Is “Josaphat” a linguistic corruption of “bodhisattva”?

Type: supporting argument (off-timeline) · Assessment: sound for the pivotal name (Josaphat ← bodhisattva) — the phonetic steps are each independently attested and the decisive slip is a documented Arabic scribal phenomenon; only the mentor’s name (Barlaam) has an unsettled ultimate etymon · advanced by nineteenth–twentieth-century philology (Laboulaye, Liebrecht, Lang, Asmussen).

See also (on another doctrine): the same accident — a single mis-copied letter, as a legend crossed languages, manufacturing a saint’s identity — shaped another figure in this family: why is St Christopher dog-headed — and “Canaanite”? There a descriptive adjective (canineus, “dog-like”) was misread as an ethnonym (Cananeus, “Canaanite”), the two a hair apart in Latin minuscule; here a Buddhist title (bodhisattva) was worn into a personal name, the pivotal slip an Arabic bāʾ/yāʾ one dot apart. Same mechanism, opposite direction: a category made into a lineage, a title made into a person.

The claim

— 1913 Catholic Encyclopedia, “Barlaam and Josaphat.” The claim is that the saint’s name is not an accidental look-alike but a worn-down form of the Buddha’s title bodhisattva, carried down a language-by-language chain: Sanskrit bodhisattva → Middle Persian Bōdisav → Arabic Būdhāsaf / Yūdhāsaf → Georgian Iodasaph → Greek Ioasaph → Latin Iosaphat/Josaphat. A parallel chain is proposed for the hermit: Arabic Bilawhar → Georgian Balahvar → Greek Varlaam/Barlaam.

Who advances it

The identification was worked out independently by Édouard Laboulaye (1859) and Felix Liebrecht (1860), stated flatly by the 1913 Catholic Encyclopedia, and reconstructed language-by-language in the twentieth century above all by David Marshall Lang (with F. C. Conybeare, Ilia Abuladze) and, for the Iranian and Arabic legs, by Jes P. Asmussen in the Encyclopædia Iranica.¹ It is now the settled scholarly account, not a fringe conjecture; the question this page tests is whether every link is genuinely documented or whether a stage is reconstructed only to bridge a gap.

The evidence for

Each step is attested, not merely posited. The name did not have to be re-derived from bodhisattva by a modern scholar hunting for a resemblance; the intermediate forms actually turn up in the texts along the route. The Buddha’s title survives as Middle Persian Bōdisav / Bwdysdf in the Manichaean fragments recovered from Turfan — the same missionary community whose Uighur, Parthian, and Middle Persian copies of the tale prove they were the ones who carried it west — so the Iranian stage rests on manuscript witnesses, not on a bare reconstruction.² The Georgian Iodasaph is fixed by the surviving Balavariani, the Greek Ioasaph by Euthymius the Athonite’s version, the Latin Iosaphat by the 1048 translation and the Golden Legend — each a datable text one can open.

The pivotal slip is a known scribal mechanism, and both forms survive. The one step that looks like special pleading — the change of the initial b to y — is in fact the best-supported of all. In the Arabic script the letters bāʾ (ﺑ) and yāʾ (ﻳ) share an identical skeleton and are distinguished only by dots below the line; adding or misreading a single dot turns Būdhāsaf into Yūdhāsaf. This is not a hypothesised error to save the theory: the Arabic manuscripts of the Kitāb Bilawhar wa-Būdhāsaf preserve both variants, Būdhāsaf and Yūdhāsaf, side by side — exactly the split the dot-ambiguity predicts, and it is the y-form that fed the Georgian, Greek, and Latin.³ The tašḥīf (dot-confusion) of bāʾ and yāʾ is one of the commonest and best-documented copyist errors in Arabic transmission generally, so the mechanism is independently ordinary, not a one-off invoked here.

Adversarial assessment

What holds. For the load-bearing name — Josaphat — the chain is documented at every joint and the crux is a routine, textually-witnessed scribal fact. Nothing about it requires a coincidence, a missing link, or a form reconstructed purely to reach the answer. The plot-parallel (the sheltered prince, the four encounters, the Great Renunciation) already makes a genetic relationship likely; the name-chain converts that likelihood into a demonstrated genealogy, because a random look-alike would not also drag along the whole Middle-Persian-to-Arabic phonetic history and the surviving b/y doublet.

What is less certain. The parallel derivation of Barlaam is weaker, and honesty requires saying where. The western half of that chain is secure — Arabic Bilawhar → Georgian Balahvar → Greek Barlaam is a clean phonetic descent through the same texts as Josaphat. What is not settled is Bilawhar’s own ultimate origin. The traditional guess derives it from Sanskrit bhagavan (“Lord,” an epithet of the Buddha), but that step is unproven and phonetically loose; Almuth Degener has instead proposed Sanskrit purohita (a priestly title) through a hypothetical Middle Persian intermediate.⁴ This gap, however, does not touch the core identification: Barlaam is the invented Christian ascetic-teacher who has no counterpart in the Buddha’s life at all, so his name need not go back to any Buddhist word for Josaphat to be the Buddha. A second, smaller caveat: the exact sixth-/seventh-century Middle Persian form is partly reconstructed from the Manichaean spellings rather than preserved in a dated colophon — but the Manichaean attestation of the Bōdisav name-form is real, so this is a matter of precise vocalisation, not of whether the stage existed.

Assessment

Sound. The identification of Josaphat as a corruption of bodhisattva is not a suggestive resemblance but a chain documented at each stage, turning on a scribal ambiguity (bāʾ/yāʾ) that the surviving Būdhāsaf/Yūdhāsaf doublet directly confirms; it is rightly treated as established. The rating attaches to that pivotal name. The mentor’s name (Barlaam ← Bilawhar) descends just as cleanly through the medieval texts, but its ultimate Sanskrit etymon — bhagavan or purohita — is genuinely open; that residual uncertainty is a footnote to a secondary name, not a crack in the core claim, and should not be advertised as if the whole identification rested on it.

Sources