Romans: Text, Readers, and the History of Interpretation

By Stephen Westerholm

A wide-ranging study of the interpretation of Paul’s letter to the Romans throughout history, from Origen to Karl Barth. 

In anticipation of his Illuminations commentary on Paul’s letter to the Romans, Stephen Westerholm offers this extensive survey of the reception history of Romans. After two initial chapters discussing the letter’s textual history and its first readers in Rome (a discussion carried out in dialogue with the Paul-within-Judaism stream of scholarship), Westerholm provides a thorough overview of over thirty of the most influential, noteworthy, and representative interpretations of Romans from nearly two thousand years of history. Interpreters surveyed include Origen, John Chrysostom, Augustine, Peter Abelard, Thomas Aquinas, Martin Luther, John Calvin, John Locke, Cotton Mather, John Wesley, and Karl Barth. 

Bearing in mind that Paul did not write for scholars, Westerholm includes in his study interpreters like Philipp Jakob Spener and Richard Baxter who addressed more popular audiences, as well as an appendix on a remarkable series of 372 sermons on Romans by beloved British preacher Martyn Lloyd-Jones. A further aim of the book is to illustrate the impact of this New Testament letter on Christian thought, supporting Westerholm’s claim that “the history of the interpretation of Romans is, in important areas and to a remarkable extent, the history of Christian theology.”

• Language: English
• Publisher: Eerdmans
• Release date: Sep 29, 2022
• ISBN: 9781467465045

Stephen Westerholm

 Stephen Westerholm is professor emeritus of early Christianity at McMaster University, Hamilton, Ontario. His other books include Reading Sacred Scripture: Voices from the History of Biblical Interpretation (with Martin Westerholm), Justification Reconsidered: Rethinking a Pauline Theme, and Understanding Paul: The Early Christian Worldview of the Letter to the Romans.

Book preview

Preface

Commentators on the Bible today are latecomers to a conversation that has been ongoing for centuries. We do well not to barge in at once with our own views. Common courtesy, a decent respect for the opinions of others, and the realization that wisdom was not born, and will not die, with us all demand that we inform ourselves of what others have been saying before we enter the conversation.

Part of the aim of [the Illuminations series of commentaries] is to alert the reader to the history of conversation surrounding biblical texts and to invite the reader to join that conversation.¹ It was, indeed, this distinguishing feature of the series that tempted me to enter the crowded field of commentators on Romans. The studies that follow in this book represent my efforts to inform myself of the earlier conversation before embarking on the commentary itself.

A commentary requires a text to interpret; I begin with the text and textual history of Romans. But even this study overlaps with the history of interpretation: already the earliest extant Greek manuscript of the epistle shows a scribe who interprets while copying the text. We all learned long ago that translation necessarily involves interpretation. If we did not already know, we here learn that, at least for some scribes, copying Romans did so as well.

What do we know about the first readers of Romans? Alas, we do not know how they understood or responded to the text. Still, since Paul wrote to a particular audience what he thought they needed to hear, anything we can learn about them may illuminate what he meant to say and how he wanted to be understood. Hence, discussions of Paul’s audience form a standard feature of commentaries on Romans. Rather than simply repeating standard fare, I have chosen to discuss Paul’s audience in conversation with a contemporary stream of Pauline interpretation, the Paul within Judaism school. Scholars associated with this school have a distinctive view of Paul’s intended readership, so that my interaction with them naturally addresses what needs to be said on that issue. But these scholars also raise important questions about the content of the epistle; these, too, I address.

The heart of the book is, of course, the history of interpretation, from Origen to Barth, found in chapter 3–7. Extensive as it is, the survey is necessarily limited to a selection from among the most influential, noteworthy, or representative treatments (commentaries, annotations, series of homilies, early English translations) of the epistle as a whole. Commentators later than Barth figure as dialogue partners in every recent academic commentary; they require no summary here.² The earlier history is often neglected; here it takes center stage.

Paul’s primary readership is not scholarly, nor was it intended to be. With that in mind, my preparation for the commentary includes consulting treatments of the epistle written for a popular audience: Orthodox, Catholic, and Protestants of several stripes. As one sample of such a reading, the appendix looks at a remarkable series of sermons treasured today in segments of Reformed Christianity, though little known elsewhere: the Doctor—Martyn Lloyd-Jones—looks at the apostle in fourteen volumes containing 372 sermons on Romans.

Translations of the biblical text are generally my own; but in the discussion of individual interpreters, I have retained the readings found in their work (in the relevant cases, the readings found in the English translations of their work).

Chapters 1, 2, and 7, and parts of chapters 5 and 6, were written at a time when, because of pandemic restrictions, I did not have physical access to a library. I am very grateful to Janice Adlington, head of collections and content strategy at McMaster University Library, for providing me, repeatedly, with electronic access to literature without which I could not have continued my work.

I am also very grateful to C. L. Seow, general editor of the Illuminations commentary series, for encouraging me to publish separately material originally designed for the introduction to the Romans commentary; to Trevor Thompson and Laurel Draper of Eerdmans for expert and friendly assistance at various stages in the submission of the manuscript and its preparation for publication; and to Ryan Davis for his excellent editorial work.

This book is dedicated to my grandchildren—of whom, at the time of writing, there are eight: Evelyn, Elijah, Sven, Abigail, Caleb, Hannah, Isaac, and Anders. A ninth, completing the starting lineup of our baseball team, is expected just in time for spring training. I would love to dedicate a book to each of them individually. But once their number exceeded two, it became clear that I could not keep up; the collective dedication will have to serve the purpose. If it seems strange that I should dedicate studies such as these to young people not yet ten years old, I can only say that, alas, I did not write The Little Engine That Could or Little Brown Bear (though I now know these classics nearly by heart), or (for those a little older) All-of-a-Kind Family or Treasures of the Snow. And progress on my proposed All I Really Need to Know I Learned at a Baseball Game has, I fear, been imperceptible. On the other hand, I am glad to point out an intimate connection between my grandchildren and the composition of this book: they have provided most of the diversion and fun I have had while working on the project. Whether they will ever read it is perhaps not very likely; better, in any case, that they should learn by heart (as Luther thought we all should) Paul’s Letter to the Romans. It will tell them how to play the game.

1. C. L. Seow, Scott C. Jones, Judith H. Newman, and Loren Stuckenbruck, To the Reader, in Job 1–21: Interpretation and Commentary, by C. L. Seow, Illuminations (Grand Rapids: Eerdmans, 2013), xii.

2. Such a summary would, in any case, overlap significantly with part 2 of my Perspectives Old and New on Paul: The Lutheran Paul and His Critics (Grand Rapids: Eerdmans, 2004).

Part 1

Text

CHAPTER 1

Text and Textual History

Of those who read Romans, only a tiny fraction concern themselves with its textual history. This is as it should be. Attention given to ancient manuscripts and their variant readings may be attention diverted from what a text is saying. To be sure, when Romans speaks, distraction may be welcome. One way of defend[ing] oneself against God’s Word, Søren Kierkegaard wryly noted, is to take Holy Scripture, lock your door—but then take ten dictionaries, twenty-five commentaries, then you can read it, just as calmly and coolly as you read newspaper advertising. And if it ever occurs to you to wonder whether you are, in fact, living as the text requires, the danger is still not very great. Look, perhaps there are several variations, and perhaps a new manuscript has just been found—good Lord!—and the prospect of new variations.¹ Point taken. But it remains true that millions who read Romans for its content unwittingly rely on the few whose efforts furnish the text they read. Their essential, preliminary, and (for some of us, at least) intriguing work merits attention here; for the moment, we postpone a look at what the letter says.

Translators and other scholars of the New Testament generally work with the Greek text of Novum Testamentum Graece (at the time of writing, in its twenty-eighth edition;² also referred to as the Nestle-Aland edition and, hence, NA²⁸) or The Greek New Testament published by the United Bible Societies (at the time of writing, in its fifth edition [UBS5, 2014]). The text of these editions is identical, although they cite differently the evidence on which the text is based. Neither edition reproduces the wording of any single ancient manuscript; rather, where the manuscripts disagree, the editors have chosen the reading that seems superior. Since some manuscripts provide the better reading at one place, and other manuscripts at another, the resulting text is referred to as eclectic. Students of Romans may well want to know the evidence on which, the grounds by which, and the degree of probability with which these decisions are made.

When Paul sent his letter to the Romans, in all likelihood he kept a copy for himself. Handwritten copies are seldom identical, so the text of the copy he retained may not have been exactly the same as that in the letter sent to Rome. Furthermore, it is not at all impossible—it has often been suggested—that the apostle sent copies to other churches, perhaps without the specific references to Rome in 1:7, 15; further copying can only have led to further variations in the text. By the end of the first century, the process of collecting Paul’s letters had begun; again, it is not at all impossible—it has also been suggested—that Paul himself initiated the first such collection, using retained copies of letters he had sent.³ But whether or not the initiative was his, the first collection, with its own inevitable divergences from the letters as Paul sent them, naturally became the archetype from which other copies were made.

In short, the history of the text of Romans is not simply the story of what happened when the letter Paul sent to Rome was copied and recopied. Variations found in our many manuscripts may, in some cases, go back to differences among versions of the letter originating from the apostle himself;⁴ still others were introduced when the letters were collected. From that point on, as copies of the letter multiplied, so, too, did variations in its text.

To keep things in perspective, we should add that the text of this two-thousand-year-old letter has proven remarkably stable: with one exception, differences between ancient copies pertain only to the wording of individual verses.⁵ Even in these cases, much evidence for the text of Romans is available to scholars on which to base judgment. Using that evidence (summarized below), they can generally identify errors introduced accidentally in the process of copying. Frequently, they can also distinguish earlier readings from attempts of scribes to improve the text that lay before them. Thus, where manuscripts differ, scholars often agree on the older form of the text. Nonetheless, at particular points, judgments diverge; and, at particular points, the relation between the oldest reading in surviving manuscripts and the text of the letter Paul sent to Rome—or one of the other apostolic forms of the letter!—remains a matter for (more or less educated) speculation.⁶

1. PAPYRUS 46

Around the year 200 of our era, a professional but blundering and not always attentive scribe produced a manuscript containing the letters of Paul.⁷ One could wish he—or possibly she⁸—had performed the task more professionally. And one can only regret that, although much of the codex has survived, much has been lost, particularly at its beginning and end. Papyrus 46 is nonetheless of extraordinary significance, representing perhaps the oldest known copy of the Greek text of our epistle, and certainly the oldest of any extent.⁹ A study of its nature and content reveals much about the early history of the Pauline text.

1.1. THE NATURE OF THE PAPYRUS

The earliest New Testament manuscripts were all papyri (i.e., written on material made from the papyrus plant). That P46 is a codex (resembling books as we know them) is itself significant. Jews used scrolls for their scriptures; indeed, in the Greco-Roman world at large, and until the fourth century, literature of any kind was predominantly written on scrolls. Yet, for reasons about which we can only speculate, the early Christians produced almost all their biblical manuscripts as codices rather than scrolls.¹⁰ (One plausible proposal suggests that the codex was first adopted because it allowed a collection of Paul’s letters to be contained in a single manuscript.)¹¹ Furthermore, our papyrus regularly uses abbreviations for such hallowed names (nomina sacra) as God, Lord, Jesus, and Christ: this feature, too, is distinctive of manuscripts in use among early Christians, and provides further, perhaps surprising, evidence of cohesion in their ranks.¹² The manuscript was probably prepared for public reading: although, as is customary with ancient Greek texts, words generally follow each other without intervening space, at a number of points, the scribe of P46 left space to mark a pause in sense; and there are other markings designed to facilitate reading as well.¹³

Fifty-two papyrus sheets were folded together to produce a codex with 104 leaves, written on both sides (thus 208 pages). Parts of 86 leaves have survived: 56 are now kept in the Chester Beatty Library in Dublin, 30 at the University of Michigan in Ann Arbor.¹⁴ The loss of seven leaves at the beginning resulted in the loss of the opening chapters of our epistle (Rom 1:1–5:16). The loss of leaves at the end means that we cannot be certain which of the letters attributed to Paul but missing here (2 Thessalonians, Philemon, the Pastoral Epistles) were originally contained in the manuscript.¹⁵ Twelve leaves (24 pages) contain the extant text of Romans. On the page containing Romans 5:17–6:4, the end of each line has been lost; on the other side of the same leaf, the beginning of each line is missing (6:5–14). Only isolated words, or even letters, have been preserved on the two sides of the leaf containing Romans 14:9–15:9. On the other hand, much has survived from 8:15 to 14:8 and from 15:11 to the end of the epistle. After 16:23—with which, in this manuscript, the letter abruptly ends—a rough calculation of the number of lines in the text (1,000) is recorded for the purpose of paying the scribe.¹⁶

1.2. VARIANT READINGS, UNINTENDED AND DESIGNED

His patron should have demanded a refund. A few examples must serve to illustrate the many scribal lapses that mar our manuscript. (The errors increased as our scribe worked his way through Romans;¹⁷ my impression is that he was liable to fits of exhaustion.¹⁸) At 6:14, the word for sin (ἁμαρτία) was left without its final letter. Conversely, at 8:38, the first two letters of οὔτε were mistakenly copied twice (a typical scribal error, known as dittography). The Greek letter omicron appears twice in succession (at the end of one word and the beginning of the next) in the text of 11:7; our scribe wrote the letter thrice. An omission of final sigma from οὓς at 8:30 resulted in the following bizarre text: But those whom [God] predestined, these he also called. And he did not call [ου καὶ ἐκάλεσεν], these he also justified. After our scribe had written the first word of 12:4, which ended with Greek rho (καθάπερ), his eye, returning from his own text to the one he was copying, fastened mistakenly on the rho at the end of the second word (γὰρ); thus, overlooking the second word, he proceeded to write the third (the scribal error of parablepsis [overlooking] due to homoioteleuton [similar ending]). A similar skipping at 16:15, this time from sigma (at the end of αὐτοῖς) to sigma (at the end of πάντας), resulted in the omission of πάντας (all) before saints.

At 11:29, he misread κλῆσις (calling) as κτίσις (creature)—however implausible, in its context, the resulting text. At 16:5, he—but he was not quite alone in this—mistook ἀπαρχὴ (firstfruits) for two words, ἀπ᾽ (i.e., ἀπό, from) and ἀρχή (beginning); realizing, however, that, after the preposition ἀπό, ἀρχή must be read as ἀρχῆς, he modified the text accordingly. The resulting text, again, is scarcely intelligible. A scribal blunder renders 15:13 meaningless as well.¹⁹

To be sure, accidental mistakes like these are easily recognized: other manuscripts, produced much later than P46, preserve the uncorrupted text that must have lain before our slovenly scribe. Thus, even the mistakes of our earliest Pauline manuscript provide evidence of a truer text from a still earlier date. Of greater significance is the recognition that, once such surface errors have been discounted, the basic text of our papyrus is of supreme quality;²⁰ "the Vorlage [parent text] which lay before the scribe of P46 preserved a text of perhaps unequalled quality.²¹ Clearly, in disputed cases, the evidence provided by our earliest manuscript weighs heavily in favor of readings it shares with other manuscripts; moreover, with surprising frequency it alone (or in combination with a very few others) among all extant witnesses preserves the true wording of the Pauline archetype.²² As examples of texts in which P46, with little other support, preserves what is likely the genuine reading, Michael Holmes cites 8:24 (ὃ γὰρ βλέπει τίς ἐλπίζει; lit., For what one sees, who hopes?) and 13:12 (ἀποβαλώμεθα; lit., let us throw off, a stronger verb than ἀποθώμεθα, the verb found in most manuscripts here and in similar exhortations [e.g., Eph 4:22, 25; Heb 12:1]).²³ In such cases, if we start with the reading of P46, we see that alternative readings found in other manuscripts represent the kind of improvements" that scribes typically introduced in dealing with difficult or unusual texts. Where one variant can be seen as an attempted improvement of another—in terms of its grammar, clarity, theology, and so on—the unimproved (more difficult) text is likely to be original.²⁴

Of no less interest are the many readings in our papyrus that are neither genuine nor accidental: here, it is P46 that shows attempted improvements. Some of these readings are unique to our manuscript: in these cases, our scribe was presumably the one who saw reason to alter the text. In other instances, readings unlikely to be genuine are found in other manuscripts as well: in these cases, the text was presumably altered at some point in the first or second century before being reproduced in P46 and elsewhere. Here, then, we may catch a glimpse of an early reader [or of early readers] of Romans actively and thoughtfully engaged with the content of the text being read.²⁵ In each instance, we see how, in limited but real ways, liberty was taken with the text of the epistle.

At times, the scribe of P46 (and, sometimes, other scribes as well) had little use for Paul’s rhetorical repetitions. The apostle’s If children, then heirs—heirs of God and fellow heirs with Christ is reduced, in P46, to If children, then heirs of God and fellow heirs with Christ (8:17). Here, the scribe’s eye did not simply skip from the first occurrence of heirs (κληρονόμοι) to the second, since he also omitted the particle μὲν that follows the second; as elsewhere, he deliberately eliminated redundancy. At 8:23, what appears to be the genuine text reads (literally), And not only [the creation], but also ourselves [καὶ αὐτοὶ], who have the firstfruits of the Spirit, we also ourselves [ἡμεῖς καὶ αὐτοὶ] groan in ourselves; P46 omits the first also ourselves. At 12:14, Paul exhorted the Romans, "Bless those who persecute [you]; bless and do not curse. One blessing was quite enough for our scribe: Bless those who persecute and do not curse. Paul took up, at the beginning of 15:27, the same words with which he began 15:26 (they were pleased"); here, P46 is not alone in omitting the repeated words.

Elsewhere, our scribe, alone or with others, substituted what one might have expected Paul to write for what he appears to have written. (Again, the text more likely to be original is that which invited scribal improvements.) Κακός (evil) is the usual opposite of ἀγαθός (good; so, e.g., at 2:9–10; 16:19); at 9:11, P46 and others substitute it ([good or] evil [κακόν]) for the unusual φαῦλον ([good or] base). In the quotation from Isaiah in 9:27, P46 (with most other witnesses) replaces Paul’s ὑπόλειμμα (remnant) with the synonymous κατάλ[ε]ιμμα, preferred because it is the term used in the quoted (Septuagint) text.²⁶ Where Paul wrote, For Isaiah says (10:16), P46 reads, As it is written in Isaiah; as it is written is Paul’s more frequent way of introducing a quotation. Since the verbs in 11:14 are in the future tense, the scribe of P46 (and others) thought the same should be true of the verb that concludes v. 13: I glorify (δοξάζω) becomes I will glorify (δοξάσω). At 1:13, Paul wrote that he had often (πολλάκις) been minded to come to Rome; in a similar context (15:22), P46 and others repeat the adverb, replacing Paul’s unusual τὰ πολλὰ (lit., the many [times]). With pedantic truth, P46 and others correct Paul’s I am going to Jerusalem, serving [διακονῶν] the saints [there] (15:25) to I am going to Jerusalem in order to serve [διακονῆσαι] the saints [there].

In places, the alterations found in P46, alone or with others, were presumably meant to simplify the text. At 6:12, Paul wrote, Do not let sin rule in your mortal body, so that you obey its [i.e., the body’s] desires. Papyrus 46 and others read, Do not let sin rule in your mortal body so that you obey it [sin]. And why bother saying It is necessary to submit [ἀνάγκη ὑποτάσσεσθαι] when one can simply say, Submit [ὑποτάσσεσθε]! (13:5; so P46 and others)? Paul continues, in 15:20, the sentence he began in 15:18: … thus aspiring to proclaim the gospel where Christ has not been named. He might, more simply, have started a new sentence—and so, in P46 and other witnesses, he does: Thus I aspire to proclaim the gospel where Christ has not been named.

In general, the scribe of P46 was more inclined to remove a perceived difficulty than to add something by way of explanation;²⁷ but at 8:34, he added ἅμα δὲ (together) before Christ Jesus, perhaps wanting to make clear that it is not only God who justifies and does not condemn, but also Christ Jesus.

Then there are times when our scribe, or one of his predecessors, having read the text with a woodenness foreign to the apostle, concluded, That’s not quite what we want to say, now, is it? At 8:17, Paul wrote, … if we suffer with [Christ], in order that we may also be glorified with [him]; the first verb could also be translated, suffer the same thing as. Presumably, it was the thought that we do not suffer just as Christ did that led the scribe of P46 to rework the text: … if we suffer, in order that we may be glorified with [Christ]. According to 8:15, believers have received "the spirit of adoption; on the surface, this contradicts 8:23, where believers are said to groan while eagerly awaiting the [still future] adoption, the redemption of our body. P46 was not alone in resolving the contradiction by deleting adoption from the latter verse: eagerly awaiting the redemption of our [body]. (In P46, the word body" is missing from the mutilated beginning of the line, but must at one time have been present.)

Here, mention may be made of perhaps the best-known variant in Romans: at 5:1, did Paul write, We have [ἔχομεν] peace with God, as the majority of commentators believe, or Let us have [ἔχωμεν] peace with God, as the majority of manuscripts read? While this particular verse is not preserved in our papyrus, P46 does demonstrate that short o (omicron) and long ō (omega) were routinely confused at the time it was written. (In fact, the confusion began already in the pre-Christian era, as the two Greek vowels became indistinguishable in pronunciation.)²⁸ In our manuscript, omega appears for omicron in 6:2; omicron for omega in 9:17, 29; 10:14; and 11:25 (in 9:17, omicron was first written, then corrected to omega). The interchange of the two vowels is so common that, where manuscripts disagree whether o or ō should be read, the text critic is at times reduced to considering which variant makes better sense in the context.²⁹

Readers may well be pardoned if they find the above list of divergent readings underwhelming. Though even minor variations can be of significance for those who study the history of the text, students of Romans may justifiably conclude that, for our understanding of the epistle, little is at stake in any of these cases; and although many more readings in P46 could be mentioned, with one exception (to be discussed below), none of the variants passed over is of noticeably greater importance than those here treated.

And yet the point is worth making: there are many different readings—and not all of the secondary readings were accidental in their origin. At times, scribes may have been attempting to restore—as Marcion attempted to restore—the true reading of a sacred text they thought had been corrupted;³⁰ the effective result, even then, has been further variation. Michael Holmes sums up the situation by saying that, if we look at the bigger picture (the ‘macro’ level of the sentence, paragraph, and letter), we find stability in the transmission of the text; if we look at the details (the ‘micro level’ of the word, clause, and phrase), we find fluidity.³¹ The evidence of P46 illustrates the fluidity common to a number of manuscripts in the early period.³² To be sure, some copyists even then understood their task (as some translators understand theirs) to require word-for-word faithfulness;³³ but, in the early period at least, other copyists exercised a measure of freedom (as some translators paraphrase) to better convey the text’s message. On the one hand, then, the epistle was acknowledged to be of great importance, warranting repeated reproduction. On the other hand, some copyists clearly felt that the word of the Lord (or of his apostle) was not tied to a particular wording. The same conviction has encouraged Christians, unlike Muslims, to translate their Scriptures into other languages, from the second century to this day.³⁴

But more can be learned from a look at variant readings in P46. To illustrate the kinds of readings found in P46, there was no need to itemize the manuscripts with which our papyrus agreed (when it agreed with any). When, however, comparisons are made,³⁵ P46 is found most frequently to align with manuscripts traditionally labeled Alexandrian.³⁶ Nonetheless, not infrequently it shares readings with manuscripts labeled Western rather than Alexandrian (e.g., from the examples given above, at 6:12; 8:23; 13:5, 12; 15:25, 27). (More will be said about these labels and manuscripts below.) From these facts, differing conclusions have been drawn. The evidence of P46 suggests to some scholars that, at the time it was written, these different textual traditions had yet to take shape; and when they did, they distinguished themselves by selecting differently from among the reservoir of readings current in the second century.³⁷ Alternatively, if the beginnings of different textual traditions do go back to the second century,³⁸ then P46, though basically Alexandrian, shows influence (or contamination) from elsewhere as well. The debate between these alternatives is ongoing.³⁹

1.3. PAPYRUS 46 AND THE END OF THE EPISTLE

What, finally, does the evidence provided by P46 contribute to an understanding of the one major textual issue related to Romans, the ending of the epistle—according to Kurt Aland,⁴⁰ the most difficult problem facing New Testament textual criticism? As noted above, in P46, the letter ends abruptly at 16:23. This seems a most unlikely ending to the letter, though perhaps not impossible as a postscript attached to the main body of the letter. And, indeed, the doxology with which Romans concludes our Bibles today (16:25–27) is placed, in P46, at the end of chapter 15, giving the text that follows (16:1–23) the appearance of a postscript.

The content of chapter 16, with its many greetings, raises its own questions. How could Paul have known so many believers in Rome when, as 1:10–15; 15:22–23 make clear, his mission had not yet taken him to Rome? Furthermore, in 16:3–5, Paul greets Prisca (Priscilla) and Aquila and the church in their house; but according to Acts, they left Rome before Paul met them, moving first to Corinth, then to Ephesus (Acts 18:1–3, 18–19, 24–26; see also 2 Tim 4:19). And in sending greetings to Epaenetus, Paul explicitly identifies him as the first convert in Asia (Rom 16:5). Putting these clues together, T. W. Manson proposed that Paul sent the first fifteen chapters of our Letter to the Romans (thus without the greetings of chapter 16), whereas he sent the sixteen-chapter form to the church in Ephesus, with greetings attached to members he knew well.⁴¹ In this way, Paul himself was the source of both the fifteen- and the sixteen-chapter form of the letter.

Although Manson’s proposal initially met with enthusiasm, it is generally agreed today that Romans 16 was written to believers in Rome.⁴² In no letter to a church he had founded did Paul include anything like the list of greetings in this chapter; indeed, in writing to such churches, it would hardly have been politic for him to greet some members while (necessarily) excluding others. Conversely, in writing to a community he had never visited, he would want to establish as many points of contact as possible. He need not personally have met all those whose names he mentions; some might be leaders in the community about whom he had learned from Prisca and Aquila. In any case, it is not unlikely that a number of people whom Paul had met in his travels now found themselves in Rome: there was much traffic to Rome in the first century, as later. Prisca and Aquila, at least, would merely have returned to the city from which they had been driven (Acts 18:1–2); the same might be true of others in Paul’s list as well.

There is no reason to doubt, then, that Paul’s Letter to the Romans included chapter 16 (with or without the doxology of 16:25–27). On the other hand, it seems likely that an ancestor of P46 contained only fifteen chapters, with the doxology serving as its conclusion. The scribe of P46 (or of one of its more proximate ancestors) copied this manuscript; but, finding a different manuscript that included 16:1–23, he added these verses after the doxology. The result is the strange text of P46: Romans 1:1–15:33 + 16:25–27 + 16:1–23.

Papyrus 46 suggests that Romans once existed in a fifteen-chapter form; other evidence indicates that a fourteen-chapter form was current in the second century as well.⁴³ Indeed, in most manuscripts containing all sixteen chapters, the doxology (16:25–27) appears immediately after 14:23. The same explanation might well apply to them as to P46: the doxology was placed after 14:23 to give a satisfactory conclusion to an abbreviated (in this case, fourteen-chapter) form of the letter; when comparison was made with sixteen-chapter manuscripts, chapters 15 and 16 were added after the doxology.

We can only speculate about the process by which a sixteen-chapter letter was reduced to the shortened forms. Perhaps Marcion, the usual suspect, removed the final two chapters from his Apostolikon;⁴⁴ if so, however, it is strange that there is broad second-century evidence for the omission of these chapters, whereas other omissions attributed to Marcion have had minimal impact on the textual tradition;⁴⁵ moreover, the fifteen-chapter version of the letter is left unexplained. Perhaps Paul⁴⁶ or a later editor,⁴⁷ desiring a broader circulation for the letter, decided to publish a version without those parts that pertained peculiarly to Rome. This would account for the omission of chapter 16, and of 15:14–33 as well; it is not clear, however, why 15:1–13 would have been omitted from the fourteen-chapter version, nor (again) why there should have been both a fourteen- and a fifteen-chapter form of the letter in circulation. Perhaps the ending of a particular copy was mutilated and the damaged text became the ancestor of other manuscripts.⁴⁸ But the problem remains: Were there two such damaged manuscripts, mutilated at different points, which nonetheless served as parent texts for other manuscripts?

And the doxology? Everyone agrees that it repeats important themes from earlier parts of the epistle;⁴⁹ but among the letters attributed to Paul, the language of the doxology most closely resembles that of Ephesians. Hence, whereas some scholars see no difficulty in ascribing the doxology as well as Ephesians to Paul,⁵⁰ others find Pauline authorship dubious in both cases.⁵¹ Here, we confine our attention to the text-critical issues.

If Paul composed the doxology, he naturally intended it for the end of the (sixteen-chapter) letter, where, indeed, it is found in important early witnesses.⁵² If chapter 15, or chapters 14 and 15, were then deliberately omitted, presumably those responsible for the omission saw fit to attach the concluding doxology to their abbreviated forms of the text. On the other hand, if someone other than the apostle added the doxology to the letter, Paul presumably ended his epistle at 16:23.⁵³ On this hypothesis, the Western tradition, finding 16:23 an unsatisfactory ending, moved 16:20 (with slight variations in wording) to a position after v. 23, providing a more suitable conclusion to the letter (16:24). Other (Alexandrian) manuscripts, equally dissatisfied with 16:23 as an ending, but retaining 16:20 in its place, added the doxology (vv. 25–27). Perhaps the doxology was thus composed (on this hypothesis, by someone other than Paul) to conclude chapter 16; alternatively, it may have been composed as an ending for the fourteen-chapter form of the letter (in most manuscripts, it appears after 14:23), then moved to the end of fifteen- and sixteen-chapter forms as well.

And, typically, in manuscripts of the Byzantine tradition, nothing is lost: in some, the two—originally distinct—attempts to end the letter on a more satisfactory note than 16:23 are both preserved (v. 24 and vv. 25–27). (Note that the King James Version, translating a Byzantine text, includes both of these endings.) Nor is that all. As we have seen, Western manuscripts moved v. 20 to a position after v. 23; Byzantine manuscripts kept the verse at v. 20 but added v. 24 as well. (Again, the King James Version has both.) Furthermore, Alexandrian manuscripts placed the doxology (16:25–27) directly after v. 23. In most Byzantine manuscripts, the doxology appears at the end of chapter fourteen; but a few of the Byzantines kept it both there and at the end of chapter sixteen.⁵⁴ Repeatedly, listeners to the letter in this form find themselves on the verge of saying Amen, as listeners to Beethoven’s best-known symphony are on the verge of applauding, only to discover that, no, after all, there are more endings to come.

2. OTHER EARLY GREEK MANUSCRIPTS

Papyrus 46 is one of the most important manuscripts of the New Testament.⁵⁵ Other Greek manuscripts dated prior to the fourth century are much more fragmentary: P27, P40, P113, P118, and (on parchment [treated animal skin] rather than papyrus) 0220. Papyrus 27 (now in the Cambridge University library) was once part of a codex (like P46 and all the manuscripts discussed below), since, although only a single leaf survives, it contains writing on both sides: scrolls, read as they were unrolled, were written on one side only. Very fragmentary portions of Romans 8:12–22, 24–27 are found on one side of the leaf; with each line, the beginning is lost, and only (parts of) one to four words are found at the end. On the other side, parts of 8:33–39; 9:1–3, 5–9 are preserved, with the end of each line lost and only (parts of) one to four words found at the beginning. As with P46, various nomina sacra (God, Lord, Christ, and Spirit) are abbreviated. Enough of the manuscript has been preserved to allow the judgment that its text is basically Alexandrian,⁵⁶ an earlier representative of that tradition.⁵⁷ There is one singular reading. At 8:21, instead of saying that creation "will be set free from its bondage to corruption, our manuscript reads has been set free (the final letters of the aorist passive form have been preserved); is the point that what will certainly take place in the future may be regarded as (as good as) done already? On the positive side, Royse notes that P27 does not share several errors of P46.⁵⁸ Nor does it share P46’s free approach to the text: the Alands classify it as a strict manuscript—that is, one that reproduced the text of its exemplar with greater fidelity."⁵⁹

Papyrus 40 (kept at the University of Heidelberg) contains seven fragments from four leaves, showing parts of the following passages: Romans 1:24–27, 31–2:3; 3:21–4:8; 6:4–5, 16; 9:16–17, 27. On several fragments, only (parts of) individual words are preserved; somewhat more can be read of 3:21–24 and 4:2–6—and it is here that problems abound. Three omissions of more than one word can be attributed to parablepsis: the scribe’s eye skipped from πίστεως (faith) at the end of 3:30 to πίστεως in the middle of 3:31, so that he omitted the intervening words; from Abraham in 4:1 to Abraham in v. 2; from δικαιοσύνην (righteousness) at the end of 4:5 to the same word near the end of v. 6. On the other hand, words at the beginning of 4:4 are written twice (dittography). Apart from such singular readings, P40 agrees almost throughout with Vaticanus (B), so that its place is with the Alexandrian tradition.⁶⁰ The Alands classify it as a free text,⁶¹ somewhat misleadingly, since its singularities are due rather to scribal carelessness than to liberties deliberately taken with the text.

Papyrus 113 (Ashmolean Library, Oxford) is a small fragment of a single leaf. Parts of four lines can be seen on one side of the fragment, each line containing from three to six decipherable letters; on the other side, parts of five lines are visible, each with two to four letters. Amazingly, this is sufficient for the text of the former side to be identified as coming from Romans 2:12–13; that of the latter side, from 2:29. This limited evidence shows a text identical with that of all the earlier and weightier witnesses [both Alexandrian and Western].⁶²

Papyrus 118 (University of Cologne) is made up of four small fragments from a single leaf, written on both sides. Parts of Romans 15:26–27, 32–33; 16:1, 4–7, 11–12 are preserved. The text (such as it is) is generally Alexandrian. Apart from two minor scribal errors, it never departs from more than one of the three great Alexandrian manuscripts discussed below: Vaticanus, Sinaiticus, and Alexandrinus.⁶³

Papyrus 10 (Harvard University), from the fourth century, is a single papyrus sheet on which Romans 1:1–7 has been written, apparently as a schoolboy’s exercise.⁶⁴ Its numerous errors are of no textual significance. On the other hand, its minority reading Christ Jesus in 1:1, supported by Vaticanus, may well represent the original text (it is adopted in NA²⁸).

Later papyri containing short portions of Romans need not be discussed here: P26, from around 600; P31, seventh century; P61, around 700; P94, fifth/sixth century; and P99, around 400, but containing only the first verse of Romans. Manuscripts written on parchment in capital letters (majuscules, or uncials) are later than the earliest papyri; still, the oldest majuscule on Romans (0220, Museum of the Bible, Washington, DC), dates from the third century. A single leaf, it contains parts of Romans 4:23–5:3 on one side, parts of 5:8–13 on the other. In publishing the fragment, Hatch identified its text as Alexandrian, noting that, where variants occur, 0220 agrees with Vaticanus more than with any other manuscript.⁶⁵ The Alands classify it as a strict manuscript.⁶⁶ Of particular note is that this, our earliest manuscript containing Romans 5:1, reads ἔχομεν (we have peace) in the indicative mood; and at 5:2, it (like Vaticanus) omits the words by faith, found in most manuscripts.

3. MANUSCRIPT GROUPINGS

Accidental errors—for example, the unintended omission, or duplication, of one or more words, or even of entire lines—are bound to occur whenever a text is copied by hand. Since the resultant text generally yields little sense, such mistakes are easily recognized and corrected, often by the copyists themselves. Mistakes they leave unaltered may be corrected by later readers, either by an intelligent guess or by consulting a different manuscript. When the manuscript on which the mistake was made is corrected, scholars use an asterisk to distinguish the original reading (e.g., D*) from the corrected text (e.g., Dc). Where more than one reader has corrected a manuscript, superscripted numerals (D¹, D², D³) distinguish the correctors.

Texts that have been deliberately altered are a different matter. Not only does the new reading make sense; to the scribe who introduced it, the result was a superior text—and later copyists would see no obvious reason to alter it. Hence, variant readings of this sort were generally copied, in good faith, for centuries. Textual critics have long grouped together manuscripts on the basis of the variants they share—though it should be said at once that the boundaries between these groupings are far from rigid. Manuscripts are labeled Alexandrian, Western, or Byzantine because they predominantly show variants belonging to a particular textual tradition. But most manuscripts are, in fact, mixed: in varying degrees, they contain variants found in other traditions as well. Recognition of cross-pollination between the traditional manuscript groupings has led some scholars to discard the groupings entirely;⁶⁷ but, as we shall see, the groupings serve a purpose, provided we keep in mind that borders between them are not tightly circumscribed.⁶⁸

As a general rule, the Alexandrian tradition has proven more resistant to intentional change than other groups.⁶⁹ Presumably, scribes in this tradition had access to superior manuscripts; and though not immune to mistakes, they normally did not deliberately alter the text they copied. Some scholars credit them with possessing a superior text-critical sense as well.⁷⁰ Where readings found in Alexandrian manuscripts differ from those in other textual traditions, it is frequently easy to see how variants in the other traditions represent intended improvements of the Alexandrian reading, so that the latter is judged to be their starting point. The designation Western for certain manuscripts and particular variants is misleading: Western readings are found in Eastern as well as Western sources.⁷¹ Since, however, no other designation for these texts has gained widespread acceptance, the traditional title is generally (as here) retained. Variants labeled Western often go back at least to the second century and may, at times, represent the original form of the text, but frequently they instantiate the freedom exercised by (some) copyists in improving the text. The Byzantine tradition also preserves ancient variants; a few of those found only in Byzantine manuscripts may be original. But this tradition tends to show improved readings and, at times (as with the endings of Romans), combines (conflates) different variants in a single text, that nothing may be lost.

Turning to Romans, we may now inquire to what extent these broad characterizations prove true in our epistle.

3.1. THE ALEXANDRIAN TRADITION

The primary witnesses to the Alexandrian tradition in Romans (in addition to P46) are the uncials Vaticanus (designated by the letter B or the number 03), Sinaiticus (the Hebrew letter א or the number 01), Alexandrinus (A, 02), and Ephraemi (C, 04), as well as the tenth-century manuscript numbered 1739.⁷² Before we look at the group as a whole, something should be said about each of these important manuscripts.

In the judgment of the Alands, Vaticanus (B) is by far the most significant of the uncials.⁷³ Written about the middle of the fourth century, it may be the earliest extant manuscript of the entire Bible (though parts of Genesis and the Psalms as well as the end of the New Testament are now missing); even more importantly, for parts of the Bible—Old Testament (Septuagint) and New—it provides our finest witness to the original Greek text. Its quality is not the same throughout, however, and, in the Pauline epistles, it is not at its best. Its text in Romans is certainly more closely related to the other Alexandrian manuscripts than to any other group;⁷⁴ but like P46, B not infrequently shares readings with Western rather than other Alexandrian witnesses. In such cases, it is not always clear whether the text of B has been contaminated by a Western variant, or whether the reading in question survives in B and the Western manuscripts from an early period before the Alexandrian and Western traditions took their separate paths.

At 6:21, B and the three primary Western manuscripts (D F G) add on the one hand (μὲν) to the end of those things is death, intending to balance the δὲ (on the other hand, but) at the beginning of v. 22. This appears to be a secondary improvement of the text. At 14:21, B, together with the Western and the Byzantine textual traditions, expands the phrase by which your brother stumbles by adding or is offended or weak (see the King James Version; note, however, that SBLGNT and THGNT75 adopt the lengthier reading). At 5:2, where B and the three primary Western manuscripts lack by faith (τῇ πίστει), the omission may well be original: it finds additional early support from uncial 0220 and the Sahidic translation (more on this version below).

Again like P46, B has a good number of readings largely or completely unsupported by other Greek manuscripts.

At 1:32, B has participles (ποιοῦντες, doing, and συνευδοκοῦντες, approving) rather than finite verbs (do, approve); it is noteworthy that these readings of B appear to be known already at the end of the first century (1 Clement 35.6).⁷⁶ At 5:17 (abundance of grace and of the free gift of righteousness), B omits free gift (δωρεᾶς), perhaps deeming it redundant. At 9:3 (my brothers, my kin), B omits my brothers (τῶν ἀδελφῶν μου), perhaps thinking it redundant; here, however, the omission might simply be an oversight, the scribe’s eye passing from the article (τῶν) before brothers to the identical article before kin. At 14:13 (stumbling block or hindrance), B omits stumbling block (πρόσκομμα) as redundant. Conversely, for love of Christ (so some manuscripts) or love of God (so others) at 8:35, B has love of God which is in Christ Jesus, anticipating the language of 8:39.

Among the unique readings we may place several obvious scribal mistakes.

In chapter 4, the last nine words of v. 4 and the first four words of v. 5 were copied twice (dittography). At 15:13, B omits εἰς τὸ περισσεύειν, presumably because the preceding word also ended in -ειν; the eye of our scribe skipped from one word ending with -ειν to the next (homoioteleuton). At 16:8, the article τὸν is omitted—since the same three letters are the last letters of the preceding name. Here, our scribe wrote letters once that should have been written twice (haplography).

Sinaiticus (א), like Vaticanus, is from the mid-fourth century and once contained the entire Bible (parts of the Old Testament have been lost). In Romans, it belongs together, and much more firmly than B, with the other Alexandrian manuscripts,⁷⁷ but is not impervious to cross-pollination from other textual traditions.

At 8:21, א shares with the primary Western witnesses the conjunction διότι (because, or perhaps simply that) in place of the synonymous ὅτι. At 12:2, א shares with the Byzantine textual tradition the pronoun your [mind] (ὑμῶν), left implicit in the other Alexandrian and the Western manuscripts. Similarly, at 13:9, to the list of commandments from the Decalogue, א and the Byzantine manuscripts add You shall not bear false witness.

There are a number of singular readings, including a few blunders.

At 1:8, א* omits διὰ Ἰησοῦ Χριστοῦ (through Jesus Christ); since this is not a phrase that a scribe would choose to omit, our scribe’s eye presumably leapt from the ending of the preceding word (μου, my) to the similar ending of Χριστοῦ (homoioteleuton). For some reason (or, better, by some accident), א* omitted 11:30 entirely. At 14:23, our scribe’s eye skipped from the first ἐκ πίστεως (from faith) to the second, thus omitting the intervening words.

Alexandrinus (A), a fifth-century uncial once containing the entire Bible, is closely related to other Alexandrian manuscripts in Romans (and the other Pauline epistles),⁷⁸ although, in the Gospels, it represents the Byzantine tradition. Like Vaticanus and Sinaiticus, however, it at times departs from readings found in most of the Alexandrians, sharing variants with the other textual traditions.

With P46 and the primary Western manuscripts (D F G), but against the other Alexandrians (and most Byzantine manuscripts), A lacks the infinitive εἶναι (to be) at 6:11 (consider yourselves [to be] dead to sin). Against the other Alexandrians and the Western manuscripts, A shows a number of Byzantine readings, of which three may be cited. At 2:8, A and the Byzantines add μὲν (on the one hand) to the phrase not obeying the truth, highlighting the contrast with the following but [or on the other hand] obeying unrighteousness. At 10:17, they read word of God rather than word of Christ. With the Byzantines, A places the doxology after 14:23 (it retains it after 16:23 as well).

A has a number of distinctive readings because our scribe either thought too much or paid too little attention to what he was doing.

As instances of the former category, we may cite 2:5 (in the day of wrath and revelation [ἀποκαλύψεως] of the righteous judgment of God becomes, in A, in the day of wrath and recompense [ἀνταποδώσεως] of the righteous judgment of God); 3:9 (borrowing language from 1:16; 2:9, 10, A adds first [πρῶτον] with Jews to Paul’s charge that both Jews and Greeks are under sin); and 6:17 (A notes that when the Romans obeyed the gospel teaching, they did so, not simply from the heart [ἐκ καρδίας], but from a pure heart [ἐκ καθαρᾶς καρδίας]). Among our scribe’s blunders, we may note dittography at 2:23 (the letters ασ in παραβάσεως, transgression, are written twice); a lengthy parablepsis by which our scribe skips from διὰ τοῦ in 5:10 to διὰ