conclusion.

Paul complicates the picture further. His letters are the earliest Christian texts we have, and they are strikingly uninterested in Jesus’ earthly biography. Paul focuses on revelation, scripture, and the meaning of Christ’s death, not on teaching, parables, or miracles. For mythicists, this supports the idea of an originally celestial figure later historicized. For mainstream historians, it reflects the genre and purpose of Paul’s writing rather than ignorance of a historical person. Methodologically, this is one of the tightest knots in the debate, because the same data supports different priors.

What all of this shows is that the absence of a clean counter‑tradition does not automatically privilege historicity, but neither does narrative diversity collapse it. The evidence is underdetermined. Historians fill that gap with judgments about what kinds of explanations are more economical given known patterns of ancient religious development. That is why the debate persists. Not because one side is ignoring data, but because the data allows more than one constrained reconstruction.

This is also why the disagreement remains methodological rather than theological. The question is not whether early Christians believed these things. That is certain. The question is whether belief requires a historical anchor or can arise from scriptural exegesis, visionary experience, and communal mythmaking alone. Different weighting of those factors produces different conclusions without either side stepping outside historical practice.

When you strip away theological commitments and cultural reflexes, the distance between these two statements shrinks dramatically. Both agree that what we have direct access to is belief. Both accept that narratives were shaped by followers. Both deny that miracles or resurrection can be established historically. The disagreement turns on whether the minimal hypothesis of a historical individual is more economical than the hypothesis of invention. That’s a methodological judgment, not a revelation.

This is where Cleopatra Selene II becomes a useful foil. She lived at roughly the same time as Jesus, but the evidentiary landscape is entirely different. Coins bear her image. Inscriptions name her titles. Buildings and administrative records anchor her existence firmly in the material world. There is little narrative about her inner life, but no serious historian doubts she existed. That’s elite history. Power leaves artifacts. Peasants do not.

Jesus, if he existed, belonged to the class of people Rome routinely erased. That fact cuts both ways. It explains the absence of material evidence, but it also lowers the threshold at which historians infer existence. The standards shift not because of bias toward Christianity, but because ancient history has to work with what survives.

I’ll be transparent about my own leanings. I tend toward the mythicist side, not because I think it’s proven, but because I’m comfortable with Bayesian reasoning and deeply suspicious of how Western cultural saturation affects our priors. When a figure dominates two thousand years of art, politics, and morality, it becomes very hard to imagine that figure not existing. That psychological weight matters. It doesn’t decide the question, but it bends intuition.

What keeps me from treating either statement as settled is that both are hypotheses constrained by method. Neither is a theory in the scientific sense. Neither escapes uncertainty. And importantly, neither implies the theological claims people usually care about. Divinity, resurrection, cosmic significance all sit entirely outside this debate.

So when people argue as if denying the first statement is an attack on faith, or affirming it is a concession to religion, they’re missing the point. The real question is not who Jesus was, but how historians reason when evidence is sparse, bias is real, and cultural gravity is enormous.

In that sense, the conversation with my friend about King Arthur and Robin Hood never really ended. It just got heavier. And maybe that’s the honest place to leave it. Not with certainty, but with clearer boundaries around what can be claimed, what cannot, and how much of our confidence comes from habit rather than data.

References

• Carrier, Richard. On the Historicity of Jesus: Why We Might Have Reason for Doubt. 2014.
• Carrier, Richard. Proving History: Bayes’s Theorem and the Quest for the Historical Jesus. 2012.
• Ehrman, Bart D. Did Jesus Exist? The Historical Argument for Jesus of Nazareth. 2012.
• Ehrman, Bart D. Jesus Before the Gospels: How the Earliest Christians Remembered, Changed, and Invented Their Stories of the Savior. 2016.
• Goodman, Martin. Rome and Jerusalem: The Clash of Ancient Civilizations. 2007.
• Grabbe, Lester L. An Introduction to the Hebrew Bible. 2010.
• Meier, John P. A Marginal Jew: Rethinking the Historical Jesus. 1991.
• Atwill, Joseph. Caesar’s Messiah: The Roman Conspiracy to Invent Jesus. 2005.
• The Holy Bible. The Gospels according to Mark, Matthew, Luke, and John.
  (Standard critical editions: NA28 / UBS5; English translations vary.)
• Tacitus. Annals. Book 15, Chapter 44.