Hinam Press · The Stayed Hand

The Source Dossier

The Stayed Hand argues from documents, and a book that argues from documents owes the reader the documents. The printed book keeps one in its own binding — the Aqedah pizmon Et Shaʿarei Ratzon, with the author’s translation. The rest of the exhibit room lives on this page: the Molekh question, the case for the northern Ur, the Tanakh witnesses and their editions, the Mesha stele with every lacuna left open, Philo of Byblos, and Diodorus on Carthage — each with its rights and its editions stated. Corrections reported through the book’s errata page reach this dossier too.

Appendix A: The Molekh Question

[Working apparatus: the technical argument deferred from Chapters 1 and 2, kept out of the main line so the chapters could move. To verify at the §10 pass: Eissfeldt (1935) thesis details; Heider and Day rebuttals; Dewrell’s synthesis; the Punic mlk inscription corpus (KAI; the molk adam / molk immor formulae); the MT vocalization argument.]

The word “Molekh” appears eight times in the Tanakh, and a lawyer likes to see the whole docket before the argument starts: five times in the Molekh laws of Leviticus (18:21; 20:2, 3, 4, 5 — the statutes Chapter 1 read); once in Shlomo’s apostasy list, where the text calls him “Molekh, the abomination of the Ammonites” (1 Kgs 11:7); once at the demolition of the Tophet (2 Kgs 23:10, “so that no man might pass his son or his daughter through the fire to Molekh”); and once in Yirmiyahu’s third denial (32:35, where the high places of Baal and the passing to Molekh stand in one verse). The reading handed down — and the one the nineteenth-century critics inherited — took the word for a name: a god to whom children were passed in fire. (Not every ancient reader did. The translators of Alexandria are the great exception, and the note two paragraphs on prints their Greek.) The traditional account added a philological flourish that the critics would later repurpose: the vowels. The consonants are מלך (the ordinary word for king, melekh), and the vocalization Molekh looks like the Masoretes’ known practice of dressing shameful names in the vowels of boshet, “shame,” as they did to Ashtoret and to the boshet-names in Shmuel. On that account the god’s real name was simply “the King,” and the tradition defaced it in the pointing. (Chapter 1 met the same consonantal ambiguity from the other side, in Isaiah’s la-melekh pun at 30:33, where the ear cannot tell the pyre’s royal victim from its god. The ambiguity is ancient, and nothing here turns on it: king or Molekh, the fire is the fire.) Then, in 1935, Otto Eissfeldt published a short monograph that redrew the map. In the Punic votive inscriptions, the very stelae of Chapter 2’s precinct, the consonants mlk carry a sacrificial sense: a molk-offering, with compounds the scholarship reads as molk adam (a molk of a human) and molk immor (a molk of a lamb: the substitution visible in the vocabulary itself). The formulas deserve a moment, because the substitution economy of this whole book is visible inside them: the stelae distinguish a molk adam (a molk of a human) from a molk immor, a molk of a lamb, the same dedication vocabulary with the victim swapped, which is to say the Carthaginian precinct kept, in its own technical language, the record of the choice this book’s Chapter 2 watched Diodorus narrate. Eissfeldt argued that biblical “Molekh” was the same word — the rite’s name, “to pass one’s seed as a molk” — misread as a demon once the living institution was dead and the Masoretes, centuries later, pointed an opaque technical term with the shame-vowels they kept for idols.

The stones themselves can be seen. A case of them stands in the Louvre: Punic stelae from Cirta (Constantine), bought for the museum out of the collection of Lazare Costa, the very corpus whose publication by Berger and Chabot set this argument running. Crescents, raised hands, and the triangular sign of Tanit cut into soft local stone. And their labels deserve to be read in their own characters. The dedications run in a fixed votive frame: to the lady, to Tanit, face of Baʿal, and to the lord, to Baʿal Ḥammon — that which [name] son of [name] vowed, because he heard his voice and blessed him. Into that frame, certain stones insert the disputed term: מלך אמר, mlk ʾmr (CIS I 307) — on Eissfeldt’s reading, a molk of a lamb; מלך אדם, mlk ʾdm — a molk of a man; and only a single bare mlk ever entered the argument. On late stelae from Ngaous, in Roman Algeria, the same institution surfaces in Latin dress: morchomor, molchomor (Latin ears spelling the Punic phrase), beside the talion formula anima pro anima, soul for soul, and the two words that could stand as this book’s epigraph from the other side of the sea: agnum pro vicario. A lamb as the substitute.¹²

Two things in Eissfeldt’s monograph deserve to be read from the monograph itself, because the secondary literature has flattened both. The first is that he did not accept the shame-vowels story at all. The derivation of mōlekh from melekh by way of boshet he took to be Geiger’s hypothesis and no more, and he said what he thought of the supports under it: the claims advanced with it — “for instance that of the identity of ‘Moloch’ with Milkom, the god of the Ammonites” — “have long been abandoned,” and the arguments of Geiger and his followers “prove on closer inspection to be brittle.”³ That is worth saying plainly because this book has just rehearsed the defacement story as the traditional account, and the scholar most often credited with dismantling the god considered the defacement story unproven. The vowels are contested at both ends.

The second is a paragraph almost nobody quotes, and it happens to state the argument of this book’s Chapter 8 thirty-seven years before Weinfeld’s footnote said it again. Eissfeldt observes that only the older laws contain the demand for the human firstborn, and then asks why the younger ones do not. His answer is a containment reading: Deuteronomy and the Holiness Code “stem from a time of determined struggle against every possibility of child sacrifice, and are themselves participants in that struggle,” and for that reason “make no mention whatever” of the firstborn demand. Why not simply restate it with its redemption clause attached? Because — and this is the sentence — they “renounced that demand because it, even in the mitigation afforded by the possibility or the obligation of redemption, could have been understood as a justification of child sacrifice, and would then have hindered their efforts toward the abolition of this sacrifice.”⁴ A law so dangerous that even its merciful clause was too dangerous to print: the redemption price could be heard as a tariff, and a tariff implies a thing being priced. Chapter 8 argued the centralization statutes were a quarantine; Chapter 10 argued the five shekels were the mercy. Eissfeldt’s paragraph says the quarantine was so strict that in the latest codes it swallowed the mercy too: the demand and its remedy struck together, because you cannot cancel an institution while publishing its rate card.

The stones themselves can now be read here rather than described, because the standard epigraphic edition prints them. Three from Donner and Röllig will do. The first is the frame with the term inside it, from Constantine, and it is worth setting out whole, because every element this book has been circling is in four lines:

לאדן לבעל חמן ולרבת לתנת פען בעל נדר אש נדר חמלכת בבעשתרת בן נבל מלך אדם בשערם בתם כ שמע קלא ברכיא lʾdn lbʿl ḥmn wlrbt ltnt pʿn bʿl ndr ʾsh ndr ḥmlkt bbʿshtrt bn nbl mlk ʾdm bshʿrm btm k shmʿ qlʾ brkyʾ “To the lord, to Baʿal Ḥammon, and to the lady, to Tinnit face-of-Baʿal: the vow which Ḥimilkat son of Boʿashtart, son of Nabal, vowed — a molk ʾadam, of his own flesh, complete. Because he heard his voice, he blessed him.” — KAI 105

Molk ʾadam: a molk of a man. And then the phrase that has kept scholars up at night, בשערם בתם — of his own flesh, complete — which reads, on Eissfeldt’s own construction of it, as the stone distinguishing this offering from the other kind: not a substitute, but the vower’s own, and entire. I print his reading because this is his argument; I do not print it as settled, because it is not.⁵ Beside it stands the other kind, from Malta:

נצב מלך אמר אש שם ארש לבעל [חמן] אדן [כ שמע קל] [דב]רי ntzb mlk ʾmr ʾsh shm ʾrsh lbʿl [ḥmn] ʾdn [k shmʿ ql] [db]ry “Stele of a molk ʾimmor which Arish set up for Baʿal [Ḥammon], lord, [because he heard the voice] of his [wor]ds.” — KAI 61 B

Molk ʾimmor: a molk of a lamb. Two stones, one formula, one word changed, and the word that changes is the victim. Everything this book argues about the ram on Moriah and the five shekels in the kohen’s palm is legible in the difference between those two inscriptions, with the difference that Carthage kept both options open on the same shelf, priced, while the Torah closed one of them.⁶

And a caution the editors themselves would insist on. KAI’s restorations are printed in brackets, and where the stone is simply gone the edition prints dots rather than guesses, including at the end of the Mesha stele’s line 17, where it declines to supply the noun that many translations confidently render. Donner and Röllig put the warning in their preface, in a sentence any book quoting them ought to reproduce: at all such places, they write, one must be emphatically warned against thoughtless adoption.⁷

The thesis has everything a scholarly legend needs: a stroke of comparative brilliance, a scandalous payoff, and a genuine evidentiary base. The Punic usage is real, and it permanently joined the biblical and Carthaginian dossiers, which is why this book’s Chapter 2 stands partly on it. But the strong form was answered — in Heider’s monograph of 1985 and in Day’s of 1989 — and for a generation that answer was the field’s. Their two books assembled the counter-case: the Tanakh’s syntax repeatedly treats Molekh as a recipient (la-Molekh, “to Molekh,” parallel to “to Baal” and never naturally to a rite-name: one does not whore “after” a sacrificial category, and Lev 20:5 speaks of straying after the Molekh); a deity Malik/Milku is independently attested: in Mesopotamian god-lists, where Malik is equated with Nergal, lord of the underworld (the identification runs from Lagrange 1903 through Heider’s reassessment, The Cult of Molek, pp. 32ff.), and at Ugarit, where a mlk appears with what look like chthonic characteristics, giving biblical Molekh a possible divine pedigree among the gods of the dead; and the “Molekh of the Ammonites” in Kings is most naturally a cousin or conflation of Ammonite Milkom, the same divine family under the same royal title. On this reading the Masoretes defaced a real name, and the fire’s recipient was a king indeed — the king below.

Then the answer was answered. The most recent monograph on the question is Dewrell’s, of 2017, and the section of it that takes up this point is headed “‘Molek’ Disappears Once Again.” Neither of the two exhibits the revival rests on, he argues, survives inspection. The Malik dossier is thinner than Heider made it: the god-lists stand, and the Ugaritic mlk does look like evidence for such a deity, but the Mari half of the file does not survive — the mālikū there are deceased rulers rather than a god, and Heider’s headline witness, a Standard Babylonian list that calls Malik “the king of Mari,” is a parody text whose whole joke is pairing gods with cities they had nothing to do with, so that for the joke to land Malik must not have belonged to Mari. And of the deity that does survive, Dewrell says there is no evidence whatever that he was ever associated with child sacrifice. The other exhibit is Lev 20:5, the syntax verse — and it is textually damaged in exactly the clause the argument needs clean, with the Septuagint’s plural common noun a live candidate for the older reading. His conclusion is Eissfeldt’s: the thesis “remains the most convincing explanation” of the biblical phrase, and the offerings, at least in Jerusalem, went to Israel’s own G-d. What he keeps from the middle of the field is the Punic link, which he says still has much to commend it, and the separation this book needs anyway: the firstborn-consecration texts of Chapter 3 are a third thing, an inner-Israelite matter, not to be merged with either.⁸

And the standard English commentary’s judgment deserves the last word before this appendix rests, because it is the sober middle of the field. Tigay’s excursus on the question lays the ambiguity out in both directions: maʿavir ba-ʾesh, “make pass through fire,” may be shorthand for sacrifice (it is how the prophets describe the Ben-Hinnom offerings, and the “passing” of firstborn in Exodus 13 ends in sacrifice or redemption), or it may name a nonlethal rite of consecration or purification, of the fire-passing kind attested elsewhere; and his verdict, on the lethal-versus-nonlethal question and the Molekh connection alike, is this: “Modern scholarship has not been able to resolve the question…”⁹

One distinction has been doing silent work through all of this, and it should be made explicit, because the appendix has been treating “Eissfeldt’s thesis” as a single object when it is two. Paul Mosca, whose 1975 Harvard dissertation is the study every later monograph on this question builds on, separates them in his opening sentence: Eissfeldt’s “two basic theses — that Punic mlk and Hebrew מלך were cognate sacrificial terms and that in ancient Israel children were sacrificed not to some foreign deity named Moloch or Melech, but to [Hashem] himself — have been repeatedly attacked in the intervening decades. The aim of the present study is to reexamine these controversial theses in the light of the new evidence that has been steadily accumulating.”¹⁰

Two theses, and this book’s fortunes are not equally tied to them. The first — that molekh names a rite rather than a god — is the one the paragraphs above leave open, answered in 1985 and 1989 and reopened in 2017, and Chapters 1 and 2 were built so that nothing turns on it. The second — that the children in the valley were being given to Israel’s own G-d — is the one Chapters 1 and 3 rest their weight on, and it is on far firmer ground. It is what Yirmiyahu’s triple denial presupposes; it is what ibn Kaspi saw in the fourteenth century and Gray saw in 1925; it is what Yeḥezkel’s statutes that were not good is trying to explain; and it is what Mosca’s dissertation set out to defend and what Dewrell’s monograph, forty-two years later, concluded again. Heider and Day stand on the other side of it — Day ends by concluding that Molekh and Israel’s G-d were never identified — and I have not counted the field. What holds this half up is the prophets, read as Chapters 1 and 3 read them. A reader who wants to shelve the whole Eissfeldt program under “interesting but superseded” should notice which half is being superseded.

The dispute over the first thesis costs this book nothing, and that is by design. If Eissfeldt was right, the Torah bans a rite the whole Levantine world practiced under a technical name; if Heider and Day are right, it bans the same rite directed to a named chthonic god; and the newest monograph in the dossier says Eissfeldt was right after all. On Eissfeldt’s reading, on Heider’s and Day’s, and on Dewrell’s alike, the practice is real, Northwest Semitic, domestic to the region, and the thing the quarantine laws quarantine. Chapters 1 and 2 were built so that nothing in them turns on the vocalization of three consonants — but the reader deserved the whole docket, and here it is.

Appendix B: Ur of the Chaldees

[Working apparatus: the geographical question deferred from Chapter 4, split out of Appendix A on 2026-08-04 and given a letter of its own, having no business under the Molekh question’s roof. One open item, disclosed in the footnotes rather than buried: Woolley’s Ur of the Chaldees (1929) is not on this book’s shelf, so the Woolley claims run only as far as Rendsburg and Weinfeld carry them. A second closed on 2026-08-08: Firestone’s talmudic cross-reference for Kutha does not sit at the daf he gives, but the sugya was run down one masekhet over, and it is now adopted with the correction shown.]

Chapter 4 left this question here, and it divides much as the Molekh question of Appendix A does: a majority view, a stubborn minority, and a book that has arranged not to need the answer. The majority view since Woolley’s excavations is the great southern Ur — Tell el-Muqayyar in lower Mesopotamia, Sumerian Ur of the ziggurat and the royal tombs — with “of the Chaldees” read as the Torah’s later-period address label, pointing the reader toward the Babylonian south (the Chaldeans are a first-millennium presence there, which is why critics take the label as anachronistic and geographers take it as disambiguation). The minority tradition is northern, and it is old, and it is a good deal stronger than its minority status suggests. The candidate is Urfa — Şanlıurfa in southeastern Turkey, ancient Edessa, some forty kilometers north of Ḥarran, where local tradition, Jewish and Christian and Muslim alike, has long pointed to Avraham’s birthplace. Gary Rendsburg, whose redaction work this book relies on in the Introduction, has assembled the case, and it turns on four things.

First, direction. Genesis has the family leave Ur for Canaan and stop at Ḥarran, and Ḥarran is not on the way: “a journey from Ur to Canaan would not pass through Ḥarran. We know the trade-routes. Anyone travelling from Ur of Sumer to Canaan would first have traveled upstream on the Euphrates” — north to the great bend, then west, nowhere near Ḥarran. From Urfa, by contrast, Ḥarran is the first stop on the road.

Second, the river. Yehoshua’s covenant speech at Shekhem, the verse the Haggadah brings and Chapter 4 opened on, puts the ancestors on the far side of it:

וַיֹּ֨אמֶר יְהוֹשֻׁ֜עַ אֶל־כׇּל־הָעָ֗ם כֹּה־אָמַ֣ר יְהֹוָה֮ אֱלֹהֵ֣י יִשְׂרָאֵל֒ בְּעֵ֣בֶר הַנָּהָ֗ר יָשְׁב֤וּ אֲבֽוֹתֵיכֶם֙ מֵֽעוֹלָ֔ם תֶּ֛רַח אֲבִ֥י אַבְרָהָ֖ם וַאֲבִ֣י נָח֑וֹר וַיַּעַבְד֖וּ אֱלֹהִ֥ים אֲחֵרִֽים׃ Va-yomer Yehoshua el-kol-ha-ʿam: koh amar Hashem Elohei Yisrael — be-ʿever ha-nahar yashvu avoteikhem me-ʿolam, Teraḥ avi Avraham va-avi Naḥor, va-yaʿavdu elohim aḥerim. “And Yehoshua said to all the people: Thus said Hashem, the G-d of Israel — beyond the River your ancestors dwelt of old, Teraḥ the father of Avraham and the father of Naḥor, and they served other gods.” — Joshua 24:2

Be-ever ha-nahar, beyond the River, and the River is the Euphrates: “Sumerian Ur cannot be the referent, as it is on the same side of the Euphrates as Canaan.” Urfa is on the other side.

Third, the family’s ethnicity, which is Chapter 17’s material. The confession the farmer makes over his basket names the father by nation:

וְעָנִ֨יתָ וְאָמַרְתָּ֜ לִפְנֵ֣י ׀ יְהֹוָ֣ה אֱלֹהֶ֗יךָ אֲרַמִּי֙ אֹבֵ֣ד אָבִ֔י וַיֵּ֣רֶד מִצְרַ֔יְמָה וַיָּ֥גׇר שָׁ֖ם בִּמְתֵ֣י מְעָ֑ט וַֽיְהִי־שָׁ֕ם לְג֥וֹי גָּד֖וֹל עָצ֥וּם וָרָֽב׃ Ve-anita ve-amarta lifnei Hashem Elohekha: arami oved avi, va-yered Mitzraymah va-yagor sham bi-metei meʿat, va-yehi sham le-goy gadol atzum va-rav. “And you shall speak up and say before Hashem your G-d: a wandering Aramean was my father, and he went down to Egypt and sojourned there few in number, and there he became a nation, great, mighty and many.” — Deuteronomy 26:5

Arami oved avi: an Aramean father. And “the Aramean heartland is located in northern Mesopotamia. In fact, the Bible refers to the region in which Urfa is found as Aram Naharaim.”

Fourth, and prettiest, the family tree is a map: Teraḥ’s father Naḥor and grandfather Serug are both names of towns in the region — Naḫur, known from the Akkadian sources, though Rendsburg grants in the same breath that its precise site in upper Mesopotamia is unknown, and Serug, well known from the later Syriac sources and surviving as modern Turkish Suruç, some forty-six kilometers southwest of Urfa.¹¹

And the epithet, which looks like the fatal objection, he turns into an argument. If the Chaldeans belong to the Babylonian south, why would a northern town be called Ur Kasdim? Because the Chaldeans were not native to the south either — they migrated there in the early to mid-first millennium, and Xenophon still knows them in the north, a warlike people on the Armenian frontier. And because the epithet is doing the work epithets do: “the great Ur required no further appellation; the expression Ur of the Chaldees implies that we are not speaking of the famous city.” One says London, and London, Ontario. His verdict: “various lines of evidence converge to support the tradition of the Jews, Christians, and Muslims of Turkey, as well as the scholarly arguments put forward by Bush and Gordon: Ur-Kasdim is north of Ḥarran and should be identified with the city of Urfa.”¹²

The epithet had been troubling readers a long time before Woolley, and the twelfth century answered it in the same key. Ibn Ezra, at the genealogy that first carries the family to the place, notices who the Chaldeans are in the Torah’s own reckoning — the descendants of Naḥor, Avraham’s brother, through his son Kesed (Gen 22:22) — and draws the inference:

ואברם נולד במקום נקרא אור כשדים וכמוהו על כן באורים כבדו ד’ (ישעי’ כד טו) והקרוב שהי’ לו שם אחר כי כשדים מבני נחור אחי אברם רק משה כתב השם הנודע בימיו והעד כי שם נולד אברם ששם מת הרן אחיו וכתוב בארץ מולדתו ויתכן שאור עשו הכשדים במקום הנזכר, ונקרא כן כמו “אשר אור לו בציון” (ישעי’ א, ט) Ve-Avram nolad be-makom nikra Ur Kasdim, u-khemohu ʿal-ken ba-urim kabedu Hashem (Yeshayahu 24:15); ve-ha-karov she-hayah lo shem aḥer, ki Kasdim mi-benei Naḥor aḥi Avram — rak Moshe katav ha-shem ha-noda be-yamav. Ve-ha-ʿed ki sham nolad Avram: she-sham met Haran aḥiv, ve-khatuv be-eretz moladto. Ve-yitakhen she-ur ʿasu ha-Kasdim ba-makom ha-nizkar, ve-nikra ken kemo “asher ur lo be-Tziyon” (Yeshayahu 1:9). “And Avram was born in a place called Ur Kasdim… And it is likely that it had another name, for the Kasdim are of the sons of Naḥor, Avram’s brother; only, Moshe wrote the name known in his own day. And the proof that Avram was born there is that his brother Haran died there, and it is written in the land of his birth.” — Ibn Ezra to Genesis 11:26

A later address written onto an older place, entered on the record by a commentator with no stake in Woolley and no access to a spade.¹³ And the observation under it is the same family-tree observation Rendsburg makes, made about eight hundred and fifty years earlier: the Chaldeans belong to Naḥor’s house, and Naḥor’s house is in the north with the rest of the relatives.

A second road reaches the same north, and it never passes through Urfa. Moshe Weinfeld, whom Chapter 8 called as a hostile witness, gets there from strata rather than from trade-routes. Of Ur Kasdim he writes flatly that “this is an anachronism, because the Chaldaeans appear in the Mesopotamian documents for the first time in the ninth century B.C.E., and they became rulers of Babylon only in the seventh and sixth centuries.” Then he goes past the label to the itinerary: “it seems that in the original version of the Patriarchal narratives the Patriarchs came from Haran; Ur of the Chaldaeans, as the place of their origin, was incorporated in a later stage.” On the old (J) layer, he holds, the call of Genesis 12:1 is delivered at Ḥarran, and Ur Kasdim is not where the family starts at all.¹⁴ Rendsburg goes north by roads and toponyms; Weinfeld goes north by layers. Neither is arguing this book’s case, and the pairing is the one Chapter 8 already ran with Weinfeld and Eissfeldt. Weinfeld also states the anachronism objection more fairly than the parenthesis at the head of this appendix does, and he supplies the sentence that belongs beside the verdict below: for the purpose he is actually pursuing, “Ur and Haran serve the same purpose: to indicate that the father of the nation comes from a known civilization.”

The ancient interpreters had the Ur/Ḥarran problem long before the source critics did, and they spent real ingenuity on it. Kugel collects their solutions under a heading that states the problem by itself: “Abraham Was in Ur—or Was It Haran?” Genesis 11:31 takes the family out of Ur and settles them at Ḥarran; two verses later the voice says go out from your land, and from your birthplace, and from your father’s house, which sounds like Ur and is being spoken at Ḥarran. Philo of Alexandria read Genesis 12:1 as a flashback and put the call back in Ur. Stephen, at his trial in Acts, took the same position on the record: the God of glory appeared to Avraham “when he was in Mesopotamia, before he lived in Haran.” Jubilees solved it from the other end, letting Avram sit fourteen years in Ḥarran and then ask whether he should “return to Ur of the Chaldeans, who are [now] earnestly requesting that I return to them” — whereupon the call arrives, and go out from your land is heard as an order not to go back.¹⁵ Chapter 4 called Jubilees as a witness to the burning house; here the same book is caught doing geography, and doing it because the itinerary would not sit flat.

One counterweight, since this book prints its counterweights. The tradition at Urfa is real, and Rendsburg is entitled to it, but the Muslim tradition is not all northern. Firestone, cataloguing the Islamic Abraham legends, locates Avraham’s birth at Kutha, “an ancient town in Iraq on a canal joining the Tigris and the Euphrates rivers.” Iraq, that is, and not Anatolia.¹⁶ So the southern identification has an old religious tradition of its own standing behind it, the Bavli’s as well as Islam’s, and the argument from the tradition of Turkey’s Jews, Christians and Muslims is one tradition among several rather than the consensus of the three faiths. The Jewish limb of that tradition is the thinnest of the three. It comes down through Rendsburg’s own source, George Bush, who had it in 1839 at second hand from a Mr. Wolff, and nothing more recent stands behind it here.

Woolley’s spade carried the question for a century in the other direction, on evidence that was about a magnificent city rather than about this family’s road.¹⁷ The verdict is that the identification is open, that the south holds the scholarly default, and that nothing in Part II leans on it: the Torah’s own itinerary runs the family through Ḥarran regardless, the furnace of the midrash burns wherever Ur was, and the argument of Chapter 4 needs only what every reading grants — that the family came out of the heart of the old world, and left.

Appendix C: The Sources

This is the exhibit list.

A book that argues from documents owes the reader the documents. What follows is the primary material the case rests on, set out in the original language with a translation of my own, so that the reader can check the argument against the evidence rather than against my account of it. Where a rendering carries weight — and in this book most of them do — you are entitled to see what it renders.

Nothing under copyright is reproduced here. Every text below is either ancient or medieval and long out of copyright, or is quoted from an edition old enough to be in the public domain; where a modern scholar’s reading governs a word, the scholar is credited and the words remain the author’s.

A. The Tanakh witnesses

Every verse quoted in this book is given in the Masoretic text, with full vowels and cantillation, from the witness known as Miqra according to the Masorah, cross-checked against Mechon-Mamre and the Biblia Hebraica Stuttgartensia. Where the printed English Bibles number a verse differently — Exodus 22:28 here is 22:29 in many of them — the Hebrew numbering governs, and the notes say so at the point it matters.

B. Rabbinic texts — where to find them

Every rabbinic passage this book quotes is on Sefaria, in the editions the notes name, free, in Hebrew or Aramaic and in English, a search box away. So they are not reprinted here. Reprinting them would have added forty pages to a book that can instead point, and would have set several modern editions’ own vocalization in type along the way. The chapter arguing from a passage quotes what it needs at the point of argument and names the locus in its note; that note is the address, and the text is waiting at it.

C. The Mesha stele (KAI 181), lines 4–19

Mesha king of Moab cut this monument at Dibon in the ninth century BCE, within years of the campaign 2 Kings 3 reports from the other side. The stone is in the Musée du Louvre, Département des Antiquités orientales, AO 5066 (= MNB 752 = KAI 181): basalt, 125 × 69 × 37 cm, from Dhiban, acquired 1873, Sully salle 303. It is quoted in Chapter 2, where it stands beside 2 Kings 3:27, and again in Chapter 14, where its ḥerem clause is the book’s best epigraphic witness that the ban was a regional vocabulary and not an Israelite peculiarity.

The Moabite below is set as G. A. Cooke prints it in A Text-Book of North-Semitic Inscriptions (Oxford: Clarendon Press, 1903), no. 1, p. 1 — a public-domain edition, in square Hebrew characters, of the lines the argument uses. Cooke’s sigla travel with his text: · is the word-divider cut on the stone; | is the stone’s own vertical stroke; a bare . stands for one illegible letter; [] enclose restoration. Cooke also marks a doubtful reading with a dot above the letter, which the book’s Hebrew face cannot carry; rather than normalise the doubt away silently, the dotted letters are itemised here: ימה and וישׁ (l. 8), ואבן (l. 9), עטרת and the final י (l. 10), העם (l. 11), וא[ס] (l. 12), מחרת and the final וא (l. 14), ואח (l. 15), the letters of ג[ב]רן (l. 16), את (l. 18). The Moabite is printed without line numbers, so that it matches Cooke’s text letter for letter; the transliteration beneath carries Cooke’s marginal numbers 4–19, one line to one line of the stone.

שע · כי · השעני · מכל · ה . לכן · וכי · הראני · בכל · שנאי | עמר י · מלך · ישראל · ויענו · את · מאב · ימן · רבן · כי · יאנף · כמש · באר צה | ויחלפה · בנה · ויאמר · גם · הא · אענו · את · מאב | בימי · אמר · . . . . וארא · בה · ובבתה | וישראל · אבד · אבד · עלם · וירש · עמרי · את · [אר] ץ · מהדבא | וישב · בה · ימה · וחצי · ימי · בנה · ארבען · שת · ויש בה · כמש · בימי | ואבן · את · בעלמען · ואעש · בה · האשוח · ואב[ן] את · קריתן | ואש · גד · ישב · בארץ · עטרת · מעלם · ויבן · לה · מלך · י שראל · את · עטרת | ואלתחם · בקר · ואחזה | ואהרג · את · כל · העם · . הקר · רית · לכמש · ולמאב | ואשב · משם · את · אראל · דודה · וא[ס] חבה · לפני · כמש · בקרית | ואשב · בה · את · אש · שרן · ואת · א[ש] מחרת | ויאמר · לי · כמש · לך · אחז · את · נבה · על · ישראל | וא הלך · בללה · ואלתחם · בה · מבקע · השחרת · עד · הצהרם | ואח זה · ואהרג · כל . · שבעת · אלפ . ג[ב]רן · ו . . ן | וגברת · ו . . ת · ורחמת | כי · לעשתר · כמש · החרמתה | ואקח · משם · א . . לי · יהוה · ואסחב · הם · לפני · כמש | ומלך · ישראל · בנה · את יהץ · וישב · בה · בהלתחמה · בי | ויגרשה · כמש · מפני [ו]

Transliteration. The scheme is the book’s own, as it is printed in the chapters (maqom, mlk ʾmr, bi nishbaʿti); it departs at two points from the chapters’ reader-first practice, using q for ק where the chapters use k, and the half-ring ʾ for א where the chapters use a straight apostrophe and confine it to word-internal position. Two further departures are deliberate, and both are because this is a transcription and not a reading: alef is marked in every position, word-initial included, and ש is always sh, since Moabite orthography does not distinguish śin and no etymological judgment should be smuggled in by a transliteration. Cooke’s sigla are carried through unchanged, so the two blocks can be checked letter against letter. No vowels are supplied anywhere.

4 shʿ · ky · hshʿny · mkl · h. lkn · wky · hrʾny · bkl · shnʾy | ʿmr 5 y · mlk · yshrʾl · wyʿnw · ʾt · mʾb · ymn · rbn · ky · yʾnp · kmsh · bʾr 6 tzh | wyḥlph · bnh · wyʾmr · gm · hʾ · ʾʿnw · ʾt · mʾb | bymy · ʾmr ·. 7 wʾrʾ · bh · wbbth | wyshrʾl · ʾbd · ʾbd · ʿlm · wyrsh · ʿmry · ʾt · [ʾr] 8 tz · mhdbʾ | wyshb · bh · ymh · wḥtzy · ymy · bnh · ʾrbʿn · sht · wysh 9 bh · kmsh · bymy | wʾbn · ʾt · bʿlmʿn · wʾʿsh · bh · hʾshwḥ · wʾb[n] 10 ʾt · qrytn | wʾsh · gd · yshb · bʾrtz · ʿṭrt · mʿlm · wybn · lh · mlk · y 11 shrʾl · ʾt · ʿṭrt | wʾltḥm · bqr · wʾḥzh | wʾhrg · ʾt · kl · hʿm ·. 12 hqr · ryt · lkmsh · wlmʾb | wʾshb · mshm · ʾt · ʾrʾl · dwdh · wʾ[s] 13 ḥbh · lpny · kmsh · bqryt | wʾshb · bh · ʾt · ʾsh · shrn · wʾt · ʾ[sh] 14 mḥrt | wyʾmr · ly · kmsh · lk · ʾḥz · ʾt · nbh · ʿl · yshrʾl | wʾ 15 hlk · bllh · wʾltḥm · bh · mbqʿ · hshḥrt · ʿd · htzhrm | wʾḥ 16 zh · wʾhrg · kl. · shbʿt · ʾlp. g[b]rn · w. n | wgbrt · w. 17 t · wrḥmt | ky · lʿshtr · kmsh · hḥrmth | wʾqḥ · mshm · ʾ. 18 ly · [the Name] · wʾsḥb · hm · lpny · kmsh | wmlk · yshrʾl · bnh · ʾt 19 yhtz · wyshb · bh · bhltḥmh · by | wygrshh · kmsh · mpny [w]

At line 18 the stone spells the Tetragrammaton — a Moabite king naming Israel’s G-d on his own victory monument. The Hebrew line above prints it exactly as Cooke prints it. This book bars any Latin-character spelling of the Name, and the bar is not qualified for non-Israelite sources; the transliteration therefore stops at the word and marks the slot [the Name]. Square brackets do double duty in this section and the difference is worth stating: elsewhere they enclose letters inside a word and mark a restoration — ג[ב]רן, [gr]t — whereas here they enclose an English phrase standing in for a whole word, which is an editorial substitution and not a claim about what survives on the stone. The book’s reverential substitute is deliberately not used in that slot: Mesha did not say it, and putting it in his mouth would be a small falsification of the evidence.

Translation (the author’s own, made from the Moabite above; it is not adapted from any published rendering, and neither of the two English versions entered it — Cooke’s own of 1903, which is public domain, and K. A. D. Smelik’s in The Context of Scripture, which is not). Stone-line numbers stand in ⟨ ⟩; [] marks English resting on a restored letter, as in the Moabite; ⟨…⟩ marks a gap the edition leaves empty, where this appendix declines to guess; and a word left in transliteration is a word this appendix declines to translate, for the reasons given below.

⟨4⟩ …]shaʿ — for it was he who delivered me from all the ⟨…⟩, and because he let me gaze on all who hate me. As for Omri ⟨5⟩ the king of Israel: he bore down on Moab for days on end, because Kemosh had grown angry at his ⟨6⟩ country. His son came after him, and he in his turn said: Moab I will bear down on. In my own time he spoke so ⟨…⟩ ⟨7⟩ yet I gazed on him, and on his household, and Israel is destroyed — destroyed for all time.

Now Omri had seized the la[nd] ⟨8⟩ of Medeba and occupied it through his own lifetime and half of his son’s, forty years in all; and in my ⟨9⟩ time Kemosh returned it. I rebuilt Baʿal-Meʿon, putting the reservoir in it; and I rebui[lt] ⟨10⟩ Qiryathen.

The Gadites had settled in Ataroth’s territory long before, and Israel’s king ⟨11⟩ fortified Ataroth to his own use. I gave battle to that town, seized it, and put to death its whole populace — ⟨12⟩ a ryt for Kemosh, for Moab. From there I carried away the ʾarʾel belonging to its dwd, dra[g]ging ⟨13⟩ it into Kemosh’s presence at Qeriyoth; and I resettled it with the people of Sharon and those of ⟨14⟩ Maḥarath.

Then Kemosh spoke to me: go, seize Nevo out of Israel’s hand. So I ⟨15⟩ marched through the night, engaged the place from first light through to midday, seized ⟨16⟩ it, and put every soul in it to death: seven thousand — m[e]n with ⟨…⟩, women with ⟨…⟩ ⟨17⟩ and girls; for to Ashtar-Kemosh I devoted it. From there I carried away the ⟨…⟩ ⟨18⟩ belonging to [the Name], dragging them into Kemosh’s presence.

Israel’s king had fortified ⟨19⟩ Yahatz, holding it as his base while he warred on me; and Kemosh dislodged him ahead of me. [And] …

The clause the book leans on hardest, line 17, standing alone as Cooke prints it:

כי · לעשתר · כמש · החרמתה ky · lʿshtr · kmsh · hḥrmth “for to Ashtar-Kemosh I devoted it.”

What the stone will bear. hḥrmth is a Hiphil, first person singular, of ḥ-r-m, with a feminine singular object suffix — vocalised in Chapter 14 as heḥeramtiha, “I devoted it,” the town being feminine. That is the same root, the same stem and the same construction as the biblical war-ḥerem, on a neighbouring king’s monument; Chapter 14’s claim that the ban was a regional koine rests on an identical verb in an identical grammatical form, not on a consonantal root spotted at a distance. The devotion is expressed with ל־, to a divine recipient, which is exactly what the ban-as-sacrifice reading hears. But the same stone supplies its own counterweight five lines earlier: at Ataroth the thing done is a ryt לכמש ולמאב — for Kemosh and for Moab. The datives are two, one a god and one a nation, and one does not sacrifice to a people. Whatever ryt names, its beneficiaries are split between the deity and the body politic.

ryt (l. 12) is a hapax and is left untranslated here on purpose. Nothing in the consonants requires “sacrifice.” Cooke derives it from r-ʾ-y, to see, and makes it a thing looked at (his note, p. 11); Smelik prints “as a sacrifice(?)” with his own question mark, which Chapter 14 quotes and footnotes; others have gone to r-w-y, to drink one’s fill. An edition that translates a hapax silently is making an argument in the guise of a rendering. ʾarʾel and dwd (l. 12) are left standing for the same reason: ʾarʾel is related to the biblical ariel, the altar-hearth (Ezek 43:15; Isa 29:1–2), and dwd is either a divine title or “its beloved.”

The word-divider inside לעשתר · כמש is cut on the stone and is printed by Cooke, by Lidzbarski and by Smend and Socin alike. “Ashtar-Kemosh” as one compound deity is therefore an editorial reading, exactly as Chapter 14’s footnote already says; it is kept here because it is the form the book has already printed and already flagged.

rḥmt (l. 17) is literally wombs, and the war-spoil idiom is Judges 5:30, רַחַם רַחֲמָתַיִם לְרֹאשׁ גֶּבֶר, a womb or two per man, where it plainly means a girl taken as plunder. It is rendered “girls.” Under no reading is it infants.

The tempo is evidence too. The ḥerem at Nevo occupies a single morning — from the break of dawn to noon, after a night march, at the end of an assault. The stone gives it no calendar, no recurrence, and no tariff. That is Chapter 14’s distinction between the Tophet and the ban, confirmed by the stone’s own timekeeping.

And a plain statement the reader is owed: there is no child offering anywhere in these sixteen lines. No wall, no son, no burning, no ketzef. The stele corroborates Mesha’s theology — Kemosh angry, Kemosh saving, Kemosh driving the enemy out — and says nothing whatever about the act on the wall that Chapter 2 sets beside it. Chapter 2 says so; it is repeated here, where the text is printed, because a reader looking at sixteen lines of Moabite will want to know why the thing the chapter is about is absent from them.

The apparatus, where the editions disagree. Cooke states at p. 4 n. 1 that his text follows Lidzbarski’s Handbuch (1898), 415 f., corrected from Lidzbarski’s Ephemeris i (1900), 1–10, which has not been collated. The three divergences that matter:

Line 4. Cooke prints ה. לכן — h, one illegible letter, lkn — and argues in his note (p. 8) for השלכן, “assailants,” adding that hmlkn is the less probable reading. Lidzbarski, Handbuch, 415, and Smend–Socin, 12, both print המלכן outright, “the kings.” The translation above prints the gap rather than the guess. Any rendering of the line as “from all the kings” — which is the sense Chapter 2’s summary implies — rests on Lidzbarski and Smend–Socin and not on the edition this appendix prints, and must say so. Cooke’s own note at p. 8, it should be said, places Lidzbarski among those who prefer השלכן.

Line 16. What stands on Cooke’s page is כל. · שבעת · אלפ. ג[ב]רן · ו. ן | וגברת · ו., running over into ת · ורחמת. That is: all of it, with the ה of כלה restored only in Cooke’s note; seven thousand, with the ן of אלפן a gap; then ג[ב]רן, men, itself half-restored; then a category that is a blank ending in ־ן; then וגברת, women, whole; then a category that is a blank ending in ־ת; then ורחמת, whole. Lidzbarski leaves the same two categories blank. The “resident aliens” (גרן/גרת) that the modern reading puts there are Cooke’s proposed restoration in his note (p. 12), where he himself records the alternative that גר means a child; and Smend–Socin, p. 12, print those gaps filled, without brackets, as מגברן · ומבנן | וגברת · ובנ, the word running over to ת at the head of l. 17 — of men and of boys, and women and daughters. Where Cooke prints כל. and argues כלה, Lidzbarski prints כלה and Smend–Socin print כלם; where Cooke prints אלפ. and argues אלפ[ן], Lidzbarski prints אלפן and Smend–Socin print אלף. Chapter 14’s sentence “There are no infants on Mesha’s list” survives literally — no editor restores sucklings, and sons and daughters are not infants — but it is standing on restored letters, and one respectable 1886 restoration of those very gaps puts children in the massacre. Chapter 14’s footnote itemises gbrn and grn, gbrt and [gr]t as though the sojourners stood on the stone. They do not.

Lines 17–18. Cooke prints א. and argues in his note (p. 12) for א[ת כ]לי, the vessels, rejecting א[רא]לי because that noun stands in the singular at l. 12 and takes את before it. Lidzbarski prints אר[א]; Smend–Socin print ארא with their doubtful-letter mark over the second alef, giving the altar-hearths. The English above takes neither. The following word is secure, and it is the fact worth having: a Moabite king carries off the cult furniture of Israel’s own G-d, naming Him, and drags it before Kemosh.

Rights. The inscription is ninth-century BCE and carries no copyright. The printed text is Cooke 1903, no. 1, p. 1, read from the clean Cornell exemplar (archive.org cu31924096083104; the other exemplar, the Wellcome copy, carries a previous owner’s pencil vocalisation over the Moabite, including a penciled ש filling the lacuna in l. 4, none of which is print). Cooke was printed pre-1929 and its author died in 1939, so it is public domain in the United States and in the UK and EU alike; text and notes may both be reproduced. The collating editions — Lidzbarski, Handbuch der nordsemitischen Epigraphik (Weimar, 1898), 415, and Smend and Socin, Die Inschrift des Königs Mesa von Moab (Freiburg i. B., 1886), 12 — are public domain on both grounds. KAI 181 is cited as the standard reference by number only; none of Donner and Röllig’s German, apparatus or restorations is reproduced. The Louvre record is a factual museum record cited by inventory number, with no museum photograph reproduced. Smelik’s translation in Hallo and Younger, The Context of Scripture 2.23, is in copyright; nothing of it is reproduced in the English printed here, and the one sentence Chapter 14 takes from it stays quoted and footnoted. Cooke’s own 1903 English is public domain and could lawfully be reprinted; it is not, because the front matter claims these translations as the author’s: the rendering above was made independently of Cooke’s English, and its independence was checked by hand, line against line.

D. Philo of Byblos, apud Eusebius, Praeparatio Evangelica 1.10.44, with the 4.16.11 parallel

Eusebius of Caesarea quotes this one paragraph of Philo of Byblos twice, and sources it twice: at 1.10.43–44 as from Philo’s treatise On the Jews, at 4.16.11 as from Book I of the Phoenician History. Both cannot be right. Both frames are printed below, and not only the fragment, because printing the paragraph alone would conceal what the doubling proves: our witness was not always sure what he was quoting. The passage is good evidence for the protocol and thin evidence for its provenance. Attridge and Oden argue the attribution problem at p. 93 n. 148; the point is theirs and is cited, not quoted.

The Greek is set from E. H. Gifford’s edition (Oxford: E Typographeo Academico, 1903), tomus I — public domain, with its apparatus criticus and its marginal Stephanus pagination — at pp. 54–55 (= Steph. 40 c 1 – d 4) and p. 203 (= Steph. 156 d 3 – 157 a 2). A second public-domain witness, Migne’s Patrologia Graeca 21 (Paris, 1857), reprinting Viger’s text of 1628, is collated at cols. 85 and 273.

PE 1.10.44 — Gifford I:54–55. Eusebius’ frame first:

Ὁ δ᾽ αὐτὸς ἐν τῷ περὶ Ἰουδαίων συγγράμματι ἔτι καὶ ταῦτα περὶ τοῦ Κρόνου γράφει· […] Καὶ μετὰ βραχέα φησίν· Ho d’ autos en tō peri Ioudaiōn syngrammati eti kai tauta peri tou Kronou graphei; … Kai meta brachea phēsin; “The same author, in his treatise On the Jews, writes this also about Kronos: […] And a little further on he says:”

Then the fragment, which Gifford sets in smaller type with the marginal source-name ΦΙΛΩΝΟΣ:

Ἔθος ἦν τοῖς παλαιοῖς ἐν ταῖς μεγάλαις συμφοραῖς τῶν κινδύνων ἀντὶ τῆς πάντων φθορᾶς τὸ ἠγαπημένον τῶν τέκνων τοὺς κρατοῦντας ἢ πόλεως ἢ ἔθνους εἰς σφαγὴν ἐπιδιδόναι, λύτρον τοῖς τιμωροῖς δαίμοσι· κατεσφάττοντο δὲ οἱ διδόμενοι μυστικῶς. Κρόνος τοίνυν, ὃν οἱ Φοίνικες Ἦλον προσαγορεύουσι, βασιλεύων τῆς χώρας, καὶ ὕστερον μετὰ τὴν τοῦ βίου τελευτὴν εἰς τὸν τοῦ Κρόνου ἀστέρα καθιερωθείς, ἐξ ἐπιχωρίας Νύμφης Ἀνωβρὲτ λεγομένης υἱὸν ἔχων μονογενῆ, (ὃν διὰ τοῦτο Ἰεδοὺδ ἐκάλουν, τοῦ μονογενοῦς οὕτως ἔτι καὶ νῦν καλουμένου παρὰ τοῖς Φοίνιξι) κινδύνων ἐκ πολέμου μεγίστων κατειληφότων τὴν χώραν, βασιλικῷ κοσμήσας σχήματι τὸν υἱὸν βωμόν τε κατασκευασάμενος κατέθυσεν. Ethos ēn tois palaiois en tais megalais symphorais tōn kindynōn anti tēs pantōn phthoras to ēgapēmenon tōn teknōn tous kratountas ē poleōs ē ethnous eis sphagēn epididonai, lytron tois timōrois daimosi; katesphattonto de hoi didomenoi mystikōs. Kronos toinyn, hon hoi Phoinikes Ēlon prosagoreuousi, basileuōn tēs chōras, kai hysteron meta tēn tou biou teleutēn eis ton tou Kronou astera kathierōtheis, ex epichōrias Nymphēs Anōbret legomenēs huion echōn monogenē, (hon dia touto Iedoud ekaloun, tou monogenous houtōs eti kai nyn kaloumenou para tois Phoinixi) kindynōn ek polemou megistōn kateilēphotōn tēn chōran, basilikō kosmēsas schēmati ton huion bōmon te kataskeuasamenos katethysen.

“The ancients had a custom: in the great calamities of peril, in place of the ruin of all, those who held power over either a city or a nation would hand over the beloved one among their children for slaughter — ransom to the avenging spirits; and the ones so given were slaughtered by a rite kept secret. Kronos, then — whom the Phoenicians address as Ēlos — was reigning over that country, and afterwards, when his life had ended, was consecrated into the star of Kronos. There was a nymph of that country called Anōbret, and by her he had an only son (whom for that reason they called Iedoud; among the Phoenicians an only son still bears that name today). And when the greatest dangers of war had taken hold of the land, he decked his son out in royal array, built himself an altar, and sacrificed him.”

PE 4.16.11 — Gifford I:203. The frame, which sources the same words to a different book:

Ταῦτα μὲν οὖν ἐκ τῆς προδηλωθείσης κείσθω γραφῆς. Ἐκ δὲ τοῦ πρώτου συγγράμματος τῆς Φίλωνος Φοινικικῆς ἱστορίας παραθήσομαι ταῦτα· Tauta men oun ek tēs prodēlōtheisēs keisthō graphēs. Ek de tou prōtou syngrammatos tēs Philōnos Phoinikikēs historias parathēsomai tauta; “So much, then, may stand from the writing set out above. But from the first book of Philo’s Phoenician History I shall set down this:”

Ἔθος ἦν τοῖς παλαιοῖς ἐν ταῖς μεγάλαις συμφοραῖς τῶν κινδύνων ἀντὶ τῆς πάντων φθορᾶς τὸ ἠγαπημένον τῶν τέκνων τοὺς κρατοῦντας ἢ πόλεως ἢ ἔθνους εἰς σφαγὴν ἐπιδιδόναι λύτρον τοῖς τιμωροῖς δαίμοσι· κατεσφάττοντο δὲ οἱ διδόμενοι μυστικῶς. Κρόνος τοίνυν, ὃν οἱ Φοίνικες Ἦλ προσαγορεύουσι, βασιλεύων τῆς χώρας, καὶ ὕστερον μετὰ τὴν τοῦ βίου τελευτὴν εἰς τὸν τοῦ Κρόνου ἀστέρα καθιερωθείς, ἐξ ἐπιχωρίας νύμφης Ἀνωβρὲτ λεγομένης υἱὸν ἔχων μονογενῆ (ὃν διὰ τοῦτο Ἰεοὺδ ἐκάλουν, τοῦ μονογενοῦς οὕτως ἔτι καὶ νῦν καλουμένου παρὰ τοῖς Φοίνιξι), κινδύνων ἐκ πολέμου μεγίστων κατειληφότων τὴν χώραν, βασιλικῷ κοσμήσας σχήματι τὸν υἱὸν βωμόν τε κατασκευασάμενος κατέθυσε. Ethos ēn tois palaiois en tais megalais symphorais tōn kindynōn anti tēs pantōn phthoras to ēgapēmenon tōn teknōn tous kratountas ē poleōs ē ethnous eis sphagēn epididonai lytron tois timōrois daimosi; katesphattonto de hoi didomenoi mystikōs. Kronos toinyn, hon hoi Phoinikes Ēl prosagoreuousi, basileuōn tēs chōras, kai hysteron meta tēn tou biou teleutēn eis ton tou Kronou astera kathierōtheis, ex epichōrias nymphēs Anōbret legomenēs huion echōn monogenē (hon dia touto Ieoud ekaloun, tou monogenous houtōs eti kai nyn kaloumenou para tois Phoinixi), kindynōn ek polemou megistōn kateilēphotōn tēn chōran, basilikō kosmēsas schēmati ton huion bōmon te kataskeuasamenos katethyse.

The English is word for word the same, with two exceptions, and they are the whole difference between Gifford’s two printings: for “Ēlos” read “Ēl”, and for “Iedoud” read “Ieoud”. The paragraph is not set twice here, because Eusebius did not quote two passages; he quoted one passage twice and attributed it to two different works.

Migne’s Greek, and what a reader will find if he checks. Migne (PG 21, col. 85, and again at col. 273 for the 4.16.11 parallel) sets the fragment inside guillemets, with heavier commas, no parentheses and no ἄνω τελεία — and, in both passages, reads Ἰσραὴλ where Gifford reads Ἦλον and Ἦλ:

« … Κρόνος τοίνυν, ὃν οἱ Φοίνικες Ἰσραὴλ προσαγορεύουσι, βασιλεύων τῆς χώρας … ἐξ ἐπιχωρίας νύμφης, Ἀνωβρὲτ λεγομένης, υἱὸν ἔχων μονογενῆ, ὃν διὰ τοῦτο Ἰεοὺδ ἐκάλουν … » … Kronos toinyn, hon hoi Phoinikes Israēl prosagoreuousi, basileuōn tēs chōras … ex epichōrias nymphēs, Anōbret legomenēs, huion echōn monogenē, hon dia touto Ieoud ekaloun …

So the Viger vulgate carries “Israel” throughout, and it is Gifford’s Ἦλον/Ἦλ that is the departure. The reading is not Migne’s invention: Gifford’s apparatus assigns Ἰσραὴλ to codex I at 40 c 5, and at 156 d 7 records the same manuscript’s reading in the same clause as ΙΗΛ under a supralinear bar — the standard nomen sacrum contraction of Ἰσραήλ. That a scribe assimilated an unfamiliar ΗΛ to a sacred contraction he wrote every day is the obvious account of it, but it is an inference from the page and Gifford nowhere says it. What is not inference, and what the reader is owed flatly, is this: check this book against Migne and you will find Israel where we print Ēl. Attridge and Oden, who collate independently, assign the Israel reading to four witnesses where Gifford assigns it to one.

The son’s name. This matters because Chapter 2 hears yaḥid under it, and the text is not as settled as a single printed form suggests. Gifford prints Ἰεδοὺδ at 1.10.44 (I:54 = Steph. 40 d 1) and Ἰεοὺδ at 4.16.11 (I:203 = Steph. 156 d 10). His apparatus at 40 d 1 records Ἰεδα͞υδ (ΑΗ), Ἱροὺδ (ΒΟ), Ἰεοὺδ (I); at 156 d 10 it records Ἰουεοὺτ (ΑΗ). Migne prints Ἰεοὺδ in both places. Attridge and Oden print Ἰεοὺδ at 1.10.44 — but as an emendation: their apparatus credits it to Eissfeldt, Kleine Schriften 3 (1966), 409–10, and gives the manuscript readings as Ἰεδοὺδ (A), Ἱεροὺδ (N), Ἱροὺδ (BODV). The smooth form that makes yaḥid audible at 1.10.44 is therefore codex I’s there, and the modern standard edition’s conjecture — proposed, as it happens, by the same scholar whose molk thesis Appendix A weighs. Chapter 2 prints Ἰεούδ, which is the correct citation form of the oxytone Gifford prints at 4.16.11 and Migne prints in both places (an oxytone takes the grave only before another word in its phrase, and the acute when cited alone); it is not what Gifford prints at I:54, the page Chapter 2 cites. The force of the supralinear bar over the -αυ- of Ἰεδα͞υδ at 40 d 1 is not explained by anything Gifford says, and is recorded without a guess; where Gifford read Ἰεδα͞υδ in his ΑΗ, Attridge and Oden read Ἰεδοὺδ in their A.

The instability of the name is itself the argument, and the argument is not this book’s. Løkkegaard proposed reading the variants as a vacillation between Hebrew yḥyd, “only,” and ydyd, “beloved”; Attridge and Oden restate and credit him at p. 94 n. 150; Levenson develops it at pp. 30–31, where Ἰεοὺδ answers to yaḥid and Ἰεδοὺδ to yadid, μονογενής more probably renders yaḥid, and the Septuagint itself renders yaḥid with love-words — at Gen 22:2, τὸν υἱόν σου τὸν ἀγαπητὸν ὃν ἠγάπησας. Philo’s ἠγαπημένον and the Septuagint’s ἠγάπησας are the same verb. Two collations are in play here and must never be merged: Gifford’s sigla are A, H, B, O, I; Attridge and Oden’s are A, B, D, G, N, O, V; three of the letters occur in both systems and denote different manuscripts.

Where the argument lives, and why the standard English hides it. λύτρον is a bare accusative in apposition to τὸ ἠγαπημένον τῶν τέκνων, itself the object of ἐπιδιδόναι. No preposition governs it. The recipients stand in a plain dative, τοῖς τιμωροῖς δαίμοσι, again with no preposition. The only εἰς in the sentence is εἰς σφαγήν, and it governs the manner of death, not the exchange. That is identical in all three printings. The sentence’s one preposition of substitution is ἀντί, and its object is τῆς πάντων φθορᾶς, the ruin of all. Against that, LXX Gen 22:13: καὶ ἀνήνεγκεν αὐτὸν εἰς ὁλοκάρπωσιν ἀντὶ Ἰσαὰκ τοῦ υἱοῦ αὐτοῦ (Swete, I:35).

Both sentences turn on ἀντί, and the objects are opposite — the ruin of all on the wall, Yitzḥak on the mountain. The two uses of the preposition are not identical: Philo’s ἀντί names the outcome bought off, the Septuagint’s names the victim replaced. What the pairing shows is one vocabulary of exchange, running in opposite directions.

And Gifford’s own English destroys the preposition. He renders ἀντὶ τῆς πάντων φθορᾶς as a purpose clause — “in order to avert the common ruin” in Book 1, “in order to avert the general destruction” in Book 4 — silently converting a preposition of exchange into a statement of intention. Every English rendering of this passage in general circulation descends from that one. This book translates afresh because the standard English does not say what the Greek says. Copyright is not the reason.

Smaller choices, stated. σφαγήν and κατεσφάττοντο are kept on one English root (“for slaughter” … “were slaughtered”), against Gifford’s “for sacrifice” … “slain,” which breaks a link Philo built: the children are handed over for the cutting, and then they are cut. λύτρον is set off with a dash rather than smoothed to “as a ransom,” because the Greek supplies no “as” and the word stands in bare apposition. δαίμοσι is rendered “spirits”; Chapter 2 has “avenging powers” for the same dative, and the difference is deliberate: the chapter characterizes what the δαίμονες are in the transaction, this appendix says only what the word says. βασιλικῷ σχήματι is “royal array”: σχῆμα is bearing or guise, not a garment, and the “royal robes” of the Introduction and Chapter 2 specify clothing the Greek does not name. (The substantive reading of the detail is Grottanelli’s, reported by Levenson at p. 27 — the king puts his own rank on the child, so that in ransoming the city he ransoms himself as well.) μονογενής is “an only son” in both places, so that the reader can see it is one word twice; “only-begotten” would import the Nicene register into a Phoenician myth and prejudge Chapter 22’s argument. εἰς τὸν τοῦ Κρόνου ἀστέρα is “into the star of Kronos”: the identification with Saturn is right and belongs in a note, but the Greek does not make it and καθιερωθείς is one verb. ἐπιδιδόναι and διδόμενοι keep Philo’s move from compound to simplex. Ἀνωβρὲτ is a nymph “of that country,” a local one; her name is unexplained, and the proposals to make her an imported goddess are surveyed and rejected by Attridge and Oden at p. 93 n. 149 (Renan’s ʿyn, “spring”; the ʿAnat rabbat reading Clemen reports).

Both Ἦλον and Ἦλ carry a smooth breathing, so the transliteration is Ēlos and Ēl with no initial h: the Greek is transcribing Semitic ʾĒl, with alef and not he. At 1.10.44 the mark on p. 54 is the psili-plus-perispomeni of ἦν three lines above and is plainly not the dasia-plus-perispomeni of Ὧι on the same page. At 4.16.11 the mark cannot be resolved at the scan’s resolution, and the reading rests instead on Attridge and Oden, who print both Ἦλ and Ἦλον with smooth breathing while printing Ἱροὺδ and Ἱεροὺδ with rough. That caveat is general: the Gifford scan’s page image is a 400 ppi one-bit JBIG2 mask and the Migne scan’s is 600 ppi one-bit; anything “read” above those figures is interpolation, and JBIG2 is a pattern-matching codec that can substitute similar glyph bitmaps. Where a diacritic matters, it is corroborated from a second printing rather than from magnification.

Two further collation notes. Gifford’s cross-reference at I:54 gives the parallel as 156 d 2 – 157 a 2, while his own marginal lettering and source-line at I:203 begin the fragment at 156 d 3; this appendix pins d 3, against the printed marginalia. And Attridge and Oden’s critical text reads ἐπὶ τὸν τοῦ Κρόνου ἀστέρα where Gifford prints εἰς τὸν — which is the AH reading Gifford himself records at 40 c 6.

Rights. Gifford (Oxford, 1903) is public domain in the United States as a pre-1929 publication, and Gifford died in 1905, so the edition is out of copyright in the UK and EU as well; its Greek, its apparatus criticus and its marginal Stephanus pagination may all be reproduced, and are. Migne PG 21 (1857) is public domain everywhere by age; its Greek is Viger’s of 1628. Swete’s Old Testament in Greek, vol. I (Cambridge, 1901), is public domain. Mras’s Eusebius (GCS, 1954) is in copyright and is not the source of anything printed here. Attridge and Oden, Philo of Byblos (1981), is in copyright: it is cited above at pp. 62, 93 nn. 148–149 and 94 n. 150 for readings, sigla and scholarship, and nothing of its English or its prose is reproduced; its apparatus as printed is the editors’ own editorial work, and the readings above are reported in this book’s formatting and not in theirs. Levenson’s Death and Resurrection of the Beloved Son (1993) — cited above at pp. 26–27 and 30–31 — is in copyright, and is cited and not quoted — with one hazard worth stating, since it governs any future English of this passage: Levenson reproduces Gifford’s English in full at his p. 26, so Gifford’s translation must be reckoned with even where only Levenson is on the desk. The English above was composed from the Greek and then read against both Gifford’s translation and Levenson’s page; the longest shared run is five words, in clauses where any faithful English converges.

E. Diodorus Siculus 20.14.1–7, with Kleitarchos and the Suda

The Agathocles crisis of 310 BCE, the substitution scandal, and the bronze image: the passage Chapter 2 leans on for the proposition that Carthage treated substitution as fraud. The Greek is the Teubner text of C. Th. Fischer (Leipzig, 1906), vol. 5, pp. 192–93, with his apparatus criticus at the foot of both pages — Vogel edited the earlier volumes; Book XX is Fischer’s. Eusebius copies the whole of it, and his text is collated below from Gifford (Oxford, 1903), p. 210 (= Steph. 161 a 3 – b 7). Fischer’s marginal section numbers are given as ⟨1⟩–⟨7⟩; his face prints ϑ and ϰ, given here in ordinary letterforms; ⟨ ⟩ inside the Greek are his own supplements.

Transliteration scheme, built from the Greek this book already prints (hypobolimaios, lytron, monogenēs, Kleitarchos, Kronos): rough breathing = h, smooth unmarked; η = ē and ω = ō, other vowels plain; υ = y standing alone and u inside a diphthong; χ = ch, θ = th, φ = ph, ψ = ps, ξ = x; γ before a velar = n; initial or doubled ρ = rh/rrh; accents unmarked; elisions kept. Iota subscript is written on the line as an adscript i (λάθρᾳ = lathrai, δημοσίᾳ = dēmosiai), so that the case endings survive and the Greek can be checked from the Latin letters.

⟨1⟩ διόπερ οἱ Καρχηδόνιοι, νομίσαντες ἐκ θεῶν αὐτοῖς γεγονέναι τὴν συμφοράν, ἐτράπησαν πρὸς παντοίαν ἱκεσίαν τοῦ δαιμονίου καὶ νομίσαντες μάλιστα μηνίειν αὐτοῖς τὸν Ἡρακλέα τὸν παρὰ τοῖς ἀποίκοις, χρημάτων πλῆθος καὶ τῶν πολυτελεστάτων ἀναθημάτων ἔπεμψαν εἰς τὴν Τύρον οὐκ ὀλίγα. dioper hoi Karchēdonioi, nomisantes ek theōn autois gegonenai tēn symphoran, etrapēsan pros pantoian hikesian tou daimoniou kai nomisantes malista mēniein autois ton Hēraklea ton para tois apoikois, chrēmatōn plēthos kai tōn polytelestatōn anathēmatōn epempsan eis tēn Tyron ouk oliga. ⟨2⟩ ἀποικισθέντες γὰρ ἐκ ταύτης εἰώθεισαν ἐν τοῖς ἔμπροσθεν χρόνοις δεκάτην ἀποστέλλειν τῷ θεῷ πάντων τῶν εἰς πρόσοδον πιπτόντων· ὕστερον δὲ μεγάλους κτησάμενοι πλούτους καὶ προσόδους ἀξιολογωτέρας λαμβάνοντες μικρὰ παντελῶς ἀπέστελλον, ὀλιγωροῦντες τοῦ δαιμονίου. διὰ δὲ τὴν συμφορὰν ταύτην εἰς μεταμέλειαν ἐλθόντες πάντων τῶν ἐν τῇ Τύρῳ θεῶν ἐμνημόνευον. apoikisthentes gar ek tautēs eiōtheisan en tois emprosthen chronois dekatēn apostellein tōi theōi pantōn tōn eis prosodon piptontōn; hysteron de megalous ktēsamenoi ploutous kai prosodous axiologōteras lambanontes mikra pantelōs apestellon, oligōrountes tou daimoniou. dia de tēn symphoran tautēn eis metameleian elthontes pantōn tōn en tēi Tyrōi theōn emnēmoneuon. ⟨3⟩ ἔπεμψαν δὲ καὶ τοὺς ἐκ τῶν ἱερῶν χρυσοῦς ναοὺς ⟨αὐτοῖς⟩ τοῖς ἀφιδρύμασι πρὸς τὴν ἱκεσίαν, ἡγούμενοι μᾶλλον ἐξιλάσεσθαι τὴν τοῦ θεοῦ μῆνιν τῶν ἀναθημάτων πεμφθέντων ἐπὶ τὴν παραίτησιν. epempsan de kai tous ek tōn hierōn chrysous naous ⟨autois⟩ tois aphidrymasi pros tēn hikesian, hēgoumenoi mallon exilasesthai tēn tou theou mēnin tōn anathēmatōn pemphthentōn epi tēn paraitēsin. ⟨4⟩ ᾐτιῶντο δὲ καὶ τὸν Κρόνον αὑτοῖς ἐναντιοῦσθαι, καθ’ ὅσον ἐν τοῖς ἔμπροσθεν χρόνοις θύοντες τούτῳ τῷ θεῷ τῶν υἱῶν τοὺς κρατίστους ὕστερον ὠνούμενοι λάθρᾳ παῖδας καὶ θρέψαντες ἔπεμπον ἐπὶ τὴν θυσίαν· καὶ ζητήσεως γενομένης εὑρέθησάν τινες τῶν καθιερουργημένων ὑποβολιμαῖοι γεγονότες. ēitiōnto de kai ton Kronon hautois enantiousthai, kath’ hoson en tois emprosthen chronois thyontes toutōi tōi theōi tōn huiōn tous kratistous hysteron ōnoumenoi lathrai paidas kai threpsantes epempon epi tēn thysian; kai zētēseōs genomenēs heurethēsan tines tōn kathierourgēmenōn hypobolimaioi gegonotes. ⟨5⟩ τούτων δὲ λαβόντες ἔννοιαν καὶ τοὺς πολεμίους πρὸς τοῖς τείχεσιν ὁρῶντες στρατοπεδεύοντας ἐδεισιδαιμόνουν ὡς καταλελυκότες τὰς πατρίους τῶν θεῶν τιμάς. διορθώσασθαι δὲ τὰς ἀγνοίας σπεύδοντες διακοσίους μὲν τῶν ἐπιφανεστάτων παίδων προκρίναντες ἔθυσαν δημοσίᾳ· ἄλλοι δ’ ἐν διαβολαῖς ὄντες ἑκουσίως ἑαυτοὺς ἔδοσαν, οὐκ ἐλάττους ὄντες τριακοσίων. toutōn de labontes ennoian kai tous polemious pros tois teichesin horōntes stratopedeuontas edeisidaimonoun hōs katalelykotes tas patrious tōn theōn timas. diorthōsasthai de tas agnoias speudontes diakosious men tōn epiphanestatōn paidōn prokrinantes ethysan dēmosiai; alloi d’ en diabolais ontes hekousiōs heautous edosan, ouk elattous ontes triakosiōn. ⟨6⟩ ἦν δὲ παρ’ αὐτοῖς ἀνδριὰς Κρόνου χαλκοῦς, ἐκτετακὼς τὰς χεῖρας ὑπτίας ἐγκεκλιμένας ἐπὶ τὴν γῆν, ὥστε τὸν ἐπιτεθέντα τῶν παίδων ἀποκυλίεσθαι καὶ πίπτειν εἴς τι χάσμα πλῆρες πυρός. εἰκὸς δὲ καὶ τὸν Εὐριπίδην ἐντεῦθεν εἰληφέναι τὰ μυθολογούμενα παρ’ αὐτῷ περὶ τὴν ἐν Ταύροις θυσίαν, ἐν οἷς εἰσάγει τὴν Ἰφιγένειαν ὑπὸ Ὀρέστου διερωτωμένην ēn de par’ autois andrias Kronou chalkous, ektetakōs tas cheiras hyptias enkeklimenas epi tēn gēn, hōste ton epitethenta tōn paidōn apokyliesthai kai piptein eis ti chasma plēres pyros. eikos de kai ton Euripidēn enteuthen eilēphenai ta mythologoumena par’ autōi peri tēn en Taurois thysian, en hois eisagei tēn Iphigeneian hypo Orestou dierōtōmenēn τάφος δὲ ποῖος δέξεταί μ’, ὅταν θάνω; / πῦρ ἱερὸν ἔνδον χάσμα τ’ εὐρωπὸν χθονός. taphos de poios dexetai m’, hotan thanō? / pyr hieron endon chasma t’ eurōpon chthonos. ⟨7⟩ καὶ ὁ παρὰ τοῖς Ἕλλησι δὲ μῦθος ἐκ παλαιᾶς φήμης παραδεδομένος ὅτι Κρόνος ἠφάνιζε τοὺς ἰδίους παῖδας, παρὰ Καρχηδονίοις φαίνεται διὰ τούτου τοῦ νομίμου τετηρημένος. kai ho para tois Hellēsi de mythos ek palaias phēmēs paradedomenos hoti Kronos ēphanize tous idious paidas, para Karchēdoniois phainetai dia toutou tou nomimou tetērēmenos.

Translation (the author’s own, made from Fischer’s Greek; no published English was opened while it was made, and where a clause already stands in Chapter 2’s footnoted rendering that wording is carried across so the book does not translate the same Greek two ways).

⟨1⟩ And so the Carthaginians, reckoning that the calamity had come upon them from the gods, turned to supplication of the divine power in every form it took; and reckoning that Herakles above all bore them a grudge — the Herakles among the colonists — they sent off to Tyre a great deal of money and no small number of their costliest dedications.

⟨2⟩ For as settlers sent out from that city they had been accustomed, in earlier times, to remit a tenth to the god of everything that came in as revenue; but afterwards, having got great wealth and taking in more considerable revenues, they were remitting next to nothing, holding the divine power cheap. This calamity brought them to a change of mind, and they called to mind every one of the gods in Tyre.

⟨3⟩ They sent, too, for the supplication, the golden shrines out of their sanctuaries, images and all, supposing that they would the better appease the god’s wrath now that the offerings had gone as a petition to be let off.

⟨4⟩ They laid it to Kronos’ charge as well that he had set himself against them, in that in earlier times they had sacrificed to this god the best of their sons, but later, secretly buying children and rearing them, they would send them to the sacrifice; and when an inquiry was held, some of those who had been consecrated were found to have been supposititious.

⟨5⟩ Taking thought of this, and seeing the enemy encamped against their walls, they fell into dread of the gods, as men who had abolished the honors their fathers had paid them. In their haste to set the faults right, they picked out two hundred of the most conspicuous children and sacrificed them in public; and others who stood accused gave themselves up of their own will, and these were not fewer than three hundred.

⟨6⟩ They had among them a bronze Kronos, its hands stretched out palm upward and sloping toward the ground, so that the child laid upon them rolled off and fell into a chasm full of fire. And it is likely that Euripides took from this source the tale he tells of the sacrifice among the Taurians, where he brings on Iphigeneia, questioned by Orestes: “And what grave will receive me, when I die?” “Holy fire within, and a broad chasm of earth.”

⟨7⟩ And the story current among the Greeks too, handed down from ancient report, that Kronos used to make away with his own children, seems to have been preserved among the Carthaginians in this custom.

The words the chapter rests on. ὑποβολιμαῖος, from ὑποβάλλω, “to slip in underneath,” is in Attic usage the supposititious child — the baby smuggled into a household and passed off as a genuine heir, the standing figure of Athenian inheritance litigation. It is not a cultic term: with the fire at their backs the Carthaginians reached for the vocabulary of forged succession. Its strangeness in this setting is itself attested, since codices B and O of Eusebius corrupt it to πτολεμαῖοι. The English “supposititious” is Liddell and Scott’s own gloss for the word and is Chapter 2’s already. καθιερουργημένων is a perfect passive — “those who had been consecrated by slaughter,” the act complete: the fraud was found out in victims already offered. ζήτησις is the word for an official investigation; Diodorus is describing an audit, not a leak. διορθώσασθαι τὰς ἀγνοίας is the language of correcting a text or an account, ἄγνοια being the technical softener for a defective cultic performance (compare the Septuagint’s ἀγνοήματα), so the two hundred are booked as a rectification entry. καταλελυκότες τὰς πατρίους τῶν θεῶν τιμάς uses the political verb for dissolving a constitution: the charge is repeal, not vague impiety.

Two more, on the machine and on the myth. ὕπτιος is palm upward, ἐγκεκλιμένας is sloping, and ἀποκυλίεσθαι is middle-passive — the child rolls; no hand touches it. That is the most polemically shaped sentence in the passage and should be read as such. And ἠφάνιζε is the imperfect of ἀφανίζω, “cause to vanish”: Diodorus has no swallowing verb here, and 20.14.7 is not the Greek confirming the devouring myth. Fischer prints the imperfect and records the aorist ἠφάνισε in ms. F, so the tense carries a variant even though the absence of a devouring verb does not. (Booth’s public-domain English, held here only for comparison, renders the clause “that Saturn devoured his own children,” which is the error this note guards against.) τετηρημένος, from τηρέω, is the verb for observing a custom or keeping a law; Diodorus’ closing verdict is, in its own grammar, this book’s colonial-archive thesis. The Euripides couplet is not a second witness: Diodorus introduces it with εἰκός, “it is likely,” as his own literary conjecture, and the hook is the word χάσμα. Fischer notes at p. 193 that the Euripides manuscripts read πέτρας, “of rock,” where Diodorus’ tradition has χθονός; no edition of Euripides was consulted, so that divergence is reported on Fischer’s authority alone.

One crux the editors cannot settle: at ⟨1⟩ the sense the passage needs is the Herakles of the mother city, since the gifts go to Tyre, but ἄποικοι means the people sent out, not the senders. Fischer keeps ἀποίκοις and records the difficulty in his own apparatus, with Wesseling’s ἀποικισταῖς and Dindorf’s ἀποικίσασι, both meaning the founders. The translation above prints the transmitted text and declines to smuggle “mother city” into the English while the editors are still arguing about it.

Eusebius’ copy. Gifford p. 210 gives the same passage, differing at five substantive points, all confirmed on the page: Ἠιτιῶντο with iota adscript; αὐτοῖς with smooth breathing where Fischer prints αὑτοῖς; καθόσον as one word; λάθρα without iota subscript; and πατρῴους for Fischer’s πατρίους. The last of these is a chosen reading, not a slip: Gifford’s apparatus at b 1 records “πατρῴους] AH, πατρίους cet,” so the two editors split the same manuscript evidence. Punctuation and capitalisation differ throughout — Gifford sets commas after ἔννοιαν, στρατοπεδεύοντας and σπεύδοντες where Fischer has none, and a comma where Fischer sets a high stop before καὶ ζητήσεως, and he capitalises Τούτων, Διορθώσασθαι and Ἦν at sentence openings where Fischer does not. Fischer’s breathing at ⟨4⟩ is worth one line of its own, since the appendix prints the rougher form: his apparatus is set in both marks and offers what he declined, reading “3 αὑτοῖς edd.” and “5 μηνιεῖν RX. αὑτοῖς edd.” (both rough) against “19 αὐτοῖς XF Eus.” (smooth) — so the printed text at line 19 is the rough αὑτοῖς, and Eusebius’ smooth form is exactly the variant Fischer’s note ascribes to him. English cannot carry the difference and does not need to: “set himself against them” either way.

Fischer’s sigla, from his own Notarum explicatio: P = Patmius s. X/XI; X = Venetus Marc. 376 s. XIV/XV; R = Parisinus 1665 s. XI/XII; V = Vaticanus gr. 132 s. XV; F = Florentinus Laur. plut. LXX 12 s. XIV/XV; M = Venetus Marc. VII 8 s. XV; J = Paris. 1661 s. XVI; O = Claromontanus 222 s. XVI; “edd.” = the editions of Bekker, Dindorf, Rhodomann, Stephanus, Vogel and Wesseling.

Kleitarchos, apud the scholion on Plato, Republic 337 A. From Hermann and Wohlrab, Platonis Dialogi vol. 6 (Leipzig: Teubner, 1887), pp. 332–33 (the inner-margin scholia pagination is 188–89), read from the page image.

337 A. ἀνεκάγχασέ τε μάλα Σαρδάνιον] παροιμία ἐπὶ τῶν ἐπ’ ὀλέθρῳ τῷ σφῶν αὐτῶν γελώντων. 337 A. anekanchase te mala Sardanion] paroimia epi tōn ep’ olethrōi tōi sphōn autōn gelōntōn. “‘and he burst out in a great sardonic laugh’] a proverb used of those who laugh to their own destruction.”

[Between that gloss and the Kleitarchos fragment the scholion gives Timaios’ Sardinian aetiology — the islanders clubbing their aged parents into the grave-pit, laughing. It is that derivation, with the Tarrhaean’s herb below it, that the scholion’s own ruling on the spelling disqualifies, and it is omitted here only for length.]

Κλείταρχος δέ φησι τοὺς Φοίνικας, καὶ μάλιστα Καρχηδονίους, τὸν Κρόνον τιμῶντας, ἐπάν τινος μεγάλου κατατυχεῖν σπεύδωσιν, εὔχεσθαι καθ’ ἑνὸς τῶν παίδων, εἰ περιγένοιντο τῶν ἐπιθυμηθέντων, καθαγιεῖν αὐτὸν τῷ θεῷ. τοῦ δὲ Κρόνου χαλκοῦ παρ’ αὐτοῖς ἑστῶτος τὰς χεῖρας ὑπτίας ἐκτετακότος ὑπὲρ κριβάνου χαλκοῦ, τοῦτον ἐκκαίειν τὸ παιδίον. τῆς δὲ φλογὸς τοῦ ἐκκαιομένου πρὸς τὸ σῶμα ἐμπιπτούσης, συνέλκεσθαί τε τα μέλη, καὶ τὸ στόμα σεσηρὸς φαίνεσθαι τοῖς γελῶσι παραπλησίως, ἕως ἂν συσπασθὲν εἰς τὸν κρίβανον παρολίσθῃ. τὸν οὖν σεσηρότα γέλωτα σαρδάνιον ἐντεῦθεν λέγεσθαι, ἐπεὶ γελῶντες ἀποθνήσκουσι· σαίρειν δέ ἐστι τὸ διέλκειν τὸ στόμα καὶ χαίνειν. Kleitarchos de phēsi tous Phoinikas, kai malista Karchēdonious, ton Kronon timōntas, epan tinos megalou katatychein speudōsin, euchesthai kath’ henos tōn paidōn, ei perigenointo tōn epithymēthentōn, kathagiein auton tōi theōi. tou de Kronou chalkou par’ autois hestōtos tas cheiras hyptias ektetakotos hyper kribanou chalkou, touton ekkaiein to paidion. tēs de phlogos tou ekkaiomenou pros to sōma empiptousēs, synelkesthai te ta melē, kai to stoma sesēros phainesthai tois gelōsi paraplēsiōs, heōs an syspasthen eis ton kribanon parolisthēi. ton oun sesērota gelōta sardanion enteuthen legesthai, epei gelōntes apothnēskousi; sairein de esti to dielkein to stoma kai chainein.

“Kleitarchos says that the Phoenicians, and above all the Carthaginians, honouring Kronos, whenever they are bent on getting some great thing, make a vow upon one of their children: that if they come out with what they have set their hearts on, they will devote him wholly to the god. And since Kronos stands among them in bronze, his hands stretched out palm upward over a bronze oven, it is the statue that burns the little child. As the flame off the burning body strikes it, the limbs are drawn together and the mouth appears pulled back in a grin, much like people laughing, until, contracted, it slips off into the oven. That is why the grin that bares the teeth is called sardonic, since they die laughing. And sairein is to draw the mouth wide and gape.”

(τα in συνέλκεσθαί τε τα μέλη is printed unaccented in this Teubner and is reproduced as printed.)

Two things are worth noticing in the Greek. The vow construction is εὔχομαι κατά + genitive, “to vow upon” a designated person — the same construction as the Punic ndr ʾš ndr formula in Appendix A, and as Yiftaḥ’s vow — and καθαγίζω is “to devote wholly, consume entirely for the god,” the nearest thing Greek has to ḥerem. And the machine is a different machine from Diodorus’: ramp and pit there, griddle and oven here. The ramp-and-pit is Diodorus’ and Chapter 2 now describes it as he does — the palms sloping toward the ground, ἐπὶ τὴν γῆν, with the pit where the child ends rather than where the hands point.

The scholion’s own verdict, at the end of the entry, is the reason Chapter 2 flags the etymology instead of building on it:

… ἤκουσα δέ, φησὶν ὁ Ταρραῖος, ἐγχωρίων λεγόντων, ὅτι ἐν Σαρδόνι γίγνοιτο βοτάνη σελίνῳ παραπλησία, ἧς οἱ γευσάμενοι δοκοῦσι μὲν γελῶντες, σπασμῷ δὲ ἀποθνήσκουσιν· οὕτω δὲ Σαρδόνιος ἂν λέγοιτο, καὶ οὐ Σαρδάνιος. μήποτε οὖν τὸ Ὁμηρικόν, ὅθεν καὶ ἡ παροιμία ἴσως ἐρρύη, μείδησε δὲ θυμῷ σαρδάνιον μάλα τοῖον, τὸν ἀπ’ αὐτῶν τῶν χειλῶν γέλωτα καὶ μέχρι τοῦ σεσηρέναι γιγνόμενον σημαίνει. … ēkousa de, phēsin ho Tarrhaios, enchōriōn legontōn, hoti en Sardoni gignoito botanē selinōi paraplēsia, hēs hoi geusamenoi dokousi men gelōntes, spasmōi de apothnēskousin; houtō de Sardonios an legoito, kai ou Sardanios. mēpote oun to Homērikon, hothen kai hē paroimia isōs errhyē, meidēse de thymōi sardanion mala toion, ton ap’ autōn tōn cheilōn gelōta kai mechri tou sesērenai gignomenon sēmainei.

“‘I have heard from the local people,’ says the Tarrhaean, ‘that in Sardinia there grows a plant like celery, and those who taste it seem to be laughing but die in convulsions.’ On that showing, though, the word would be Sardonios, and not Sardanios. Perhaps, then, the Homeric line — from which the proverb probably flowed — ‘and he smiled in his heart a very sardonic smile’ means the laugh that comes from the lips only and goes no further than a baring of the teeth.”

Read the ruling carefully, because it is easy to get backwards. οὕτω, “on that showing,” points back to the sentence immediately before it, the Tarrhaean’s Sardinian herb; what the scholiast disqualifies on the form of the word are the Sardinia derivations, Timaios’ and the Tarrhaean’s, not Kleitarchos’. And the entry closes by referring the proverb to Homer (Od. 20.302) and explaining it as the laugh of the lips only, μέχρι τοῦ σεσηρέναι — which is Kleitarchos’ own σαίρειν gloss with Carthage detached. The honest statement is that the scholion parades four or five incompatible aetiologies, endorses none by name, disqualifies the island derivations on the spelling, and settles for the physical grin. Chapter 2’s “the etymology is ancient and contested, and this book flags it rather than builds on it” is exactly right.

The Suda confirms the outline of the dossier from a second direction. Bekker, Suidae Lexicon (Berlin, 1854), 937–38, s.v. σαρδάνιος γέλως:

… οἱ δὲ ἀπὸ τοῦ σεσηρέναι μετ’ ἀνίας. καί φασιν ἄλλοι τε καὶ Κλείταρχος ἐν Καρχηδόνι ἐν ταῖς μεγάλαις εὐχαῖς παῖδα ταῖς χερσὶ τοῦ Κρόνου ἐπιτιθέντας (ἵδρυται δὲ χαλκοῦς, προβεβλημένας ἔχων τὰς χεῖρας, ὑφ’ ᾧ κρίβανος) ἔπειτα ὑποκαίειν· τὸν δὲ συνελκόμενον ὑπὸ τοῦ πυρὸς δοκεῖν γελᾶν. … hoi de apo tou sesērenai met’ anias. kai phasin alloi te kai Kleitarchos en Karchēdoni en tais megalais euchais paida tais chersi tou Kronou epitithentas (hidrytai de chalkous, probeblēmenas echōn tas cheiras, hyph’ hōi kribanos) epeita hypokaiein; ton de synelkomenon hypo tou pyros dokein gelan. “Others derive it from grinning in pain. And Kleitarchos, among others, says that at Carthage, in the great vows, they set a child on the hands of Kronos — he is set up there in bronze, holding his hands thrust out before him, and beneath him an oven — and then light the fire under it; and the child, drawn together by the fire, seems to laugh.”

Bekker sets -ου- throughout with the Byzantine ligature; it is expanded editorially here. The same entry credits Δήμων with a further Sardinian derivation and Τίμαιος with the parent-clubbing story, so the Suda too gives Kleitarchos as one voice among several.

Rights. Every text printed above is from an edition free of copyright: Fischer’s Teubner of 1906 (public domain in the United States by age of publication, and in the EU and UK since 1 January 2019 — C. Th. Fischer, 1869–1948 — so the text and the apparatus may both be reproduced); Gifford’s Eusebius of 1903 (Gifford d. 1905); Hermann and Wohlrab of 1887; Bekker of 1854. There is no pre-1929 Loeb of Book XX: the Loeb Diodorus did not begin until 1933, and Book XX is Russel M. Geer’s, vol. X, 1954, whose public-domain status in the United States rests on non-renewal rather than on age — a statutory lapse in one country, and not a safe basis for a book distributed internationally. The public-domain English of this passage is Booth’s (London, 1700; this reprint 1814), which is not the source of the translation above. Geer’s Loeb is cited here only in this rights discussion. One disclosure the reader is entitled to, since the book’s standard is that its English owes nothing to anybody: two short runs of the English above coincide with renderings elsewhere in this book — “a bronze statue of Kronos” (Dewrell, 2017) and “appears to have been kept” (Mosca, 1975, there quoting Geer) — and both were reworded rather than defended as coincidence. “Supposititious” for ὑποβολιμαῖος stands, because it is the lexicon’s own gloss and Chapter 2’s word already.