The Historicity of Esther

"On the face of it, the story seems to be true." — Carey A. Moore, Anchor Bible commentary on Esther, 1971¹

"A work that is full of legendary features and improbable incidents and that has no external attestation should be presumed to be fictional." — Michael V. Fox, 1991²

Summary

The book of Esther tells how a Jewish orphan became queen of Persia under King Ahasuerus and saved her people from a decree of destruction, and how the festival of Purim began.  Ahasuerus is Xerxes I, who reigned 486 to 465 BC.  The question is whether the story happened.

The verdict.  Every detail of Esther that can be checked against the Persian world is right, and no checked fact flatly contradicts it except one, the queen.  But nothing outside the Bible confirms any event of the story or any of its characters except the king.  So the book passes every test it can be given and is confirmed by none.  Whether to read it as history therefore rests on two things the evidence does not settle:  how much weight to give Herodotus's queen Amestris, and what kind of book Esther is.  A careful reader on either side can see where each judgement comes from.

The main findings:

1. The book presents itself as history. It dates its events by Xerxes' regnal years and by Babylonian month names, asks of Mordecai's greatness, "are they not written in the Book of the Chronicles of the Kings of Media and Persia?" (10:2)³, and grounds a festival on the events. Michael Fox, a critic, grants that the author "almost certainly meant us to read the book as a precise report of actual historical events."²
2. The dates fit Xerxes exactly. Years 3, 7 and 12 of the book all fall inside his 21-year reign. On the calendar of his own Babylonian tablets, the month in which Esther was taken to the king, Tebet of year 7, began on 22 December 479 BC, after Xerxes had come home from the Greek war.⁴ The four-year gap between Vashti's fall in year 3 and Esther's arrival in year 7 is the war. Fox's objection that Esther was taken "in 480 B.C.E. … when Xerxes was away fighting in Greece" puts the month a year too early.²
3. The court is right in detail. Herodotus, who wrote within about fifty years of Xerxes, independently describes the royal relay post, the king's register of "benefactors" who were rewarded, a benefactor sitting at the palace gate, the seven nobles with the right of entry, the rule that others entered only when announced, decisions made over wine, the rule that a request made at the royal feast cannot be refused, and impaling as a royal punishment. Xerxes' own inscription names India and Kush as the ends of his empire, the two places named in Esther 1:1. Excavation at Susa found the monumental "Gate of Darius," finished by Xerxes, at the official entrance to the palace.
4. The names are right for the period. Of 37 named persons in the story, one is the king, and twelve more bear names found in the Persepolis tablets or on a seal of the period. Millard showed that the Hebrew text preserved these foreign names far better than the Greek versions did. None of this identifies a person: it shows a writer who knew the period's names.
5. Marduka proves less than is claimed. A Babylonian tablet names a Marduka, an accountant on an inspection tour of Susa late under Darius or early under Xerxes, and Ungnad identified him with Mordecai. But the Persepolis tablets name Marduka in at least 15 texts, as a grain supplier and in other roles, and Cameron, who edited two of them, called it "probably a hypocoristic and common Akkadian name."⁵ A name shared with a common name is not an identification.
6. Amestris is the one hard conflict. Herodotus names only one wife of Xerxes, Amestris, the mother of his heirs. He places her at Susa, feared by the king and commanding his guards, in the very winter in which Esther became queen. Defenders answer that Persian kings had "several lawful wives" by Herodotus's own report, that Herodotus never calls Amestris "queen," and that Vashti may be Amestris demoted, but these answers are possibilities, not evidence.
7. The other objections are weaker than they look. The 127 "provinces" are medinot, the word Ezra and Nehemiah use for little Judah, not satrapies. The relative clause in 2:6 can make Mordecai's great-grandfather the exile of 597 BC. Herodotus's rule that kings married only into seven families was broken at once by Darius, who by Herodotus's own account married two daughters of Cyrus. The 75,000 dead is 15,000 in the Old Greek and 70,100 in the other Greek text. The irrevocable law, though, is unattested outside Esther and Daniel.
8. The silences are explained by the gaps in the record. The Babylonian archives fall away sharply in Xerxes' second year, the Persepolis Fortification archive had already ended, and Susa has yielded almost no administrative tablets. Esther is the one Hebrew Bible book missing from Qumran, but the community's calendar left out Purim, and on the copy rates of Ezra and Chronicles a book could be missing by chance about one time in three. A Hebrew Esther existed to be translated into Greek by 114, 78 or 48 BC, and 2 Maccabees, written in the late second century BC, already knows "Mordecai's day."
9. The genre case is real but not evidence. Fox and Levenson call Esther fiction or a "historical novella," and Moore "a historical novel," because of its reversals, banquets, symmetry, coincidences and improbable numbers. Those features show literary skill. They do not show that the events did not happen, and Moore, a critic, concluded that the book has "a historical core."¹

What a conservative reader keeps and gives up.  Keep a real king, a real palace, real customs, a calendar that fits, the right names, and an author who knew Xerxes' court from close up.  Give up any claim that archaeology has found Mordecai, that Vashti has been shown to be Amestris, or that Herodotus confirms the story.  He does not mention it, and his Amestris is a real problem.

Part 1: What the book claims about itself

Esther does not open like a parable.  It opens like an annal:  "The following events happened in the days of Ahasuerus.  (I am referring to that Ahasuerus who used to rule over 127 provinces extending all the way from India to Ethiopia.)"³  The Hebrew Ahashwerosh renders the Persian royal name Khshayarsha, which Greek writes as Xerxes.  Ralph Marcus, editing Josephus, put it plainly:  "The Heb. 'Ahashwêrôsh is, in reality, a rendering of the Persian Khshayarsha of which the Greek form is Xerxes."⁶

Dated events

The book ties its plot to six dates, each by the king's regnal year and a Babylonian month name, the calendar the Jews used under Persia.

The Julian dates come from Parker and Dubberstein's tables, which convert the Babylonian months of each regnal year from the dated tablets.⁴  The approximate days add the day of the month to the first of the month and may be a day off.  Year 12, Adar, began on 23 February 473;  a leap month, a second Adar, followed on 24 March.⁷

The book's appeal to records

Esther twice points to royal records.  The king cannot sleep and has "the book containing the historical records" read to him, and finds that Mordecai had once saved his life and gone unrewarded (6:1-3).  The book ends by asking whether Mordecai's greatness is not "written in the Book of the Chronicles of the Kings of Media and Persia" (10:2).³  That a Persian king kept a written register of men who served him is independently attested;  Part 2 gives Herodotus's account of it.  Moore found the appeal itself telling:  "Only a writer acting in good faith would dare extend such an invitation to his readers."¹  The force of that is limited, since the chronicle was not a book a reader in Judea could open, but it shows what the author was claiming.

Purim

The book's last two chapters found a festival.  "Mordecai wrote these matters down and sent letters to all the Jews who were throughout all the provinces of King Ahasuerus," and the days "are known as Purim, after the name of pur" (9:20, 9:26).³  The word pur is not Hebrew, and the book glosses it:  "pur (that is, the lot) was cast before Haman" (3:7).³  It is the Akkadian word for a lot.  The festival is the book's purpose, and critics read the story as a legend written to explain a festival that already existed.  That reading is taken up in Part 6.

No divine name

The Hebrew book never names God.  A count over the whole Hebrew text finds no divine name of any kind in its 3,057 words, where Ruth, a book less than half as long, has eighteen uses of the divine name.  The count is in the site's article on the Old Testament text.⁸  Mordecai's warning to Esther, "If you keep quiet at this time, liberation and protection for the Jews will appear from another source" (4:14), is as close as the book comes.³  Fox, who reads the book as fiction, judged the silence deliberate:  "Such a violation of expectations is surely no accident."²  The silence bears on what kind of book Esther is, and Part 6 returns to it.  It bears on whether the events happened only through that question.

What the Greek versions add

Esther survives in two Greek versions, and both add six chapters with no Hebrew counterpart:  Mordecai's dream, the texts of the two royal decrees, prayers by Mordecai and Esther, and an interpretation of the dream.  Karen Jobes, who translated Esther for the New English Translation of the Septuagint, judges that the additions "were not a part of the original Greek text of either, but were inserted later into both versions."⁹  The additions supply what the Hebrew lacks, above all God:  the Greek translator added "Explanatory phrases … especially to make God's presence explicit."⁹  The Old Greek also calls the king Artaxerxes, not Xerxes, and Josephus followed it.⁶

The Old Greek ends with a colophon:  "In the fourth year of the reign of Ptolemy and Cleopatra, Dositheus, who said he was a priest and a Leuite, and Ptolemy his son brought the above letter about Phrourai, which they said existed, and Lysimachus son of Ptolemy, one of those in Ierousalem, translated it."⁹  Several Ptolemies reigned with a Cleopatra, so the note fits more than one year.  Jobes gives the three that scholars defend:  "114/13 BCE, 78/77 BCE, or 48 BCE."⁹  On any of them a Hebrew Esther existed in Jerusalem, and was thought worth translating, by the first century BC.

The additions are not evidence for the story.  They are evidence that by the second or first century BC Jews read the Hebrew book as an account of real events, and found its silence about God hard to bear.

Part 2: The Persian setting, tested

Critics and defenders agree on this part.  Moore, who called the book a historical novel, wrote that "much that the author says about Xerxes seems to be quite compatible with what we know of him from other literary and archaeological sources."¹  Levenson, who doubts the book most strongly, grants that "The author knows, for example, about its size, its postal system, and a considerable number of details about its court life … and employs a number of words and a few names of indisputable Persian origin."¹⁰  What follows sets each detail beside its outside witness.

Herodotus on the court

Herodotus of Halicarnassus wrote his Histories around 430 BC, about fifty years after the events of Esther.  He was a Greek outsider with no knowledge of Esther.  His agreements with the book are therefore independent.

The last line matters for Part 4, where the queen is discussed.

Two of these deserve a word.  The register of benefactors is the hinge of Esther's plot, and the book gets its working exactly right:  a service is recorded, the reward can be forgotten, and the record is what brings it back.  The unrefusable request at a royal feast is the hinge of Esther's own plan, which is to make her request at a banquet.  Herodotus tells his story of Xerxes, Artaynte and Amestris with the same device, at the same court, in the same reign.

Xerxes himself

Paton, the most thorough critic of the early twentieth century, granted that "the picture of his character given in Est. as a sensual and capricious despot corresponds with the account of Xerxes given by Herodotus."¹³  He added that such kings were common, so the match proves little.  That is fair.  But the dates fit more narrowly than a type can.

Herodotus says that Xerxes, after retaking Egypt, "summoned a chosen assembly of the best men among the Persians" to plan the invasion of Greece, and that "During four full years from the conquest of Egypt he was preparing the army … and in the course of the fifth year he began his campaign."¹¹  Esther's year 3 banquet, with "The army of Persia and Media … present, as well as the nobles and the officials of the provinces" (1:3), falls in those years of preparation.³  The war took Xerxes to Greece in 480 and kept him at Sardis until 479.  Herodotus then has him go "away to Susa" (9.108).¹¹  Esther arrives in Tebet of year 7, which began on 22 December 479.  William Shea worked through the dates in 1976 and concluded that "Xerxes had at least 3 months to return to Susa before Esther went in to him."¹⁴  The four silent years between Esther 1 and Esther 2 are the years of the war.

A skeptic can fairly reply that a later writer who knew Xerxes' reign and his war could have placed the dates the same way.  That is true.  The fit shows that the author knew the chronology or the events.  It does not show which.

India to Kush

Esther 1:1 has the empire stretch "from India to Ethiopia," Hebrew Hoddu and Kush.  Xerxes' own inscription at Persepolis, the "daiva" text, lists the countries "of which I was king," about thirty of them, and the list includes "Sind" and ends with "Ethiopians."¹⁵  Sind is the Persian Hindu-, the Indus country, and Kusha is Kush.  Neither is in Darius's first list of his countries, carved at Bisitun around 520 BC, which names 23 countries and ends with "Sattagydia, Arachosia, Maka."¹⁵  So "from India to Ethiopia" describes the empire as it stood from late in Darius's reign onward, which includes Xerxes, and not as it stood under Cyrus or Cambyses.

Susa

Esther is set in "Susa the citadel," and its action moves between the king's gate, an outer court, an inner court facing the king's house, and the women's quarters.  Susa was dug by Loftus in the 1850s, by Marcel and Jane Dieulafoy in 1885-86, and by Jean Perrot from 1968 to 1979.  Remy Boucharlat's summary of the results:  Darius made Susa a royal residence "soon after 520";  "The palace (the Apadāna and the Residence) occupied 5 hectares on a 12-hectare artificial platform";  the residence "is organized around three courtyards."¹⁶  Most to the point, the road into the palace "ended at the Gate of Darius, another square construction with a hypostyle central hall, which was begun by Darius and finished by Xerxes … it was the official entrance."¹⁶  "The king's gate," where Mordecai sits and where courtiers wait in Esther, was a real building of Xerxes' time, large enough to hold a court of its own.

Two cautions.  Dieulafoy, who read Esther against his trenches with enthusiasm, also reconstructed a great fortification at Susa that Boucharlat calls "pure fantasy."¹⁶  And Fox objects that "The author is mistaken about the layout of Susa during the Achemenid period," since "the royal palace lay on a mound in the north, while the courtiers and officials dwelt on the south-eastern hill.  In Esther, on the other hand, the royal palace is on the hill, and various officials live there as well."²  The objection turns on what the book means by "Susa the citadel" (1:2, 2:5), as against "the city of Susa" (3:15, 8:15), which it keeps distinct.³  If the "citadel" is only the palace mound, Mordecai lives in the wrong place.  But Darius walled the three main mounds of Susa, about 100 hectares, inside one enclosure, with the palace on the north mound and a citadel on the west.¹⁶  "Susa the citadel" can as well be that walled royal quarter, with the town outside it.  The objection is not decisive either way.

Names and words

Yamauchi counts "thirty or more personal names of Persian and Elamite origin and twelve Persian loan-words in the text of Esther."¹⁷  The Persepolis tablets, the daily administrative records of Darius's reign, contain names that match many of Esther's minor characters:  the eunuchs Mehuman, Bigtha and Carcas, the officials Meres, Marsena and Memucan, Haman's father Hammedatha, Esther's messenger Hathach, and Haman's sons Aridatha and Aridai.¹⁷  The tablets' editor, Richard Hallock, lists each of these as a personal name in his glossary.¹⁸  A seal of the period reads "seal of Parshandatha, son of Artadatha," which Alan Millard called the name of Haman's first son "Only in lacking the final aleph."¹⁹

Counted over the whole book, 37 persons are named in the story itself.  One, the king, is attested as the same man.  Twelve more bear names attested in Persian-period records.  Twenty-four are attested nowhere else.²⁰

Millard's study draws the right limit.  It showed that the Hebrew text had kept the foreign names with care, where the Greek translators mangled them, but:  "To identify the originals of the Persian names in Esther is not to prove the historicity of the story;  it serves to illustrate the care and accuracy of the Jewish copyists when faced with foreign terms."¹⁹  Matching names show that the author knew Persian names of the right period.  They do not show that these men lived.

One name may point to the date.  Yamauchi, following Rudolf Mayer, notes that in the name of Haman's son Vaizatha "the diphthong ai shifted to e between the reigns of Xerxes and Artaxerxes I," so that the form in Esther is "old and authentic."¹⁷  Mayer's argument has not been read here at first hand, so it is reported, not relied on.

Part 3: Marduka

The claim

In 1940-41 the Assyriologist Arthur Ungnad published a Babylonian tablet that names a Marduka.  Yamauchi summarises it:  "The tablet was first noted in 1904 in the Amherst collection in England, but was not published until 1942 by Arthur Ungnad after it had been sold to the Berlin Museum.  Marduka is listed as a sipir (accountant) who made an inspection tour of Susa during the last years of Darius or early years of Xerxes.  It was Ungnad's conviction that 'it is improbable that there were two Mardukas serving as high officials in Susa.'  He, therefore, concluded that this individual is none other than Esther's uncle.  This conclusion has been widely accepted."¹⁷  Yamauchi adds that "According to Gordis it is 'the strongest support thus far for the historical character of the book.'"¹⁷  Gordis's own 1981 article reasons as Ungnad did:  "That there were two officials with the same name at the same time in the same place is scarcely likely."²¹

The Hebrew Mordokhay is the Babylonian Marduka, a short form of a name built on the god Marduk.  Millard notes that the Hebrew is "a more corrupt spelling of Marduka than does the Greek Mardochaios."¹⁹  So the name itself is a real Babylonian name of the period, borne by a Jew in exile, as Zerubbabel ("seed of Babylon") and Sheshbazzar were.

The count

Ungnad's argument is an argument from rarity.  Its force depends on how many Mardukas there were.  The Persepolis archives, published after he wrote, answer that.

Richard Hallock's edition of the Persepolis Fortification tablets, 2,087 administrative texts from Darius's thirteenth to twenty-eighth years, lists Marduka in its glossary as a personal name occurring in thirteen tablets:  "Marduka (m.Mar-du-ka, e.g. PF 81:2 …) PN.  Occurs also in:  PF 412, 790, 863, 941 f., 991, 1183, 1236, 1794."¹⁸  The same entry adds two texts from the Persepolis Treasury.  In these texts Marduka delivers flour, supplies grain to work crews, and sets the rations of others, in Darius's seventeenth to twenty-fourth years, 505 to 498 BC.²²  George Cameron, who edited the Treasury texts, wrote of the name:  "This name (cf. the Hebrew "Mordecai") is probably a hypocoristic and common Akkadian name."⁵

Yamauchi, who defends the book's historicity, gives a higher count and draws the right conclusion from it:  "We now have more than thirty texts … with the name Marduka or Marduku, which may refer to up to four individuals.  Although we cannot be certain, it is possible that one of these may have been the biblical Mordecai."¹⁷  The fifteen texts found in the two published editions are fewer than his thirty, which may include unpublished tablets.  The point stands either way.  In one archive, in about seven years, the name turns up in at least fifteen documents and on more than one man.

Babylonia shows the same.  Matthew Stolper's index to the Murashu archive of Nippur, a generation later, lists nine entries under Marduka, two of them different men named in the same text.²³

What it does and does not show

The Amherst tablet shows that a man named Marduka held office in the Persian administration and had business at Susa within a few years of Xerxes' accession.  That is consistent with Esther, which has Mordecai "sitting at the king's gate," the place of officials.  But the identification rests on the premise that two Mardukas in Persian service at Susa would be unlikely, and the Persepolis tablets have removed that premise.

The critics' answers are fair.  Levenson:  "the notion that Marduka in the Persian inscription and the Mordecai of the book of Esther are the same individual strains credulity and cannot be used to support the historicity of the biblical book."¹⁰  Fox is sharper:  if the inspector really is Mordecai, "it could just as well prove that the Mordecai of our book did not exist," since the tablet's man was an accountant on tour, not a royal favourite.²  Moore, a critic, still allowed that "Mordecai may very well have been a historical personage."¹  J. Stafford Wright went further and proposed that Mordecai was also the eunuch Matacas whom Ctesias names among Xerxes' "confidential advisers."²⁴  Nothing but a loose likeness of sound supports that.

The honest result:  the tablet confirms that the name was real and in use among officials at the right time.  It does not confirm Mordecai.  Apologetic books that say "Mordecai has been found" overstate the evidence.

Part 4: Amestris, Vashti and Esther

The objection

This is the critics' strongest point, and it comes straight from Herodotus.  Carey Moore states it:  "according to ii 16 and ii 7, Esther was queen between the seventh and twelfth years of Xerxes' reign, but according to Herodotus, Amestris was queen then."¹  Fox:  "Xerxes, according to Herodotus, was married to one Amestris (VII 114, IX 112).  She was still his wife in 479, the year Xerxes spent in Ionia after his defeat at Salamis (IX 112)—this was the year after he supposedly married Esther—and Amestris remained his wife in her old age (VII 114)."²

Herodotus mentions Amestris three times.  She was the daughter of Otanes (7.61).¹¹  In her old age she had "twice seven children of Persians who were men of renown" buried alive as an offering for her own life (7.114).¹¹  And in the longest story, Xerxes falls for his brother Masistes' wife at Sardis, arranges a marriage to get near her, goes "away to Susa," and there takes his son's bride, Artaynte, instead.  Amestris has woven Xerxes a robe;  Xerxes gives it to Artaynte;  and at the king's birthday feast Amestris asks for Masistes' wife, has her mutilated by the king's own spearmen, and Xerxes, bound by the custom of the feast, lets it happen (9.108-112).¹¹

Ctesias, a Greek physician at the Persian court around 400 BC, whose history survives only in a ninth-century summary by Photius, names her too:  "Xerxes married Amestris, the daughter of Onophas, who bore him a son, Dariaeus, two years afterwards Hystaspes and Artoxerxes, and two daughters."²⁴  He gives her a different father from Herodotus's, and like Herodotus he names no other wife.

The detail critics have not stressed makes the point sharper.  On Herodotus's own sequence this happened at Susa after Xerxes came back from Sardis in 479, the same winter in which Esther 2:16 has Esther brought to the king.  At the very time the book makes Esther queen "in place of Vashti" (2:17),³ Herodotus has Amestris at Susa, so powerful that Xerxes was afraid of her.

What Herodotus does not say

Three limits on the objection are real.

1. Herodotus never calls Amestris queen. In each of her six mentions in the Greek she is "the wife of Xerxes," Xerxeō gynē, not basileia.²⁵ That she was the leading wife, mother of the heirs, is plain from the story. That she was the only wife is not stated.
2. Persian kings had several wives. Herodotus himself says that each Persian "marries several lawful wives" and keeps "a much larger number of concubines."¹¹ Darius had at least four wives (3.88).¹¹ Herodotus names no other wife of Xerxes, but he names few women of any king.
3. Herodotus is not a court record. He wrote of Xerxes' marriages from Greek informants a generation later, and his Masistes story is a tale of court revenge that he tells for its own sake. His Amestris is a better witness than silence, but she is one witness.

The defenders' three answers

The defenders do not agree with one another, which is itself worth knowing.

Vashti was Amestris.  J. Stafford Wright and William Shea argue that Vashti and Amestris are the same woman.  Shea:  "This conclusion raises the possibility that Amestris, the only queen of Xerxes known from Greek sources, may have been Vashti, the only other queen of Xerxes known from the Bible besides Esther."¹⁴  The names are not far apart if the Persian v could appear as Greek m, and Shea argues that "The terminal consonant in Amestris is certainly a Greek addition."¹⁴  The difficulty is the timing.  Vashti is deposed in year 3, but Amestris is at the centre of the court in year 7.  Shea's answer is that the decree was "not 'divorce' in the modern sense of the word, but rather demotion from her being the chief royal wife,"¹⁴ and that "it is an overstatement of the case to say that Amestris was Xerxes' queen between his 7th and 12th years."¹⁴  Wright adds that "One does not get rid of the mother of royal children during a drunken dinner party, nor does a powerful queen submit to be put away."²⁶  A demoted Vashti who stayed a wife and mother of heirs, and stayed formidable, fits Herodotus.  But the Hebrew says the king should give her royal position "to another who is more deserving than she" (1:19), and no text joins the two names.  The equation is possible and unproven.

Esther was Amestris.  Gordis suggested that "It is not impossible that 'Esther' represented an apocopated form of the name 'Amestris.'"²¹  This fails on Herodotus's own facts:  Amestris was the daughter of a Persian commander and was Xerxes' wife long before year 7.  Fox's reply is fair:  if this were true, Esther in old age "was reported to have ordered nine boys and nine girls buried alive."²  Fox's count is slightly off:  Herodotus gives Amestris "twice seven" children, and the "nine" belongs to a separate Persian sacrifice of boys and maidens at a place called the Nine Ways, told in the same chapter.¹¹  His point stands.

Esther was a wife, not the only queen.  Gordis's other option is that "The historical Esther may not have been Xerxes' principal queen, though clearly a wife and not a concubine."²¹  This fits Herodotus best.  It strains the book, which crowns Esther "queen in place of Vashti."  Jacob Hoschander, one of Esther's most thorough defenders, rejected it as a "forced suggestion."²⁷

The seven families

Herodotus also reports that the seven nobles who made Darius king agreed that the king should not "marry from any other family, but only from those of the men who had made insurrection with him" (3.84).¹¹  Paton, Moore and Fox all use it against Esther, an orphan of a subject people.  Fox calls it a rule that "was in fact never violated."²

But four chapters later Herodotus tells how Darius himself married:  "he married two daughters of Cyrus, Atossa and Artystone … and besides them he married the daughter of Smerdis the son of Cyrus, whose name was Parmys;  and he also took to wife the daughter of Otanes" (3.88).¹¹  Only the last came from a conspirator's house.  Cyrus's line was the royal house Darius was claiming, so one can argue it counted as his own.  Even so, the rule was a pact among seven men in 522 BC, not a law of the empire, and the king who made it married outside it at once.  Xerxes' own mother, Atossa, was one of those wives.  Hoschander points to Herodotus's other report, that the Royal Judges found "a law to the effect that the king of the Persians might do whatsoever he desired" (3.31).¹¹  The seven-families objection is weak.

Where this leaves the queen

The Amestris problem is not solved.  Herodotus gives Xerxes one named wife, places her at Susa in the winter of Esther's crowning, and shows her in command of the women's quarters.  The book gives Xerxes two queens, neither of them Amestris.  Either the book has the queen wrong, or Herodotus knew only the senior wife and the book's "queen" was something less than he means by "wife of Xerxes," or Vashti is Amestris under another name.  Each of the last two is possible.  None is attested.  A defender should say so.

Part 5: The other difficulties

Paton's 1908 commentary set out the objections that later critics repeat, and he set out the case for the book fairly before he rejected it.¹³  Most of the list survives in Moore and Fox.  Each is taken here in the critics' words, then answered in the defenders', then weighed.

The law that cannot be revoked

Esther twice says that a decree written in the king's name "cannot be repealed" (1:19) and "cannot be rescinded" (8:8).³  Daniel says the same of "the law of the Medes and Persians, which cannot be changed" (Daniel 6:8, 6:15).¹²  Paton:  this "is unconfirmed by any ancient evidence.  It is a Jewish legend that is introduced here for the sake of making the decree of Purim more binding."¹³  Fox:  "There was no principle forbidding the Persians to change their laws;  that notion is found nowhere but in Esther and Daniel.  It would be impossible to run a government by such a principle."²

The defenders' evidence is thin.  Wright cites Diodorus on Darius III, who had the Athenian Charidemus executed in a rage and then regretted it, "but all his royal power was not able to undo what was done."²⁸  The man was already dead, so this shows an execution that could not be reversed, not a law that could not be repealed.  Wright himself grants that the passage allows "a different translation."²⁶  Herodotus's Royal Judges point the other way:  they found that "the king of the Persians might do whatsoever he desired."¹¹

Two things can still be said for the book.  First, Esther never has a king repeal anything.  When Haman's decree must be stopped, the king issues a second decree letting the Jews defend themselves (8:11).  The plot works whether irrevocability was law, custom, or a court's way of speaking about a king's sealed word.  Second, Darius's own inscription treats a sealed royal text as fixed and authoritative:  "this inscription as a whole has been confirmed by the impression of a seal."¹⁵  But the specific rule stated in Esther and Daniel is unattested anywhere else.  This is a real point against the book, if a modest one.

127 provinces

Esther says Xerxes ruled "127 provinces extending all the way from India to Ethiopia" (1:1).³  Herodotus says Darius "established twenty provinces, which the Persians themselves call satrapies."¹¹  Fox takes the 127 as satrapies and sees legendary inflation:  "Xerxes is said to have ruled over 127 'provinces,' meaning satrapies … whereas there were actually only 20 to 31 satrapies in the Persian empire.  Esther (as well as the book of Daniel, which speaks of 120 satrapies) exhibits the propensity of legend to inflate qualities and quantities of ancient times."²

The Hebrew word is medinah, and the Bible uses it for units far smaller than a satrapy.  Ezra calls the returning exiles "the people of the province," and Nehemiah's brother speaks of "the remnant that remains from the exile there in the province," meaning Judah, a single district of the satrapy Beyond the River.¹²  Esther itself keeps the levels apart.  The edicts go to "the satraps, and the governors, and the officials of the provinces all the way from India to Ethiopia—127 provinces in all" (8:9).³  Satraps and provinces are different things in the same verse.  Wright:  "the satrapies were divided into smaller units, and the word used in 1:1 (medinah) is used in Ezra 5:8 and Nehemiah 1:3, etc., of the province of Judah which formed part of the larger satrapy."²⁶  Hoschander made the same point in 1923:  "Judaea was a Medinah, not a satrapy."²⁷  Even Moore called the discrepancy "quite minor."¹

The royal lists do not settle the count either way.  Darius's Bisitun inscription names 23 countries, and Xerxes' list at Persepolis about thirty, but these are lists of peoples, not of administrative districts.¹⁵  Nobody has a list of Persian medinot.  The figure 127 is unconfirmed, not contradicted.  Fox's reading of it as "satrapies" is the source of the contradiction, and the book's own usage in 8:9 does not support it.

Mordecai's age

Esther 2:5-6 introduces "Mordecai … the son of Jair, the son of Shimei, the son of Kish, a Benjaminite, who had been taken into exile from Jerusalem with the captives who had been carried into exile with Jeconiah king of Judah."³  If "who" is Mordecai, he was exiled in 597 BC.  Levenson does the arithmetic:  "Mordecai is a minimum of 114 years old when the action of the book of Esther begins in Xerxes's third year."¹⁰  At Esther's crowning in Tebet of year 7 he would be at least 118.

The defenders read "who" as Kish, the last name before the clause.  Wright:  "if the relative pronoun applies to the last name in the genealogy (as in II Chronicles 22:9 and Ezra 2:61), it was Kish, Mordecai's great-grandfather, who was taken with Jehoiachin.  In Esther 2:5, 6, there are three relative pronouns, and each refers to the name that immediately precedes."²⁶  Four generations from an exile of 597 to a man in office in 480 is about thirty years a generation, which is ordinary.  Levenson calls this "questionable" and says only those "determined … to read the book as an accurate historical report" adopt it.¹⁰

Fox himself, in his commentary on the verse, calls the question open:  "It is debatable whether the clause 'who had been taken into exile' refers to Mordecai or to Kish.  The natural antecedent is Mordecai, for he is the one being identified here."²  The grammar allows both readings, and the critics' reading makes the author careless in a way that would have been obvious to any reader who knew when Jeconiah went into exile.  Fox concedes that "The author probably did not make that calculation."²  On the critics' own reading, then, the error is a slip of a genealogy, not a sign that the story was invented.  This objection is weak.

The 75,000 dead and the silence about them

Esther 9:16 has the Jews in the provinces kill "75,000 of their adversaries."³  Moore found nothing improbable in the book except the irrevocable law and "perhaps the battle fatalities for the thirteenth of Adar in ix 16."¹  Fox adds the argument from silence:  Xerxes would not "allow massive, uncontrolled battles, leading to the death of thousands, to take place throughout his empire;  and further that this catastrophe would leave no imprint in the historical record."²

Three points cut this down.

1. The number is not stable. The Old Greek reads "fifteen thousand," and the second Greek text "seventy thousand one hundred."⁹ No argument should rest on 75,000 in particular.
2. Massacres on this scale are attested. Gordis points to Mithradates VI, who in 88 BC ordered the slaughter of every Italian in Asia Minor, "men, women and children of every age," with a reported 80,000 dead in a day.²¹
3. The record for these years is thin. Caroline Waerzeggers, who studied the end of the Babylonian archives, writes that "The number of surviving tablets falls back sharply in the early reign of Xerxes (486-465 BC) when many long-established archives end more or less simultaneously."²⁹ The Persepolis Fortification archive had already stopped in 493. At Susa, "The clay tablets found … are extraordinarily few in comparison to the discoveries at Persepolis," and the royal inscriptions there "deal with the construction of buildings, but none describes political events."¹⁶ Levenson's claim that "we have a large amount of evidence about the Persian period and Xerxes in particular" is true of Herodotus's war and false of Xerxes' provinces after 484.¹⁰

William Shea suggested that the burning of Shechem around 475 BC, dated by Attic pottery, might be a trace of the fighting.  He calls it a "hypothesis," and quotes the excavator's own caution about the period:  "we simply do not know what happened."¹⁴  It should not be counted as evidence.

The silence is real, but it is the silence of a period with few records.

Decrees in every language, and foreign ministers

Paton held that "The issuing of decrees in the languages of all the provinces … was not the ordinary method of the Persian empire.  For this purpose Aramaic was employed."¹³  Aramaic was the empire's working language.  But the one royal proclamation that survives in copies went out in several.  Darius's Bisitun text was carved in Old Persian, Elamite and Akkadian;  an Aramaic version was found on papyrus at Elephantine in Egypt;  and Darius says of it, "Afterwards this inscription was sent by me everywhere among the provinces."¹⁵¹⁵  Esther's "each province in its own script" goes further than that evidence, but the practice it describes is real.

Paton, Moore and Fox also object that two foreigners in a row, Haman and Mordecai, would not have been made the king's first minister.  Wright answers that "we have no Persian or Greek references to any vizier or prime minister in Susa during the period when Haman and Mordecai are said to have held this office.  There is thus no external confirmation or disproof."²⁶  That is the fair statement.  Persian kings did raise Greeks and other foreigners to their table and their councils, but no foreign chief minister of Xerxes is attested.

Part 6: The silences, Purim, and the question of genre

Qumran

Of all the books of the Hebrew Bible, only Esther has not turned up among the Dead Sea Scrolls.  Martin Abegg, Peter Flint and Eugene Ulrich, editors of The Dead Sea Scrolls Bible, state the fact and weigh it:  "the absence of Esther from more than 200 biblical scrolls from the caves at Qumran is a bit more curious.  To be sure, the books of Nehemiah and 1 Chronicles have not been found either."  And:  "That Esther has turned up missing might be attributed to nothing more than chance coupled with the relatively small size of the book."³⁰

The chance can be priced.  The scroll copies of other late or short books are few:  one manuscript of Ezra and none of Nehemiah, one fragment of Chronicles, two of Proverbs, four of the Song of Songs, four of Lamentations.³⁰  If Esther was copied at Qumran as often as Ezra or Chronicles, the chance of finding no copy is about 37 percent.  At the rate of Proverbs it is about 14 percent.  At the rate of the Song of Songs it is about 2 percent.³¹  So the gap is unremarkable if Esther was a little-used book there, and striking if it was used like the other festival scrolls.

The editors do not think it was used.  "The Qumran calendar texts … chart festivals and holy days on the community's 364-day year.  The feast of Purim, which has its beginnings in the story of Esther, is missing."  They offer reasons the community would have rejected the book:  Purim is not in the Law of Moses, a Jewish woman marries a Persian king, God is not named, and the book's vengeance clashes with the community's rule.³⁰  Each of these is a reason for a sect to set the book aside.  None is a reason to think the book did not yet exist.  The absence tells something about Qumran.  It tells nothing about whether the events took place.

The other early witnesses

The earliest outside witness to the festival is 2 Maccabees.  Its account of Judas Maccabeus's victory over Nicanor ends with a decree to celebrate "the thirteenth day of the twelfth month, which is called Adar in the Syrian language … the day before Mardochaios' day."³²  The book opens with a letter dated "In the one hundred eighty-eighth year" of the Seleucid era, 124 BC.³²  So by the late second century BC, Jews celebrated 14 Adar as "Mordecai's day."  The Greek colophon puts a Hebrew Esther in Jerusalem by 114, 78 or 48 BC (Part 1).  Josephus, writing in the 90s AD, retells Esther as history in his Antiquities, and his list of sacred books runs "from the death of Moses, till the reign of Artaxerxes, King of Persia, who reigned after Xerxes," which takes in Esther, since he dates it to that reign.³³⁶

The silence of Jesus ben Sira, a Jerusalem teacher of the second century BC, is often cited.  His praise of famous men ends with Zerubbabel, Joshua the high priest and Nehemiah, and has no Mordecai or Esther.³⁴  But it has no Ezra either, a man nobody doubts, and no Daniel.  A praise-list that leaves out Ezra cannot show that what it leaves out did not exist.  The site's article on dating Daniel makes the same point about Daniel.

The New Testament never mentions Esther, Mordecai or Purim.  It also never mentions many other things its writers knew.  Silence in books written for other purposes weighs very little.

In the Talmud the rabbis argued whether the scroll of Esther made the hands ritually impure, the mark of a sacred book.  Samuel held that it was inspired to be read but "it was not stated that it is to be written."  Rabbi Eliezer answered that it was spoken by the Holy Spirit, because it says "And Haman thought in his heart," which only inspiration could know.³⁵  The debate was about the book's status.  Nobody in it doubted that the events had happened.

Purim and the lot

Critics since Paton have suspected that the story was written to explain a festival that already existed.  Paton:  "Purim, of course, must have had an origin, but it is not necessary that it should have been the origin recorded by Est.  Religious tales are often a secondary invention designed to explain already existing religious institutions."¹³  Moore raised "the legitimate suspicion that Purim is in its origin a pagan festival adopted at some period somewhere along the line by the Jews."¹

That is possible in principle.  But no one has found the pagan festival.  William Hallo, curator of the Yale Babylonian Collection, reviewed the candidates, the Hittite purulli and the Roman Feralia among them, and found that "each of these alleged precedents rests on little more than a dubious assonance, and none of them has anything in the least to do with the casting of lots."³⁶  What Hallo did find was the word.  An Assyrian die of the ninth century BC, inscribed for the vizier Iahali, calls itself his puru, his lot:  "twice in our little cube, and elsewhere in the Assyrian dialect, the word for lot is pûru."³⁶  Esther's gloss "pur (that is, the lot)" uses a real Mesopotamian technical word, correctly explained.

What kind of book is Esther?

The critics' deepest reason for doubting the book is not any one error.  It is the book's shape.  Esther is built on reversals:  Haman builds a stake for Mordecai and is impaled on it, plans Mordecai's humiliation and is made to lead his triumph, and plots a day of slaughter that becomes a day of feasting.  It is a book of banquets, from the king's feast in chapter 1 to the feasts of Purim in chapter 9.  Its turns depend on coincidences, such as the king's sleepless night falling just before Haman comes to ask for Mordecai's death.

Fox draws the conclusion:  the book "in its present shape is certainly a fictional creation with strongly legendary features," and "Various legendary qualities as well as several inaccuracies and implausibilities immediately throw doubt on the book's historicity and give the impression of a writer recalling a vaguely remembered past."²  Levenson:  "the book of Esther is best seen as a historical novella set within the Persian empire.  This is not to say that the book is false, only that its truth, like the truth of any piece of literature, is relative to its genre."¹⁰  Moore:  "Esther is neither pure fact nor pure fiction:  it is a historical novel."¹  Even Gordis, who argued for a historical core, granted that "the role of coincidence, all testify to literary skill of a high order.  By the same token, they raise legitimate doubts with regard to the historical character of the narrative, since the course of human events is never so artistically arranged."²¹

Adele Berlin's commentary for the Jewish Publication Society goes furthest.  As the Islamic Awareness site quotes it, "Esther's attempt to sound like a historical work is tongue in cheek and not to be taken at face value.  The author was not trying to write history, or to convince his audience of the historicity of his story (although later readers certainly took it this way)."³⁷  That reading has not been checked here against Berlin's book.

Three things should be said about the genre argument.

1. It is a judgement, not a finding. Every feature it rests on is a feature of how the story is told. A true story can be told with art. Herodotus tells the story of Xerxes, Artaynte and Amestris with a robe, a promise, a feast and a reversal, at the same court in the same years, and no historian treats Masistes and Artaynte as fictions for that reason.
2. Fox's rule cuts both ways. Fox states his method openly: "A work that is full of legendary features and improbable incidents and that has no external attestation should be presumed to be fictional."² He also states the other side of it: "There are no arguments for the book's historicity, that is to say, positive evidence that the events reported actually occurred. There are only arguments asserting the possibility that they happened."² That second sentence is true, and a defender should accept it. The first sentence is a choice of burden of proof, applied to a period whose records are mostly lost.
3. The author meant it as history. Fox again: "to read Esther as fictional, while a legitimate critical stance, runs contrary to the intentions of the author, who almost certainly meant us to read the book as a precise report of actual historical events."² If the author was writing a novella, he hid the fact very well from every Jewish and Christian reader for two thousand years.

The Islamic Awareness case

The page on Islamic Awareness that argues the book is fiction does so to defend the Quran.  The Quran puts a Haman at Pharaoh's court, and critics call that a confusion with the Haman of Esther.  The page answers that the Haman of Esther may never have existed, so there is nothing for the Quran to have confused.  The site's article on the Quran and earlier literature treats the Quranic side.

Its case about Esther is a compilation of the scholars quoted above, with Bible dictionaries and commentaries.  Its conclusion:  "it is clear that Judeo-Christian scholars do not consider the story to be a genuine historical narrative, and of little or no historical value.  Furthermore, no scholar claimed that the character Haman actually ever existed."³⁷  It grants that "the author shows familiarity and knowledge of Persian life and courtly customs."³⁷

Stated fairly, the page is right about the state of scholarly opinion, and right that the persons of Esther are unattested.  It is wrong in two details.  It says Paton and Moore "both arrived at the same conclusion that the story in the Book of Esther is not historical";³⁷ Moore in fact concluded that "the Book of Esther has a historical core."¹  And it says no scholar has claimed that Haman existed, but defenders such as Wright and Shea read the whole story, Haman included, as history.  None of this changes its main point, which is about opinion, not evidence.

Objections and responses

1.  "Herodotus names Xerxes' queen, and she is Amestris.  That settles it."  It is the strongest objection, and Part 4 grants that it is not answered.  But Herodotus calls her "wife," not "queen," reports that Persians took several lawful wives, and names no other wife of Xerxes because he names few royal women at all.  The objection shows that the book's account of the queens is unconfirmed and in tension with Herodotus.  It does not show that a Jewish wife of Xerxes is impossible.

2.  "Accurate background proves nothing.  Historical novels get the background right."  True.  Paton said it in 1908 with the Arabian Nights as his example.¹³  The background makes the book's author a person who knew Xerxes' court and calendar.  It does not make the plot true.  It does rule out the older critical claim, still repeated, that the book is a late fantasy whose author knew little of Persia, and it narrows the time and place of writing.

3.  "If the events were real, someone outside the Bible would have recorded them."  The Greek historians wrote about the war, not the provinces.  The Babylonian archives fell away after Xerxes' second year.  Susa has yielded almost no administrative texts.  A massacre in the provinces in 473 BC would be expected to leave a trace in a well-documented period.  These years are not one.

4.  "Marduka proves Mordecai was real."  No.  The name is common in the Persepolis and Babylonian records, and the Amherst tablet's Marduka was an accountant on tour.  The tablet shows that the name was used by officials at Susa in the right years.  That is all.

5.  "Vashti is Amestris, so there is no contradiction."  The equation is possible and unproven.  It needs a demoted queen who stayed formidable, and Herodotus's Amestris fits that.  But no text joins the names, the sounds are not an easy match, and the defenders themselves disagree on whether Amestris is Vashti or Esther.

6.  "Esther does not name God, so it is a secular tale and cannot be Scripture or history."  The silence bears on the book's purpose and theology.  It does not bear on whether its events happened.  Many accurate histories do not name God.

7.  "The 75,000, the 180-day banquet, the 10,000 talents and the 50-cubit stake are exaggerated numbers, the marks of a farce."  Some of these are not what they seem.  The 180 days is the length of a display of wealth before the war (1:4), and Shea ties it to the war council Herodotus describes.¹⁴  The 75,000 differs across the versions.  The 50 cubits may be the height of a stake set on a structure, or popular exaggeration, as Wright allows.²⁶  Large numbers in ancient narrative are a real difficulty, and the site's article on ancient historiography treats how ancient writers handled them.

8.  "The Qumran community knew every other book of the Bible and rejected Esther, so it was not yet Scripture or not yet written."  The community rejected Purim from its calendar.  That is a reason for it to have left Esther aside.  The Greek colophon and 2 Maccabees show that the book and its festival existed in the second century BC.

9.  "Defenders only ever show that the story might be true."  That is correct, and Fox says so.  There is no positive evidence for any event of the book.  The defender's case is that the book passes every test that can be applied and that the main contradiction, the queen, has possible answers.  That is a case for plausibility, not proof.

10.  "Critics only ever assume that the story is false."  Not quite.  The Amestris problem and the unattested irrevocable law are evidence, not assumption.  But most of the case, the genre argument and Fox's rule that an unattested story "should be presumed to be fictional," is a judgement about burden of proof, not a finding.

Issues and unanswered questions

1. Ungnad's tablet has not been read here. Ungnad's 1940-41 article in Zeitschrift für die alttestamentliche Wissenschaft and his 1942-43 note could not be obtained. The tablet's exact wording, date and place are known here only through Yamauchi, Gordis and Wright. The sources disagree on whether the tablet comes from Borsippa or elsewhere and whether it names the Marduka's superior.
2. Clines and Berlin have not been read here. David Clines's "In Quest of the Historical Mordecai" (Vetus Testamentum, 1991), the standard critique of the Marduka identification, and Adele Berlin's commentary and her 2001 Journal of Biblical Literature article, the strongest statements of the "burlesque" reading, were not obtained. Berlin is quoted here only through the Islamic Awareness page.
3. The Persepolis count. Hallock's and Cameron's editions give fifteen Marduka texts. Yamauchi reports "more than thirty," probably counting unpublished tablets. Neither count says how many men bore the name. A study of the seals and offices on each tablet could settle whether they are one man or several.
4. Vashti's name. No Old Persian or Elamite form of Vashti has been found. Until one is, both its supposed link to Amestris and its supposed meaning "beautiful" are guesses.
5. The date of writing. Nothing in this article fixes when the book was written. The Persian words and the accurate Persian setting favour the Persian period. Critics generally place it in the fourth or third century BC. The site's planned article on the biblical canon is the place for the book's reception.
6. Yamauchi's Vaizatha argument. Mayer's claim that the diphthong in Vaizatha dates the name to before Artaxerxes I was read only through Yamauchi. If it holds, it is the strongest linguistic evidence for an early date of composition.
7. Herodotus's sources for Xerxes' women. How Herodotus knew of Amestris and Artaynte, and whether his informants would have known of other wives, has not been studied here. Ctesias was read only in Photius's summary and an English translation, not in the Greek fragments.

Verdict

Esther is consistent with history but not confirmed by it.  Its author knew the court of Xerxes in detail:  its calendar, its palace gate, its post, its register of benefactors, its seven nobles, its councils over wine, its unrefusable feast-day promise, its names and its words.  Its dates fall where Xerxes' own records put them, and its four-year silence fits his war.  No checked fact contradicts it except the queen.  Herodotus's Amestris, wife of Xerxes and powerful at Susa in the very winter the book makes Esther queen, is a real problem with no attested answer.  The irrevocable law is attested nowhere else.  None of the book's persons except the king is attested, and Marduka does not change that.

The skeptic is entitled to say that a story with no outside attestation, a single sharp conflict with the best Greek source, and the shape of a well-made tale has not been shown to be history.  The conservative reader is entitled to say that a book which passes every other test, written by someone who knew the court this well, set in years whose records are mostly lost, and meant by its author as history, has not been shown to be fiction.  Between those two positions the evidence does not decide.  The decision rests on how much weight Herodotus's silence about Esther and his Amestris should carry, and on what kind of book one takes Esther to be.

Sources