HistoricalChristian.Faith

Do the traditional names of Jesus’ sisters — Mary and Salome — preserve genuine early memory?

Type: historical-tradition argument (off-timeline) · Assessment: plausible — post-canonical tradition names the two sisters of Jesus (Mark 6:3) Mary and Salome, attested in Epiphanius (Panarion 78) and, before him, in the Gospel of Philip (“Mary his sister”) and probably the Protevangelium of James (the unexplained “Salome” at 19:3). Richard Bauckham’s statistical case is striking and largely unrefuted: Mary and Salome were the two commonest Palestinian Jewish women’s names of the period (together ~47.7%), while Salome was vanishingly rare in the diaspora — so a second-century Gentile inventor would not have hit on it, and the names most likely trace to Palestinian Jewish-Christian memory. On its best case that conclusion holds, and the sisters probably played some modest part in the early church. But the names’ authenticity is neutral between the three readings of the “brethren” — sisters exist on all of them — so it supports the Epiphanian household picture only indirectly, and does not establish that the sisters were Joseph’s daughters by a former wife · argued from Richard Bauckham (Ilan 1989; Mayer 1987), with the primary attestations

Related arguments: the onomastic argument — see also Are the shared names too much of a coincidence?

See also the companion argument on whether the second-century stepbrother tradition is genuine memory or a Marian construct — the same question of what the earliest extra-canonical family tradition is worth, approached from the sisters rather than the brothers.

The claim

The Gospels record that Jesus had sisters — “are not his sisters here with us?” (Mark 6:3 ∥ Matthew 13:56) — but never name them. Post-canonical tradition does: the best-attested names are Mary and Salome. Epiphanius gives them as the two daughters of Joseph by his first wife (Panarion 78:8:1; 78:9:6), the same reconstruction that anchors the Epiphanian view that the “brethren” were Joseph’s children by a former marriage.¹ Two witnesses earlier than Epiphanius can be read to carry the same names. The Gospel of Philip — a Valentinian work of probably the third century — states: “There were three who always walked with the Lord: Mary his mother and her sister and Magdalene, the one who was called his companion. His sister and his mother and his companion were each a Mary.” Made consistent, the “her sister” of the first clause must be corrected to his sister (as the second clause shows), yielding a sister of Jesus named Mary.² And the Protevangelium of James (late second century) introduces a “Salome” at the birth of Jesus (19:3) with no explanation of who she is — not even “a woman called Salome,” but simply “Salome” — as though the readers already knew her; the best account is that she is Salome the daughter of Joseph, naturally supposed to have travelled with her father and brothers.³

Richard Bauckham — himself holding the perpetual virginity on the Epiphanian reading — argues that this convergence rests on real Palestinian Jewish-Christian memory rather than on later invention. That is the interpretation weighed here: the traditional names of Jesus’ sisters, Mary and Salome, preserve genuine early tradition — and so lend weight to the extra-canonical picture of the Lord’s household.

The evidence for — Bauckham’s statistical case

The core of the argument is an argument from onomastics, and it is genuinely striking. A study of the recorded names of Jewish women in Palestine (330 B.C.–A.D. 200) finds that of 247 women whose names are known, bearing 68 different names in all, 61 were called Salome (including the longer Salomezion) and 58 were called Mary (Mariamme/Maria) — so these two names alone account for 47.7% of the women. “Every second Palestinian Jewish woman must have been called either Salome or Mary,” their popularity owing to their being borne in the Hasmonean royal house.⁴⁵

From this Bauckham draws two conclusions. First, a bare probability: “the tradition which gives the names Mary and Salome to Jesus’ sisters has a 50% chance of being correct, even if it was not based on historical memory” — the names are so common that a random guess would land them half the time. Second, and decisively, the frequencies outside Palestine cut the other way. In Mayer’s list of 769 named Jewish women from Palestine and the diaspora together, at least 21 women named Mary lived in the diaspora, but only two named Salome — “Salome seems to have been a peculiarly Palestinian name.” So a Gentile Christian in the second-century diaspora, wishing to invent names for Jesus’ sisters, “may well have called one sister Mary, but is unlikely to have called the other Salome.” The presence of Salome therefore points back to Palestinian Jewish-Christian circles — and “if the tradition is therefore to be traced back to Palestinian Jewish Christian circles, the chances of its being an accurate historical tradition are relatively high.”⁴

Two further observations reinforce the pedigree. The names are not a late back-formation from Mark 15:40 (the women at the cross): “there would be no reason to suppose that the women mentioned in Mark 15:40 were sisters of Jesus unless the sisters of Jesus were already known as Mary and Salome,” so the tradition of the names “must antedate their identification with the women of Mark 15:40.” And the names were not widely known in the wider church — the later writers who repeat them (Sophronius, Anastasius of Sinai, Theophylact, and others) all depend on Epiphanius — which tells against invention: a fabricator’s names would be expected to spread, whereas a genuine but locally-preserved memory would remain the property of one branch of tradition, exactly as these do.⁴

What the argument does — and does not — establish

1. It shows the names are plausibly, even probably, authentic — not the sisters’ parentage. Bauckham is careful, and honesty requires the same care: “simply to know, with some degree of probability, the names of two sisters of Jesus is in itself of very little historical interest.” What follows from authentic names is only that the sisters “were known figures in some early Christian circles… must have played some part in the early Christian movement” — a modest part, since the tradition stayed local. The statistic bears on which names the tradition preserved, not on whose daughters the sisters were. That the sisters were Joseph’s children by a former wife is a separate claim, carried by the same infancy-gospel milieu but not proven by the onomastic argument.

2. The names are neutral between the three readings of the “brethren.” This is the decisive limit on the argument’s bearing. Jesus had sisters on every account of the family: on the Helvidian reading they are Mary’s own daughters, on the Epiphanian reading Joseph’s daughters by a former wife, on the Hieronymian reading kinswomen. Authentic names for them — Mary and Salome — are equally at home in all three. So the argument cannot, by itself, favour the Epiphanian household over the Helvidian one; it establishes the antiquity and Palestinian roots of a family datum, not the disputed relationship. Its support for the perpetual virginity is therefore indirect: it comes only insofar as it raises the general credit of the early extra-canonical family tradition — the tradition in which the stepbrother reconstruction also travels.

3. That indirect support is real but hostage to the tradition-cluster question. The names travel in the same second-century Palestinian-Syrian stream as the stepbrother reading, Mary’s Davidic descent, and the miraculous birth. If that stream preserves genuine memory, the sister-names are one more witness to its reliability, and the Epiphanian picture gains. But that is precisely the point left open by the companion argument on the antiquity of the stepbrother tradition: the same cluster “could preserve an accurate historical memory… [or] could be a legendary construct… We cannot be sure.” The onomastic case strengthens one thread of that cluster without settling the cluster as a whole — the sister-names are better-pedigreed than most of what surrounds them, because the statistics give them an independent check the other beliefs lack.

Assessment

Plausible. Bauckham’s statistical argument is the strongest single piece of evidence that any part of the extra-canonical family tradition rests on real Palestinian Jewish-Christian memory. It is not an argument from authority or from piety but from independent onomastic data: Salome’s near-absence from the diaspora makes it genuinely unlikely that a Gentile inventor would have chosen it, and the tradition’s failure to spread beyond one branch fits memory better than fabrication. On its own terms — are these the names the earliest tradition preserved for Jesus’ sisters? — the case is well-grounded and arguably probable, not merely possible.

But two limits keep the finding at plausible rather than higher, and they are limits on what is shown, not defects in the reasoning. First, the argument establishes the names, not the sisters’ parentage: Bauckham himself grants that knowing the names is “of very little historical interest” beyond showing the sisters were known figures who played some part in the early church. Second, and more important for this doctrine, the names are neutral between the Helvidian, Epiphanian, and Hieronymian readings — Jesus had sisters on all three — so the argument lends the perpetual virginity no direct weight. Its bearing is wholly indirect: it raises the credit of the early Palestinian family tradition in which the stepbrother reading also stands, and so far as that tradition is trusted, the Epiphanian household gains a little. That indirect support is genuine and better-secured than most of the cluster it belongs to — but it remains hostage to the unresolved question of whether that cluster is memory or construct. The sister-names are, in short, a plausibly authentic datum that strengthens the general picture of the significant role the relatives of Jesus played in the early church, without deciding the identity of the “brethren.”

One guardrail keeps the verdict honest: plausible is not a verdict of invention. Nothing here suggests the names were fabricated; the evidence leans the other way, and a fair reader may judge the names probably authentic. What the argument cannot do is convert authentic names into a proof of the Epiphanian relationship — and it is only that relationship, not the bare fact of named sisters, on which the perpetual virginity turns.

Sources