📄 The Magi in Ethiopic Tradition

Summary

Witakowski surveys how Ge'ez literature expands on the brief account of the Magi in Matthew 2:1–12. The canonical passage leaves many questions open: the visitors' homeland, number, names, timing, the appearance of the star, and how they learned of Christ's birth. The paper compares the answers given across a wide range of Ethiopic texts, both translated and locally composed. These include the Gospel translation, apocrypha (Life of Adam and Eve, Miracles of Jesus, Book of the Birth of Mary, Miracles of Mary), homiliaries, and two Gospel commentaries. It then tries to trace where each motif came from. Many of these texts cannot be dated, and their stories often contradict each other. Some elements are local inventions, such as certain lists of names. Others, the author argues, arrived from outside, mostly through Arabic intermediaries, and ultimately go back to Syriac apocryphal and exegetical writing. Examples include the virgin and child seen inside the star, gifts taken from Adam's Cave of Treasures, and a huge armed escort. The study shows that Ethiopia preserved older Eastern Christian traditions and also kept reshaping them.

Key points

• The journal header gives the source as Aethiopica 2 (1999), pp. 69–89, not 2013. The study began as a talk at the 1995 Oxford Patristics Conference.
• The Ethiopic Gospel renders Greek μάγοι as mäsäggəlan ("magicians"), with the variant säbʾa sägäl. In other words, the translators chose the popular sense over "wise men." The feast falls on 29 Tahśaś, which is 25 December in the Julian calendar and 4 January in the Gregorian.
• The Miracles of Jesus is an Ethiopic version of an Arabic apocryphal Gospel of John, translated in the 14th–15th centuries. Its details include:
  • a star shaped like a column;
  • 3000 companions;
  • a journey lasting two years;
  • barley bread given to the Magi by Mary.
• The Ethiopic Miracles of Mary grew to over 600 episodes, against 74 in the standard Arabic collection. Five of these episodes involve the Magi, and only one of the five has been published.
• The twelve Magi names given by Ibn aṭ-Ṭayyib are garbled forms of Syriac names. The author attributes the damage to diacritic loss during Arabic copying, for example Bel becoming Nil. One name, Hormizdad son of Sanatruq, dropped out entirely.

Interesting ideas

• I1. The translators' choice between the two senses of μάγοι fixed one reading for the whole later tradition, so a single word-sense decision can shape everything downstream.
• I2. Corrupted names are explained by a concrete mechanism, lost diacritical dots, which makes the variant forms predictable rather than random noise.
• I3. Contradictory stories about the same event are catalogued side by side, each tied to its source text, instead of being merged into one version.
• I4. The same feast maps to different dates depending on the calendar used, so a date always needs its reckoning system.
• I5. The author identifies a motif's origin by matching distinctive details across texts, such as the child in the star or the Cave of Treasures, not by overall similarity.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

• I1 (tokens): Record the lemma's sense choice with its alternatives and reason, so later stages can trace interpretations back to that decision.
• I2 (world): Entity linking could model systematic corruption channels, such as character confusions or transliteration chains, to match name variants to one entity with an explainable path.
• I3 (world): The text world could store competing claims about one entity with their provenance, rather than overwriting, so answers can say which source asserts what.
• I4 (world): Time normalization should attach a calendar or frame tag to dates before ordering events, avoiding silent mismatches.
• I5 (learn): When judging silver data or repair rules, score agreement on rare, distinctive structures rather than global overlap, as a sharper provenance and quality signal.