Who Wrote Isaiah 40-66?

"Even the assumption that Isaiah received supernatural revelations cannot remove this circumstance." — Wilhelm Gesenius, 1821, on the prophet's standpoint (translated)¹

"Assuming that such uniqueness is not possible, one has to conclude that chs. 40-55 were written in the 6th century B.C." — John Oswalt, defending one Isaiah, 1998²

Summary

The book of Isaiah carries the name of a prophet who worked in Jerusalem about 740-700 BC.  Its chapters 40-66 speak to exiles in Babylon and name Cyrus, the Persian king who freed them in 538 BC.  Since the late eighteenth century most scholars have assigned 40-55 to a prophet of the exile, "Second Isaiah," and 56-66 to writers after the return, "Third Isaiah."  Defenders of one author, from Oswald Allis to John Oswalt, answer that this date rests on a refusal to believe that Isaiah could foresee the future.  This article states both cases in their makers' words and tests that charge.

1. The late date is most probably right for chapters 40-55.  The prophet speaks to the exiles in person, treats Jerusalem's fall as past, and describes Cyrus as already winning.  He also appeals to "former things" that have already come true as proof that his new predictions will.  These points place the writer in the 540s BC, between Cyrus's rise and Babylon's fall in 539.

2. The main arguments for the late date do not assume that prophecy is impossible.  Gesenius, who made the theory standard in 1821, said that even granting Isaiah supernatural revelation would not remove the problem of the standpoint.  Driver, Williamson and Blenkinsopp say the same.  The "former things" argument even presupposes that prediction works.  Oswalt's charge does hit some critics, such as Duhm, but not the core of the case.

3. The late date is not proved.  The standpoint argument rests on a rule, that prophets speak to their own generation, which is drawn partly from texts dated late by the same rule.  A God who could name Cyrus in advance could also inspire a prophet to address people not yet born.  Oswalt grants that the single-author view makes the book "unique."  It is coherent, and its cost is that it needs a case without parallel.

4. The case for a separate Third Isaiah is weaker.  Its founder, Duhm, did not argue it at length, and critics still disagree whether 56-66 come from one writer or several.

5. The outside evidence shows how the book was read, not who wrote it.  The Great Isaiah Scroll, copied about 150-100 BC, runs from chapter 39 into 40 without a break.  Ben Sira, about 180 BC, praised Isaiah in words taken from chapter 61.  The New Testament names Isaiah for quotations from both halves, 12 verses for chapters 40-66 and 10 for 1-39.  All of this is at least 360 years after Cyrus.  Josephus's story that Cyrus read his own name in Isaiah is late, unsupported by Ezra, and built on Josephus's own chronology.  The Cyrus Cylinder shares royal language with Isaiah 45 but fits either date.

6. Style statistics split.  Radday's computer study found the halves too different for one author;  Adams and Rencher's found them closer to each other than to other books.  The two studies ask different questions, and neither can tell one author from a disciple steeped in his style.

7. What the date means for the argument from prophecy is narrow.  On the late date, Cyrus by name drops out of the evidence, and the fall of Babylon and the return become accurate forecasts made a few years ahead.  Isaiah 53 does not drop out:  it was copied in a pre-Christian scroll on any date, so its case against the Gospels is untouched by this question.

Part 1: The question and its history

What the book says about itself

The book of Isaiah opens with the heading of a prophet who worked in Jerusalem "during the time when Uzziah, Jotham, Ahaz, and Hezekiah reigned over Judah," roughly 740-700 BC.³  Chapters 36-39 tell of Sennacherib's siege of Jerusalem in 701 BC and end with Isaiah warning Hezekiah that his treasures will one day be carried to Babylon.  Chapter 40 then begins in a different world:  "Comfort, comfort my people ... tell her that her time of warfare is over, that her punishment is completed."³  The exiles are told to "Leave Babylon!"³  And a Persian king is named twice, as the one who will "decree concerning Jerusalem, ‘She will be rebuilt’":  "Cyrus, whose right hand I hold in order to subdue nations before him."³  Cyrus took Babylon in 539 BC, about 160 years after Isaiah's last dated oracle.

No verse in chapters 40-66 names Isaiah.  His name appears in the headings at 1:1, 2:1 and 13:1 and in the stories of chapters 7, 20 and 37-39, and nowhere after 39:8.⁴  The book as a whole carries his name, and that is the starting point of both sides.

How the theory grew

The first hint came from Abraham ibn Ezra, a Spanish Jewish commentator who died in 1164.  He opened his notes on chapter 40 by explaining why comfort follows chapter 39's warning about Babylon.  Then he added a remark about another book:  "the opinion of the orthodox, that the book of Samuel was written by Samuel, is correct as regards the first part, till the words ‘And Samuel died’."⁵  He did not say what Samuel had to do with Isaiah.  His English translator, Michael Friedländer, took the point to be that "the second part of the book of Isaiah, which contains allusions to events that took place long after the death of Isaiah, as to historical facts, is, according to I. E., not written by the same prophet."⁵  The same note shows the limit of the hint.  Ibn Ezra himself said the comfort of chapter 40 looked past the Babylonian exile to "the coming redemption from our present exile," with Babylon only "an illustration."

The modern theory began in Germany in the 1770s and 1780s.  J. C. Döderlein and J. G. Eichhorn argued that chapters 40-66 came from a later hand, and Wilhelm Gesenius's commentary of 1821 made the view standard.⁶⁷  Writers date Döderlein's first statement differently, between 1775 and the late 1780s.⁶  Bernhard Duhm's commentary of 1892 gave the theory its classic form.⁶  Duhm also split off chapters 56-66 as the work of a third writer, "Trito-Isaiah," from after the return from exile.⁸  Blenkinsopp, himself a critical scholar, remarks that Duhm took this step "(but not argued at any length)."⁷

Most critical scholars now speak of three bodies of material:  First Isaiah (1-39, with later additions), Second Isaiah (40-55, the 540s BC) and Third Isaiah (56-66, after 538 BC).  The most recent form of the theory stresses that the book is still one book.  H. G. M. Williamson of Oxford argued in 1994 that the prophet of chapters 40-55 did not merely follow Isaiah's book but edited it, joining his own work to "the literary deposit of Isaiah of Jerusalem" at chapter 33.⁶  Brevard Childs of Yale read the book as Scripture in its final shape, and denied that its unity "can be formulated in terms of single authorship."⁹  A few critics went further the other way.  C. C. Torrey of Yale argued in 1928 that the names of Cyrus and Babylon were "palpable interpolations," and that without them nothing showed the prophet "had ever heard of Jewish exiles in Babylonia."¹⁰

Against all of these, Oswald Allis (1950), E. J. Young (1958), Gleason Archer (1964), R. K. Harrison (1969), John Oswalt (1986, 1998) and Alec Motyer (1993) defended one author, the eighth-century Isaiah.  Their case is set out in Part 3.

What turns on the answer

The date matters for the argument from prophecy at one point above all.  If Isaiah wrote 44:28 around 700 BC, he named a king about 160 years before that king freed the Jews, which would be one of the most striking predictions in any literature.  If a prophet wrote it in the 540s, while Cyrus was already conquering, the name is no prediction at all.  Prophecies About Nations and Cities scores the Cyrus oracle and finds that its value rests entirely on the date, which is Test 1 of How to Test a Fulfilled Prophecy.  This article tries to settle that test, as far as the evidence allows.

The question this article asks is narrower than "who wrote it."  It is whether the case for a late date depends on denying that a prophet can predict.  Oswalt says it does:  the late date "rests squarely on the prior conviction that Isaiah of Jerusalem could not have known the future in any supernatural way."²  If he is right, a reader who allows prediction has no reason to accept the late date.  If he is wrong, even a believer has to weigh the critics' evidence.  Each side is stated below in its own words before it is weighed.

Part 2: The case for a prophet of the exile, in its makers' words

The standpoint: the exile is assumed, not predicted

The oldest and still the main argument is about where the speaker stands.  Gesenius put it in 1821.  Jerusalem lies in ruins, Babylon is at the height of its pride but near its fall, and Cyrus has already conquered many kingdoms.  "None of this is described as future," he wrote.  "It is the prophet's standpoint, the present condition of the nation and its neighbours, to which the predictions of the future are then attached."  And he answered the believer's reply in advance:  "Even the assumption that Isaiah received supernatural revelations cannot remove this circumstance.  For Isaiah, on the analogy of all the other prophecies, would still have had to start from his own time and look from there into the future."¹

S. R. Driver's Introduction of 1910, the standard English statement of the case, made the same point.  "Those whom the prophet addresses, and, moreover, addresses in person ... are not the men of Jerusalem, contemporaries of Ahaz and Hezekiah, or even of Manasseh;  they are the exiles in Babylonia."  Driver drew the contrast with Jeremiah and Ezekiel, who first predict the exile and then the return:  "In the present prophecy there is no prediction of exile :  the exile is not announced as something still future ;  it is presupposed, and only the release from it is predicted."  His rule was that "the prophet speaks always, in the first instance, to his own contemporaries."¹¹

The texts do read this way.  Jerusalem's "punishment is completed."³  The exiles are told to "Leave Babylon!"³  Later chapters speak of the temple as already burned:  "Our holy temple, our pride and joy, the place where our ancestors praised you, has been burned with fire."³

Cyrus is already on the march

The second argument is that the Cyrus of these chapters is a man already winning, not a man not yet born.  Driver:  "the prophet does not claim foreknowledge of Cyrus, but only of what he will accomplish :  he is already "stirred up," and "come"."¹¹  The verse he means reads, "I have stirred up one out of the north and he advances."³  Claus Westermann, whose commentary (German 1966, English 1969) became a standard critical treatment of chapters 40-66, read the verses after 41:25 the same way:  "They do not announce it in advance.  They look back on it, and reflect the stir which it caused throughout the whole world at the time."¹²  He dated the prophet "probably some time after 550, the year which saw the beginning of the victorious campaign of Cyrus."¹²

The argument from "former things"

H. G. M. Williamson calls a third argument "the strongest."  In chapters 41-48 the prophet points to earlier predictions that have already come true, "the former things," as grounds for trusting his new ones.  Williamson's point is that this argument "both presupposes an acceptance of the validity of predictive prophecy, and also demands that the speaker should himself be located after the fulfilment of those predictions."⁶  Brevard Childs set out its logic:  "Because the logic of the idiom demanded that the speaker be standing in between the former and latter things, the prophecy could not have been written by the eighth-century prophet Isaiah, but only by the sixth-century Second Isaiah."¹³

On this reading the "former things" include Cyrus's first victories.  Christopher North argued in 1950 that they "are the victories of Cyrus up to the fall of Sardis in 547.  These victories are described in xli. 1-5, which is contemporary, or nearly so, with the events."  When critics objected that events so recent could hardly be called "former," North replied, "I do not see why not."¹⁴

The hopes outrun the history

Gesenius had a further argument, and it is the one that most clearly rests on a view of what real prediction would look like.  The prophet's picture of the present is exact, he said, but his picture of the future is of "inspired hopes," and "reality fell far short" of them.  So:  "Had Providence seen fit to let Isaiah prophesy of Cyrus in a supernatural way, it would also have disclosed to this prophet the state of things after the exile as it truly was, not in inspired ideals that were never realized."¹  Williamson notes the same problem from the other side.  It "has frequently been maintained" that the closing chapters were written partly to answer "the apparent failure of these predictions to be borne out."⁶

Style and theology

Driver's second and third lines of argument were style and ideas.  The style of 40-66 "is very different from that of Isaiah," in ways that "stand in no appreciable relation to the subject."¹¹  And its theology, with God as creator, "the First and the Last, the Incomparable One," is fuller than that of chapters 1-39.¹¹  Both are weighed in Part 4.

Why chapters 56-66 are split off

The case for a separate "Third Isaiah" is the thinnest part of the theory, and its own makers say so.  Duhm proposed it in 1892, and Blenkinsopp remarks that Duhm took the step "(but not argued at any length)."⁷  Westermann accepted it because chapters 56-66 lean so heavily on 40-55:  "The primary element in chs. 56-66 is so obviously dependent on Deutero-Isaiah that the hypothesis of discipleship alone explains it."¹²  The usual evidence is that these chapters seem to address people back in Judah, arguing over the temple, fasting and foreigners, rather than exiles waiting to leave Babylon.  Whether one disciple, several, or the prophet of 40-55 himself wrote them later is still disputed among critics.

The critics' own answer to Oswalt

Do these critics deny that prophets can predict?  The ones quoted here say they do not.  Driver, a churchman, wrote that the late date "in no way impairs the theological value of the prophecy, or reduces it to a vaticinium ex eventu," a prophecy after the event, because the chapters were written "prior to the events which it declares to be approaching," and because the text "nowhere claims either to be written by Isaiah, or to have originated in his age."¹¹  Williamson:  the view "is not necessarily motivated by a wish to circumvent the possibility of predictive prophecy.  Indeed, there remains plenty of ‘prediction’, both general and specific, within Deutero-Isaiah itself."⁶  Westermann treats the argument from prophecy as central to the prophet's message:  "Prediction of an event is thus complemented by the event’s interpretation in retrospect."¹²

Blenkinsopp gives the critics' real reason most plainly.  "For the majority of critical scholars this view of the matter is not, as conservative polemicists argue, dictated by a disposition to rule out the possibility of predictive prophecy."  The difficulty is that on the early date "Isaiah, active in the eighth century b.c.e., was comforting his people in view of a disaster — the fall of Jerusalem and deportation — that was still a century and a half in the future, a not very plausible scenario."⁷  That is a judgment of likelihood, not a rule against miracles.  But it is a judgment, and Part 4 asks what it rests on.

Part 3: The case for one Isaiah, in its defenders' words

The dispute is about the supernatural

The defenders of one author agree that the critics' strongest card is the Babylonian setting.  E. J. Young of Westminster Theological Seminary called it "the most formidable argument which must be faced by the defenders of the unity of the book," and he placed the dispute where Oswalt later did:  "Whether Isaiah could have predicted the coming of Cyrus brings us face to face with the question of the supernatural."¹⁵  Oswald Allis put the Cyrus oracle as a dilemma.  It "either offers conclusive proof of the unique inspiration of the prophet who uttered it more than 140 years before its fulfilment," or it shows that it and "probably other and extended portions" of the book come "from a time long after the close of the ministry of the son of Amoz."⁸

John Oswalt, whose two-volume commentary in the New International Commentary series is the fullest modern defence, argues that the late date has no evidence of its own.  "This hypothesis is necessitated by an inability to accept the claims of the book, not by independent data."²  He is candid about what his own view requires.  "If Isaiah of Jerusalem did write these chapters, then he had a knowledge of the future that was more detailed than that displayed by other OT prophets.  Furthermore, this view means that these chapters are speaking to people in the future, not merely about them.  This is also unique."²  He grants that "It does require that we accept the possibility of revelation and prediction."  His conclusion is that "perhaps it is the scholarly understanding of the phenomenon of biblical prophecy that needs to be corrected, not the traditional view of the book’s authorship."²

The answer on the standpoint

R. K. Harrison of Wycliffe College, Toronto, answered the standpoint argument by pointing to the critics' own practice.  Christopher North had read the fourth Servant Song (52:13-53:12) as sufferings "viewed from a futuristic standpoint."  If a later prophet could do that, why not Isaiah?  "If pronouncements from a futuristic standpoint are denied to one, they ought to be denied to all.  But if the existence of such phenomena is in fact conceded, all a priori objections to a Proto-Isaiah’s occasionally adopting an ideal standpoint are immediately removed."¹⁶

Alec Motyer, in his 1993 commentary, argued that the chapters are less Babylonian than they look.  Babylon is named four times in chapters 40-55.  "Otherwise there is little that is exclusively or typically Babylonian about the chapters save that Cyrus is named as Babylon's conqueror (44:28; 45:1)."  There is no eyewitness detail of the city.  So "‘Babylon’ is predicted rather than presupposed, and is something foreseen in very broad terms at that."  And at one point "The mask definitely slips in 52:11 where the prophet adopts a position certainly not in Babylonia."¹⁷  The verse reads:  "Leave!  Leave!  Get out of there!"³  The speaker says "there," not "here."  Kenneth Kitchen counts the same way:  "That 40-55 is based in Babylon is simply not true," since Babylon is named only four times in those chapters and nine times in 1-39.¹⁸

Harrison added that the idolatry the chapters attack is mostly the old Canaanite kind, the worship on hills and under trees that Isaiah's own generation knew.  "If this section was written in Babylon by an unknown prophet of the exile, as liberal scholars have so commonly assumed, it is curious that the author should have been so actively preoccupied with something which had long since become a dead issue."¹⁶  And a prophet of the exile, he argued, "could hardly have failed to express himself concretely and specifically concerning the social, moral, and religious conditions of that era."¹⁶

The answer on Cyrus

Allis argued from the shape of the poem in 44:24-28.  Its lines build to the name of Cyrus as their climax, which places Cyrus in the far future of the poem rather than in its present.  "If it had been the aim of the poet to represent the desolations of Israel as already taken place, the exile as nearly ended, and Cyrus as already present, an invincible warrior on the point of attacking Babylon, the structure of the poem is ill-calculated, to say the least, to bring out and emphasize these important matters."⁸

Gleason Archer argued that the times called for a sign of this size.  "It is important to observe that the historical situation confronting Isaiah in 690 B.C. gave ample warrant for so unusual a sign as the prediction of Cyrus by name 150 years in advance of the fall of Babylon."¹⁹  Judah's sins, he said, required an exile, and a return from so far away was otherwise beyond hope.  The defenders also point to a precedent inside the Bible.  A prophet tells Jeroboam I, about 930 BC, that "a son named Josiah will be born to the Davidic dynasty" and will burn human bones on the altar at Bethel, and 2 Kings 23 records that Josiah did so "just as in the Lord’s message."³

Motyer turned the critics' Cyrus argument back on them.  The chapters say that God announced these things long before.  "Refusal to countenance the prediction of a personal name (44:28; 45:1) necessitates a date within Cyrus’s career and, in consequence, predictions ‘from the beginning’ or ‘beforetime’ must be explained by conjuring up lost prophecies."¹⁷

The answer on the "former things"

The defenders deny that the "former things" force the speaker to stand after their fulfilment.  Oswalt follows the Belgian scholar Antoon Schoors, who took the phrase to mean "statements in the past that have now been fulfilled."  On that reading, "the emphasis seems to be strictly on God’s ability to predict the future and then to fulfill that prediction when the time comes."²  Motyer read 48:3-8 as an argument about a pattern, not a date:  "the Lord himself designed the procedure of undated prediction coupled with sudden fulfilment."  "No time-factor is built into the argument."¹⁷

The witnesses

The defenders add the testimony of ancient readers.  Ben Sira, about 180 BC, praised Isaiah because "he comforted those who mourned in Sion" and "showed the things that will be and the hidden things before they even happened."²⁰  Allis read the first phrase as an echo of Isaiah 40:1 and 61:2-3, which "may be regarded as presupposing the entire Book of Consolation to be Isaianic."⁸  The Great Isaiah Scroll from Qumran runs from chapter 39 into chapter 40 without a break.  Young:  "Chapter 40 begins on the last line of the page, and there is no special indentation.  There is thus no unusual break between the two chapters.  The only place in the manuscript where there is a break (of about three lines) is after chapter 33."¹⁵  The New Testament quotes Isaiah by name about twenty times, "from both parts of the book."⁸  Archer called John 12:38-41 "the most conclusive New Testament citation," because it quotes Isaiah 53:1 and Isaiah 6:10 side by side as the words of one Isaiah:  "If it was not the same author who composed both chapter 6 and chapter 53 ... then the inspired apostle himself must have been in error."¹⁹

The book's signature

The title "the Holy One of Israel" runs through both halves and is rare elsewhere.  Oswalt counts it "13 times in chs. 1-39 and 16 times in chs. 40-66 and only 7 times elsewhere in the Bible."²¹  Archer gives 12 and 14.¹⁹  A count of the English phrase in the King James Version gives 12, 13 and 6.²²  The counts differ with what is counted, but every count shows the same pattern.

Part 4: Which arguments assume that prophecy is impossible?

Oswalt's charge can be checked argument by argument.  An argument assumes that prophecy is impossible if it would lose its force for a reader who allowed that God can reveal the future.  An argument that keeps its force for that reader is evidence, and a believer has to weigh it too.

"Cyrus is named, so the text is later than Cyrus"

As a bare inference, this does assume that prediction is impossible.  It also carries more weight in the debate than its makers like to admit.  Torrey, a critic, wrote that "the entire unequivocal evidence" for a Babylonian setting "consists of the mention of Cyrus twice by name, and the presence of the names ‘ Babylon’ and ‘Chaldæa’ in three other passages."¹⁰  But the leading critics do not use the inference in that form.  Driver's point was not that Cyrus is named, but that the Cyrus of these chapters is already "stirred up."¹¹  Westermann's was that 41:25-29 looks back on his rise.¹²  Those are claims about the text's tenses and viewpoint, and they stand or fall on the next argument.

The defenders' answers here do not supply evidence for an early date either.  Archer's "ample warrant" shows that a sign of this kind would have had a purpose in 690 BC.¹⁹  It does not show that one was given.  The Josiah precedent in 1 Kings 13:2 has the same limit.  No copy of Kings older than Josiah exists, so the verse cannot show a skeptic that names were ever predicted, as How to Test a Fulfilled Prophecy explains.  Allis's argument from the poem's climax is better, because it is about the text, but it shows only that the name belongs to the poem and was not added later.  Against Torrey that is a fair point.  It does not show when the poem was written.

The standpoint

This is the argument that matters, and it does not assume that prophecy is impossible.  Gesenius said so in 1821:  "Even the assumption that Isaiah received supernatural revelations cannot remove this circumstance."¹  What it assumes instead is a rule about how prophets speak:  they start from their own time and look forward.  Driver put it as "the prophet speaks always, in the first instance, to his own contemporaries."¹¹  Gesenius appealed to "the analogy of all the other prophecies."

The rule is real.  Jeremiah and Ezekiel announce the exile and then the return.  Isaiah 39:6-7 does the same, and the defenders read that chapter as the hinge into 40.³  What chapters 40-55 do is different in kind and in scale.  For sixteen chapters the prophet argues with, comforts and commands people who, on the early date, would not be born for a century and a half, and he never tells them that the exile he takes for granted is still to come.

Harrison's reply is fair as far as it goes.  The critics themselves allow a prophet to speak "from a futuristic standpoint" about the Servant.¹⁶  But the Servant Song looks ahead from a speaker's present to someone else's suffering.  It does not address a future generation as its own audience.  The rule the critics use is not that prophets cannot see far ahead.  It is that they speak to the people in front of them.

The rule has two weaknesses that should be stated plainly.  First, it is drawn from the prophetic books, and the passages that would break it are the ones in dispute.  Driver himself assigned Isaiah 13:2-14:23, the oracle on Babylon's fall, to the same late period.¹¹  So the rule is partly confirmed by dating the counter-examples late on the same rule.  Second, a God who could name Cyrus in advance could also inspire a prophet to speak to the exiles in advance.  Oswalt grants that this would make the book "unique."²  The standpoint argument therefore shows that the early date requires a case without parallel.  It does not show that such a case is impossible.

The "former things"

The defenders are right that the phrase does not, by itself, fix a date.  In 48:3 God says, "I announced events beforehand ... suddenly I acted and they came to pass."  The verses then announce "new events that are previously unrevealed."³  Any prophet at any date could say that God's earlier words have come true and that new ones are now given.  On Oswalt's and Schoors's reading, that is all the passages say.²

The critics' argument works only if the "former things" include Cyrus's first victories, as North held.¹⁴  The text gives some support for that.  In 41:2-3 the conqueror "pursues them and passes by unharmed;  he advances with great speed."³  In 41:25 he "advances."³  The trial speeches then ask the idols who "decreed" this beforehand.  If the conqueror's victories are offered as the proof that God foretold them, the proof must already be visible to the hearers.  Childs, who accepted the exilic date, still thought the critics had not found the earlier prophecy about Cyrus that this reading needs, and read the "former things" as the prophecies of First Isaiah.¹³  So the argument leans to a late date without settling it.  It does not assume that prediction is impossible.  As Williamson says, it "presupposes an acceptance of the validity of predictive prophecy."⁶

Style and vocabulary

Style is the one argument that has been measured, and the measurements disagree.  Yehuda Radday, who began as a believer in one author,²³ ran a battery of computer tests on the Hebrew and concluded that "the most dissimilar parts are chaps. 1-12 and chaps. 40-48.  Since Isaiah ben Amoz was beyond any doubt the author of the first, he cannot have written the latter."²⁴  L. LaMar Adams and Alvin Rencher of Brigham Young University compared the two halves with eleven other Old Testament books and found them "more similar to each other in style than to any of the control group," which they took as "a literary unity characteristic of a single author."²⁵

The two studies answer different questions.  Adams and Rencher showed that the halves are closer to each other than either is to Amos or Ezekiel.  A disciple steeped in Isaiah's language, or an editor who bound the halves together as Williamson proposes, would give the same result.  Radday showed that the halves differ from each other more than the parts within each half do.  One author writing decades apart, on new subjects, might give that result too.  Radday himself said his work "never even tried to refute the reasoning of the divisionist school from historical and ideological grounds nor to reply to the incisive questions of the unitarian one."²⁴  Harrison's protest, that critics first learn an author's style from his book and then use it to take parts of the book away from him, has some force against the older word lists, but none against Radday's controlled comparisons.¹⁶  No later study that could tell one author from a school of disciples was found for this article.  Style leans toward more than one writer, and that is all it does.  It assumes nothing about prophecy.

Theology and setting

Driver's argument from the fuller theology of 40-66 is the weakest on the critical side, because it dates a text by a theory of how Israel's ideas developed.  That theory is argued in Did Israelite Monotheism Evolve?, and it cannot be assumed here.  Harrison's argument from the old Canaanite idolatry is the weakest on the other side.  Exiles did not leave their habits behind, and Babylon's own gods appear by name:  "Bel kneels down, Nebo bends low."³  The arguments from Babylonian detail, made by Motyer and Kitchen, bear on where the writer lived, not when.  Critics disagree among themselves on the place.  Duhm put the prophet in Phoenicia, and Torrey in Palestine.²⁶¹⁰

Chapters 36-39

Both sides agree that chapters 36-39 were arranged to lead into chapter 40.  Their last chapter is out of time order.  The envoys of Merodach-baladan in chapter 39 came while he held Babylon, and Oswalt, a defender of one author, dates that visit to about 712/11 BC, about ten years before Sennacherib's siege told in chapters 36-37.²¹  The book puts it last so that Isaiah's warning, that Hezekiah's treasures and sons "will be carried away to Babylon,"³ stands just before "Comfort, comfort my people."  Young called the section "the bridge which connects chapters 1-35 and 40-66."¹⁵  The arrangement is evidence of one book.  It is not evidence of one author, since an editor would arrange it the same way.

Summary of the arguments

Oswalt's charge hits some critics.  He shows that Duhm took the title of Isaiah 14 as "proof of a postexilic date" once skeptics doubted prediction.²¹  But the arguments that carry the late date, the standpoint and the "former things," keep their force for a reader who believes in prediction.  Gesenius, Driver, Williamson and Blenkinsopp all say so in their own words, and the logic of the arguments bears them out.

Part 5: What the outside evidence can and cannot show

The Great Isaiah Scroll

The oldest copy of Isaiah 40-66 is the Great Isaiah Scroll, 1QIsaᵃ, found in Cave 1 at Qumran in 1947.  Two radiocarbon laboratories dated the same column of it.  Zurich found 2128 ± 38 radiocarbon years and Arizona 2141 ± 32, "in excellent agreement."  Arizona's calibrated range at two standard deviations is 356-291 BC (24% of the probability) or 250-103 BC (76%), and the handwriting experts' date is 150-125 BC.²⁷  A 2025 study that combined radiocarbon with a machine-learning model of handwriting found that it "dates the two halves consistently between 180–100 BCE."²⁸

So the scroll was copied about 400 years after Cyrus took Babylon in 539 BC, and about 550 years after Isaiah's last dated oracle.  It contains all sixty-six chapters, and it runs from chapter 39 into chapter 40 without a break.¹⁵  That proves the book was one book by the second century BC.  No one disputes this.  It cannot show who wrote any part of the book four centuries earlier.  Its one large gap, after chapter 33, falls where Williamson thinks the later prophet joined his work to Isaiah's,⁶ and Kitchen reads it as the midpoint of a book in two halves.¹⁸  Neither reading dates chapter 40.

Ben Sira

Jesus ben Sira wrote in Hebrew about 180 BC.  His praise of Isaiah says, in the Greek translation his grandson made, that "he comforted those who mourned in Sion" and "showed the things that will be and the hidden things before they even happened."²⁰  The words are close to Isaiah's own.  The Greek has "παρεκάλεσεν τοὺς πενθοῦντας ἐν Σιων," he comforted those mourning in Zion, and the Greek of Isaiah 61:2-3 has "παρακαλέσαι πάντας τοὺς πενθοῦντας ... τοῖς πενθοῦσιν Σιων," to comfort all who mourn, to those who mourn in Zion.²⁹  The Hebrew text of Ben Sira, preserved in a medieval copy from the Cairo Geniza, has "וינחם אבלי ציון," and he comforted the mourners of Zion, which takes up the Hebrew of Isaiah 61:2-3, "לנחם כל אבלים" and "לאבלי ציון."³⁰³

The phrase comes from chapter 61, in the part critics assign to Third Isaiah.  So by about 180 BC the whole book, including its latest chapters, was read as Isaiah's, and the prophet was praised precisely for foretelling.  That is about 360 years after Cyrus.  Like the scroll, it shows what readers believed, not who wrote the chapters.

Josephus and Cyrus

The Jewish historian Josephus, writing about AD 94, says that Cyrus freed the Jews because he read about himself in Isaiah:  "These things Cyrus knew from reading the book of prophecy which Isaiah had left behind two hundred and ten years earlier."  When Cyrus read the prophecy, "he wondered at the divine power and was seized by a strong desire and ambition to do what had been written."³¹  If the story were true, it would show that a book of Isaiah naming Cyrus existed by 538 BC.  It would not show that it existed in 700 BC, since a prophet of the 540s could have written it for Cyrus to read.

There are three reasons to doubt the story as history.  First, it is late:  Josephus wrote about 630 years after the event.  Second, his sources do not have it.  Ezra 1:1 says Cyrus acted "in fulfillment of the Lord’s message spoken through Jeremiah," not Isaiah,³ and Ralph Marcus, Josephus's Loeb translator, notes that "There is no reference to the prophets in 1 Esdras = Ezra," the Greek book Josephus was following.³¹  Third, Josephus's numbers are his own reckoning.  Two hundred and ten years before Cyrus's first year (538 BC) is 748 BC, and his "one hundred and forty years before the temple was demolished" is about 727 BC.  The two differ by exactly the seventy years of exile that Josephus counts between the fall of Jerusalem and Cyrus, though the real interval was about forty-eight years.³²  Marcus also notes that the 140-year figure does not fit Josephus's own regnal dates, which give 125 years.³¹

Ezra 1:2 does have Cyrus say that God "has appointed me to build a temple for him in Jerusalem,"³ which is what Isaiah 44:28 says Cyrus would do.  Josephus may have inferred his story from that likeness.  The story is evidence of how first-century Jews read Isaiah, and nothing more.

The New Testament

Twenty-two verses of the New Testament name Isaiah.  Twelve of them introduce words from chapters 40-66, and ten introduce words from chapters 1-39.³³  The writers plainly knew one book called Isaiah.  How much that settles depends on the reader.

As history, the citations come from the first century AD, later than Ben Sira and the Great Isaiah Scroll, and they show the same thing:  how the book was read.  They also show that the name could be used loosely.  Mark 1:2-3 introduces a quotation of Malachi 3:1 followed by Isaiah 40:3 with the words "As it is written in the prophet Isaiah."³  That is how a writer names a scroll or its leading prophet, not how he assigns every line.

As theology, the strongest case is John 12:38-41.  It quotes Isaiah 53:1, then Isaiah 6:10, and then says, "Isaiah said these things because he saw Christ’s glory and spoke about him."³  Archer argued that if chapters 6 and 53 had different authors, "then the inspired apostle himself must have been in error."¹⁹  A reader who takes John's words as a statement about authorship must hold that one Isaiah wrote both.  A reader who takes "Isaiah" as the name of the book, as Mark 1:2 suggests the evangelists could, need not.  This is a question about what the New Testament writers meant to teach, and it cannot be settled by the historical evidence in this article.

The Cyrus Cylinder

The clay cylinder on which Cyrus announced his capture of Babylon uses some of the same language as Isaiah 44-45.  In Irving Finkel's translation for the British Museum, Marduk went "seeking for the upright king of his choice.  He took the hand of Cyrus, king of the city of Anshan, and called him by his name."  Marduk "walked at his side," and made him "enter without fighting or battle" into Babylon.³⁴  Isaiah's God speaks to "Cyrus, whose right hand I hold," says "I call you by name," calls him "shepherd," and promises "I will go before you."³³

The likeness is real, and it cuts no way on the date.  The cylinder was written after 539 BC, so Isaiah 45 cannot have borrowed from it on any date.  Both use the old language of Mesopotamian kingship.  Lisbeth Fried notes that the cylinder's royal titles are "that of the Assyrian kings who ruled Babylon, Ashurbanipal in particular,"³⁵ so the idiom was current in Isaiah's own century.  Fried, who accepts the exilic date, reads Second Isaiah as doing for the Jews what the Marduk priests did for Babylon:  "Deutero-Isaiah helped to install the Persian conqueror as the legitimate Judean king."³⁵  Her reading fits the late date, but it assumes it.  The cylinder does supply one contrast.  It credits Marduk, and Isaiah's God says to Cyrus, "I call you by name ... even though you do not submit to me."³  Cyrus never acknowledged the God of Israel, and the text says he would not.

Part 6: What each answer gains and costs

For the argument from prophecy

The date of chapters 40-55 changes the lead time of their predictions, and lead time is what Test 1 measures.  The table prices each main prediction on the two dates.  The early date is taken as about 700 BC.  The late date is taken as the 540s BC, between Cyrus's victory over the Medes in 550 BC and Babylon's fall in 539.³⁶

On the late date, Cyrus by name drops out of the evidence entirely, as Prophecies About Nations and Cities concluded.  The fall of Babylon and the return remain forecasts, made a few years ahead by someone watching Cyrus win.  They were good forecasts, but a well-informed observer in the 540s could have made them.  Several details also failed on either date.  Babylon opened its gates to Cyrus without a battle, and Cyrus says he returned the gods Nabonidus had gathered into the city to their own shrines, and he asks those gods to pray for him "every day before Bel and Nabu."  Isaiah 46 has Bel and Nebo themselves "head off into captivity."³⁴³  The nations article scores those clauses.

The Servant Song of Isaiah 52:13-53:12 is not affected.  It was copied in the Great Isaiah Scroll between about 180 and 100 BC, more than a century before Jesus was born, whichever theory of its composition is true.²⁸  So the question of this article makes no difference to Isaiah 53's Test 1 against the Gospels.  Whether it passes the other tests is argued in Isaiah 53.  The same holds for every use the New Testament makes of chapters 40-66:  all of them were in a pre-Christian scroll.

For a conservative reader

If one Isaiah wrote the book, the reader gains the most striking named prediction in the Bible, a king named about a century and a half ahead, and keeps the plainest reading of the book's title and of John 12:41.  The costs are two.  First, the reader must defend a kind of prophecy with no clear parallel, a prophet speaking for sixteen chapters to people not yet born, as Oswalt grants.²  To a skeptic that looks like special pleading, because it is the one case where the usual rule is set aside, and it is set aside in the case that matters most.  Second, the grand promises of the return, which outran what happened, as Gesenius and Williamson note, are then promises Isaiah made about a future he saw clearly in other respects.¹

If a prophet of the exile wrote 40-55, the reader loses Cyrus as evidence and gains a prophet who stands where his words make most sense, and whose "former things" argument is a working example of the argument from prophecy:  a prophet pointing his hearers to oracles that had come true in their own time.  The critics quoted in Part 2 accept that this prophet predicted Babylon's fall and the return before they happened.¹¹  Nothing in the late date makes the book a forgery.  Chapters 40-66 never claim to be by Isaiah, and the book's heading in 1:1 can be read as the title of a collection that grew around his name, as other prophetic books did.  The cost is a tension with how the New Testament names its quotations, which each reader must weigh by his own view of what those writers meant to teach.

Neither verdict touches the core of the Christian argument from prophecy, which rests on texts copied before the events they are said to foretell.  What the late date removes is a single showpiece, Cyrus by name, that never could be proved to a skeptic anyway, because no manuscript older than 539 BC exists.

Objections and responses

1. "Critics date Isaiah 40-66 late only because they rule out prediction."  Some did.  Oswalt shows that Duhm took the title of Isaiah 14 as "proof of a postexilic date" once skeptics doubted prediction.²¹  But the main arguments do not rest on that premise.  Gesenius, who founded the modern consensus, said that even granting Isaiah supernatural revelation would not remove the difficulty of the standpoint.¹  Driver said the late date does not make the chapters a prophecy after the event,¹¹ and Radday began as a defender of one Isaiah and was moved by word counts.²³  What the standpoint argument does assume is a rule about how prophets speak.  A believer may reject the rule, and the rule is partly circular, as Part 4 shows.  That is the right place for the debate.

2. "Jesus and the apostles name Isaiah as the author, so the question is closed."  For a reader who takes the New Testament's way of naming quotations as a teaching about authorship, that is a real reason, and John 12:38-41 is its strongest form.³¹⁹  Other believers read "Isaiah" there as the name of the book, as Mark 1:2 uses it for a quotation that begins with Malachi.³  As historical evidence, the citations are from the first century AD and add nothing to what Ben Sira and the Great Isaiah Scroll already show.

3. "The Great Isaiah Scroll runs straight from chapter 39 into chapter 40, so the book was always one."  It shows that by the second century BC the book was copied and read as one book, which no one disputes.  A copy made four centuries after the exile cannot show who wrote what before it.

4. "A ruler called Cyrus (Kurash) ruled in Parsua around 646 BC, so the name was not unthinkable in Isaiah's day."  This is Kitchen's argument:  "a still earlier “Cyrus” (Kurash) ruled in Parsua in 646."¹⁸  It removes an objection to the early date, but at a price.  If Cyrus was a name already borne by Iranian rulers, then naming one is less improbable, and the naming is weaker evidence of foreknowledge even if Isaiah wrote it.  The argument helps the early date and costs the prediction part of its force.

5. "If a great prophet wrote 40-55, his name could not have been forgotten."  Radday, a critic, granted that the defenders of one Isaiah ask this and that it is hard to answer:  "How could the name of the ‘Unknown’ prophet have become unknown?  If the name of Obadiah, the author of one single chapter of minor importance, was carefully preserved by tradition, a fortiori should the name of one be remembered who wrote a whole book."²⁴  The critics' reply is that collections gathered under one name:  the book of Proverbs opens as "The proverbs of Solomon" and still contains "The words of Agur" and "The words of King Lemuel,"³ and Zechariah 9-14 sits under Zechariah's name with headings of its own.³  On Williamson's model the later prophet wrote as a continuator of Isaiah's book, not as a rival author.  The question remains a fair one, and it is listed among the issues below.

6. "On the late date the book dresses a forecast as a prediction, which is deceit."  Oswalt:  "If this is an after-the-fact realization, what purpose is served by attempting to make it appear a prediction?"²¹  The critics quoted here do not hold that the chapters are after the fact.  They hold that the prophet foretold Babylon's fall and the return a few years ahead, while Cyrus was rising.  Nor do chapters 40-66 claim to be Isaiah's;  the claim comes from the book's heading and from later readers.  Where the chapters say God announced things "beforehand," the critics read this as earlier oracles that had come true, and Childs as the prophecies of First Isaiah.¹³  Neither reading involves a false claim.

7. "If Isaiah 40-55 is late, then all biblical prophecy is written after the event."  That does not follow.  The critics themselves hold that the chapters were written before the fall of Babylon they announce.  And the strongest cases for prediction, such as Isaiah 53 measured against the Gospels, were copied in pre-Christian scrolls whatever their date of composition.²⁸  The late date removes one showpiece, Cyrus by name.  It does not remove the category.

8. "The Cyrus Cylinder proves Isaiah copied Persian propaganda."  It cannot, because the cylinder was written after 539 BC, and the shared language of a god who takes a king by the hand and calls him by name was Mesopotamian royal idiom already in Ashurbanipal's time.³⁵  The likeness is evidence that the writer knew how Babylonian kings were spoken of, which a Judean of either century could have known.

9. "Josephus says Cyrus read Isaiah, so the prophecy existed before Cyrus."  Josephus wrote about 630 years later, his source in Ezra credits Jeremiah, not Isaiah, and his numbers are his own reckoning.³¹³  Even if true, the story would show only that the oracle existed by 538 BC, which the late date also allows.

10. "Computers have proved two authors" or "computers have proved one."  Neither.  Radday and Adams and Rencher used different features and asked different questions, and neither design can tell one author from a disciple steeped in his style.²⁴²⁵

Issues and unanswered questions

1. No copy of Isaiah 40-66 is older than the second century BC.  Until one is found, the date of composition rests on internal evidence alone, and Test 1 for Cyrus cannot be passed.

2. The standpoint rule has not been tested independently.  It is drawn from the prophetic books, and the passages that might break it, such as Isaiah 13-14, are dated late partly by the same rule.¹¹  A study that listed every prophetic passage addressing a distant generation in the second person, with dates fixed by other means, would test it.  None was found.

3. No modern stylometric study separates "one author" from "a school steeped in one author."  The two famous studies are from 1973 and 1974.  Later critiques of them, such as Portnoy and Petersen's in the Journal of Biblical Literature (1984), could not be read for this article, and no later computational study of Isaiah's authorship was found.

4. The case for a separate Third Isaiah is thin.  Duhm's step was "not argued at any length,"⁷ and Westermann accepted it on the ground of discipleship.¹²  Whether 56-66 come from one writer, several, or the prophet of 40-55 himself later in life is open.  Blenkinsopp's commentary on 56-66 could not be read.

5. The scroll's layout at 39/40 is known here only from Young (1958) and Kitchen (2003).  The official edition (Discoveries in the Judaean Desert 32, 2010) and the photographs were not checked.

6. Several primary sources were not read:  Döderlein's commentary, whose first date writers give differently;  Morton Smith's "II Isaiah and the Persians" (1963) and Amélie Kuhrt's study of the Cyrus Cylinder (1983), known here only through Fried;  Schoors's study of the "former things," known only through Oswalt;  and the manuscript of Ben Sira's Hebrew, read here in David Kahana's edited text.

7. Why the name of the later prophet was lost has no answer from the critical side that goes beyond analogy.

Verdict

Isaiah 40-55 most probably comes from a prophet who spoke among the exiles while Cyrus was rising, in the 540s BC.  Chapters 56-66 most probably come from the decades after the return, though the case for a separate writer there is weaker.  The arguments that carry this conclusion, the prophet's standpoint and his appeal to "former things" already fulfilled, do not assume that prediction is impossible.  Gesenius, Driver, Williamson and Blenkinsopp all say so, and the logic of the arguments bears them out.  But they are not proof.  The standpoint argument rests on a rule about how prophets speak, drawn partly from texts dated by the same rule, and a God who could name Cyrus in advance could also inspire a prophet to speak to people not yet born.  The single-author view is coherent, and its best defender says plainly that it makes the book unique.

The outside evidence cannot decide between them.  The Great Isaiah Scroll, Ben Sira, Josephus and the New Testament all show how the book was read from about 180 BC on, not who wrote it four centuries earlier.  The style statistics split.  The Cyrus Cylinder fits either date.

For the argument from prophecy, the result is narrow.  On the probable date, Cyrus by name drops out of the evidence, and the fall of Babylon and the return become accurate forecasts made a few years ahead.  On either date, the oracles were copied before Jesus was born, so Isaiah 53's case against the Gospels does not depend on this question at all.

Sources