1. Writing and Publishing Acts
1
Although we lack specific information about how Acts was published, antiquity has supplied plenty of information about how works similar to Acts were published. Large narrative works such as Luke-Acts typically entailed multiple drafts and a circulation among people of some means. Examining these methods can give us a sense of how Luke may have published Luke-Acts.
The historical context helps us understand the reason that Acts is the length it is and how it may have been drafted (including likely revisions along the way).
Large histories could be long, requiring many volumes. Polybius planned nearly forty volumes (Polyb. 3.32.2) and complained that others’ histories were even longer and hence more difficult to follow (3.32.4). A book was approximately how much one could listen to in a comfortable sitting; on average, “a book of Herodotus or Thucydides is about 20,000 words, which would take around two hours to read. After the Alexandrian library reforms, an average 30–35 foot scroll would contain 10,000 to 25,000 words,” depending on the size of handwriting. Acts contains around 18,500 words (although variants can drive this estimate up or down), and so it is close to the middle of this expected range.
Authors often retained a sense of symmetry by keeping the separate books in a work to about the same length. Allowing for textual variations (and normally not including the Western text of Acts), the Gospel of Luke is about 19,500 words, and Acts about 18,500, within less than 6 percent of the Gospel in number of words. Matthew (nearly as long as Acts) and Luke are similar in word count; Mark is said to be about 61 percent the size of Matthew, 58 percent that of Luke, and 73 percent that of John. The proximity of Acts and Matthew to Luke in size suggests that certain lengths of standardized long scrolls might have been more available than others. Some scholars have thus proffered that the larger books of the NT (Luke, Acts, and Matthew) each would fill a standard literary scroll, close to the maximum length before a scroll became unwieldy.
Sometimes writers seem to have run out of space at the end of their scroll earlier than they intended, facilitating more abrupt endings. On other occasions writers more effectively budgeted the lengths of their accounts to prevent this deficiency from happening. Sometimes scrolls left blank space at the end; Luke, however, appears to have used virtually all the space available to him. Reading a work of this length, then, like “books” in multivolume histories, would have taken about two hours.
Most Greek and Roman historians began by composing a basic draft of the material in chronological order, to which a topical outline, speeches, and other rhetorical adjustments would be added later. Jewish narrative writers in Greek likewise usually began with a rough draft. This practice applied to other genres as well. Aristotle recommends sketching a work’s plot in outline, then expanding by inserting episodes; he illustrates this method with the Odyssey. Similarly, although some rhetoricians preferred extemporaneity, most urged careful premeditation, arranging the material in advance, so that they needed to add only finishing touches when they began writing their speeches. What held for speeches applied even more so to literary works, which were major undertakings.
A lengthy (by ancient standards) and expensive work such as Luke-Acts would invite careful forethought and arrangement to best communicate its message. This is especially the case if we accept the argument that Luke intended this work as a “foundation document” for the community (cf. Luke 1:4). Writing a thorough history could consume an enormous amount of time; historians thus might die before completing their works (e.g., Pliny Ep. 5.5.5). To finish a literary project sometimes required shutting out other social demands to focus on writing (1.3.3–4).
In rhetoric, the arrangement of material thoroughly informed the way that educated hearers and readers evaluated works, and so a writer could hardly afford to ignore it. Some critics advised connecting episodes to provide continuity; others, such as Polybius, allowed disjunctions in their narratives, though recognizing that not all historians agreed.
Revision during the process of composition was standard practice. For the Aeneid, Virgil first composed twelve books in prose, then revised them in verse (in whatever sequence he chose; Suet. Vergil 23). He reportedly dictated many verses in the morning, then spent the rest of the workday condensing them (Vergil 22). Even two centuries later, Virgil’s work continued to circulate in different forms; Aulus Gellius (Aul. Gel. 6.20) notes that Virgil changed the earlier reading of a passage to the later one and explains why. In other works, too, writers often added nonessential clauses or removed important ones for the sake of better sound. Although revisions can be expansive, ancient writers could employ them for abridgment. In most periods ancient rhetoric valued conciseness and brevity so long as they did not lead to obscurity of expression. Luke thus keeps his account moving by focusing on elements necessary for his story and omitting extraneous material.
Historians frequently followed an earlier historian in the first draft; Josephus seems to have followed his own Jewish War at points in his later magnum opus, the Antiquities of the Jews; 4 Maccabees adapted material from 2 Maccabees. Second Maccabees itself abridges an earlier work by Jason of Cyrene (2 Macc 2:23). Some argue that, like other Greek writers, Luke mostly follows one source at a time, at least in his Gospel, where he incorporates a large block of Q material into Mark; both he and Matthew make Mark the backbone and supplement this material from other sources. Some people viewed use of preexisting lines as plagiarism, others (when the incorporation was obvious) as flattering the source (Sen. E. Suas. 3.7). (The nature of literary borrowing changed dramatically in the modern period, after England’s copyright law of 1709.) The gospel tradition seems to have become common property of the early church and its preaching, allowing less explicit acknowledgment of any particular documents.
“Publication” most often began with public readings; the well-to-do incorporated readings as entertainment following the dinner at banquets. (They often had their own readers.) Educated members of churches or servants of house churches’ patrons probably read Luke-Acts after the Lord’s Supper, a dinner (1 Cor 10:21; 11:20–34; Jude 12). Oral recitations at banquets and other social occasions provided one form of publication, but written publication included not only the reading of a work at such occasions but sometimes its wider circulation.
Hearers (or readers) of means who heard a reading and liked a work could have transcriptions made for their own copies, thereby further circulating the work. Members of the wealthy elite could obtain copyists to “circulate” a book in this manner. Because only the elite provided “access to the channels of literary reception,” writers might hope for their elite friends to circulate their writings among other members of the elite. One sometimes secured the favor of such an individual with a dedication (see comment on ) or more elaborate literary pieces.
Christians in northeastern Mediterranean house churches may have heard Luke-Acts read often, but few would have been able to read their own copies, both because of the level of literacy required for such an activity and because of the limited affordability of a copy. (Many scholars have estimated literacy in the empire at about 10 percent. A higher estimate is possible, especially in urban areas, but reading a work such as Acts required much more than basic literacy.)
An author with especially elite connections might become more proactive, reading a work first to friends (a supportive audience), then to a wider public. Such oral recitations provided the benefit of feedback that permitted revisions to improve the work before its wider circulation. (Oral publication might also occasionally entail the danger of plagiarism. One man with an exceptionally well-trained memory, hearing an author reciting a poem, claimed that the author had stolen the poem from him. The hearer then repeated it back verbatim, a feat that the actual author could not reproduce.) Provided these oral tests proved successful, the author might publish the work or, far more often, “entrust it to a publisher who ordered the copies, assumed the production costs,” and distributed the work. Cicero’s friend Atticus, one such publisher, had many scribes in his workshops. Scribes usually wrote documents while sitting down, writing on the sheet on their lap.
Modern readers accustomed to mass printing should not overplay the distinction between circulating a draft among acquaintances and publishing a final copy. To the extent that we can make such a distinction, however, subsequent editions usually supplanted their predecessors.
Some works circulated in more than one form. Aristotle organized his teachings into both notes, which were popular and random, and the more formally arranged books, which remain extant today. For authors with adequate resources, such works might remain continually in revision, improving the rhetorical quality of a work. Authors might release these published works in various editions; thus, for example, Ovid abbreviated his earlier five-volume edition of the Amores to three volumes (Am. prol.).
An author might also release a prepublication copy but expect that it not be “published,” that is, circulated beyond the one reader, until it had been polished. The dedicatee should then receive the first copy (Cic. Att. 13.21a). After copyists had a book ready, it would be checked for corrections (13.23), and the writer might ask a friend to be certain that the revised copy was the one sent (13.48). One might even claim that one had written the book for a particular reader’s attention, not initially circulating it (Rhet. Alex. pref. 1421a.27–38, 1421b.6). At other times, well-meaning friends apparently circulated an author’s notes or draft, provoking the author to object that the work was still unfinished.
Some scholars think that Acts circulated in multiple editions, arguing that this diversity would explain some of the textual tradition. In any case, the normal process of examination and revision suggests that a significant narrative work such as Acts was not lightly composed but reflects careful composition and revision from information to cohesive narrative. It also suggests that Luke would have had ample opportunity to recognize and correct or explain some of the points that modern readers find curious had they seemed equally curious to him. Of course, no amount of revision allows one to guard against every possible misreading.
Pliny the Younger, who deliberately cultivated his literary reputation, provides abundant examples of the process noted above, although he wrote speeches rather than histories. He was not the first to offer readings even of his speeches, but this was not a common practice (Ep. 7.17.4). Readings of various other forms of literature were, however, common among the Roman upper class. In one letter, Pliny observed that most recent days had included some readings (1.13.1), and elsewhere he notes that some earlier writers (he specifies Silius Italicus) had submitted their verses to public feedback (3.7.5). Since Pliny read only speeches well, he would ask one of his freedmen to read verse (9.34.1).
Pliny complains that most potential hearers of his day preferred to sit about gossiping (1.13.1–2), but he attended most readings given by friends who invited him (1.13.5), and he praises other readings he appreciated (6.21.2). He praised one patron for restoring literature (8.12.1) by lending his home for such readings and attending them elsewhere (8.12.2). He also invited his own friends for his informal readings (9.34.1). He seemed pleased when, although he urged friends to come only if convenient, they showed up for two days of readings and “forced” him to continue on the third (3.18.4). One might fail to invite those whom one expected to be offended. A person who liked another’s reading could seek to procure a written copy to send on to a friend (6.21.7).
One reason Pliny invited friends was to secure feedback so that he could edit the works before publishing them more widely. This he considered an important service to his readers (7.17.8–10). The reading likewise gave him a sense of which passages were inappropriate in front of a real audience (5.3.8). He also solicited feedback from those to whom he sent copies of his speeches, sometimes at their request. Occasionally he needed advice whether to even publish a work. He provided the same favor to others who sent him their works, and included critical feedback to improve them. Such “editing” no doubt prevented many postpublication embarrassments.
Pliny’s own works, at least, seem to have circulated widely. He explains that only by revising and publishing his speeches can he hope for them to survive him (5.8.6–7). His friends might help him publish a work (1.2.5). He claims that he has heard that his books are being read throughout the world, although the booksellers might simply be “flattering” him (1.2.6). He knows that his works are popular in Rome and is happy to hear that his books are being sold by booksellers in Gaul (9.11.2). He is pleased when a friend assembled copies of his speeches and sent them to him for correction (4.26.1).
Pliny also hopes for others’ works to circulate. He encourages Suetonius to write more quickly so that his books will be “copied, read and sold.” He encourages another friend that he should at least offer readings of his work; once he felt the crowds’ admiration, he might prove more inclined to publish (2.10.6). Pliny warns another that, important as revisions are, one must eventually stop revising, finish the work, and then begin another one (9.35.2). He seems annoyed, however, when his wealthy rival Regulus assembled a large audience for a reading Pliny considered frivolous. Regulus then had the memoir copied and distributed in Italy and the provinces, appealing to town councils to have it read.
Such means of publication were available only to the very elite, but they offer at least some sense of how ancient works were circulated through networks interested in prominent persons’ presentations. Not likely one of the elite himself, though with sufficient education to produce a popular or midlevel historical monograph with minimal rhetorical embellishment, Luke could hope to achieve wider circulation only through his Christian network. He might depend on some members of the elite to promote it within those circles (perhaps starting with Theophilus; see comment on ) but would have to depend especially on the interest that his content would generate.
Quotations in early church fathers suggest that for some time his Gospel (which fit an established genre category and focused on Jesus, the center of the church’s attention) drew more attention than his Acts. Early Christians were interested enough, however, to preserve his work, even as a widely accepted element in the emerging catholic “canon” of apostolic readings. Over the course of time, then, his work was more widely circulated than he had likely imagined possible.
Luke may have begun by producing a rough draft of his two-volume work (or of the Gospel first, with a draft of Acts added after the Gospel was more polished). He could then receive feedback on this work through public readings at banquets (probably especially the Christian banquet that constituted part of the church’s regular meetings) and revise it accordingly. An elite sponsor—or hoped-for sponsor, such as Theophilus (see comment on )—could provide the work wider circulation in circles that could afford to have more copies made.
Burridge, “People,” 141; idem, Gospels, 118. He adds (“People,” 141n49) that scrolls were generally 8–12 in. high, with columns “2–4 inches wide, with 18–25 letters per line and 35–45 lines per column.”
Witherington, Acts, 7, suggesting also symmetry in the periods covered in Luke’s Gospel and Acts (about three decades in each).
My precise counts from the version on my computer are 19,539 and 18,496, respectively, but given textual uncertainties, even the figures provided above are rough estimates. The estimate of 19,428 words that some make for Luke, e.g., is probably closer.
Morton in Morton and MacGregor, Structure, 16, estimates that Luke and Acts are roughly the same length (cf. also p. 15); Matthew is within 1 percent of the length of either; John is within 1 percent of three-quarters this length, and Mark is close to half.
See, e.g., Bruce, Books, 12; Palmer, “Monograph (1993),” 5. Witherington, Acts, 6, follows an estimate that Acts was about 32 ft. long and the Gospel about 35 ft. long, close to the maximum 40 ft. scrolls; Metzger, Text, 5–6, estimating from a more compact writing size, suggests that each was about 31–32 ft. long, the maximum length being 35 ft.
Jos. Ag. Ap. 1.320; perhaps Dion. Hal. Demosth. 58. Luke may have included more resurrection material (Luke 24) than, e.g., Matthew because he had space available at the end.
Corn. Nep. 15 (Epaminondas), 4.6.
See, e.g., Diog. Laert. 7.2.38.
Burridge, “People,” 141.
Burridge, Gospels, 203; Aune, Environment, 82, citing Jos. Ag. Ap. 1.47–50; Lucian Hist. 16, 48; Lucian Dem. Some scholars suggest that Plutarch’s extant work on self-love was probably his unfinished draft (Van der Stockt, “Hypomnema”).
Aune, Environment, 128. Thus Josephus fully revised an earlier draft of the Jewish War into better Greek (Jos. Ag. Ap. 1.49–50); some scholars think that the earlier version was an Aramaic draft, probably circulated among Parthian Jews (cf. Hata, “Version”).
For a summary of the process (collecting material, producing draft versions, followed by finishing touches), see, e.g., Dorandi, “Copy,” 775.
Talbert, John, 64, citing Arist. Poet. 17.6–11. Aristotle’s interest in plot admittedly involves drama more than historiography (Butcher, Theory, 164–65).
Quint. Inst. 10.6.1–2. Orators should also be ready to add improvisations during a speech (10.6.5). Small, “Orator,” 204–5, observes that when preparing for speeches, Pliny (Ep. 9.36) composed section by section in his head, after each section calling in a scribe to write it down; this practice facilitated memorization and was common until the fourteenth century C.E. (see also Quint. Inst. 10.6.1).
The cost for taking slow dictation, plus papyrus (though scholars often overestimate papyrus’s cost; see esp. Skeat, “Cheap”), without counting rough drafts or travel expenses, might be about forty denarii, or in today’s U.S. equivalent, as high as four thousand dollars (see Richards, Letter Writing, 165–69, 231; inflation by the time of Diocletian’s edict, which he cites, was significant).
Thus allowing such literary techniques as foreshadowing (Quint. Inst. 10.1.21). Editing provided the writer a chance to craft the material; Epictetus’s Diatribai, e.g., undoubtedly bears less of Arrian’s stamp than the Encheiridion, where he organizes and summarizes Epictetus’s teachings.
With, e.g., Talbert, “Chance,” 230; Pao, Isaianic Exodus, 3.
Some scholars have suggested that the final form of Acts remains unfinished, perhaps because of Luke’s premature decease (either assuming that he wrote late in life or, assuming that he died young, against probably much later tradition regarding his longevity).
See esp. Wuellner, “Arrangement.” Some forms of speeches did, however, allow random sequence (Men. Rhet. 2.4, 391.19–28; 392.9–14; 393.23–24).
See, e.g., Keener, Corinthians, 29–31.
Aune, Environment, 90, cites Lucian Hist. 55; Quint. Inst. 7.1.1. Cf., e.g., perhaps Mark 1:14–39; Smith, Parallels, 131.
Aune, ibid., 90, cites Polyb. 38.5.1–8.
Cf. the opening Virgilic lines of the Aeneid removed by the final editors (LCL, 1:240–41, esp. n. 1).
Dion. Hal. Lit. Comp. 9. On expanding and abridging, see at length the discussion under genre in , , below.
E.g., Dion. Hal. Thuc. 55; Demosth. 18, 20, 24; Lysias 5; Philost. Vit. soph. 2.4.569; for much fuller treatment, see comment at Acts 24:4.
Dion. Hal. 2 Amm. 2.
Krieger, “Hauptquelle.”
See Gardner, “Mqbym.”
Aune, Environment, 139. Laistner, Historians, 83, 120, questions this common approach (with regard to Livy and Tacitus, respectively).
Aune, Environment, 65; cf. Downing, “Conventions”; idem, “Actuality”; Burridge, Gospels, 204–5. Agreements between Matthew and Luke against Mark in Markan material suggest that Luke felt free to adapt his dominant source in light of his larger repository of oral tradition and written information.
The elder Seneca lamented that even obvious, unethical plagiarism could go undetected because of the audience’s negligence (Suas. 2.19; McGill, “Seneca on Plagiarizing”). (Material in this section has been incorporated, though greatly augmented, rearranged, and revised, from Keener, John, 5–7.)
Frye, “Problems,” 275, noting the frequency of conflation before that time.
Burridge, One Jesus, 20; Alexander, “Book Production,” 86, 90; Dewey, “Oral-Aural Event,” 145–47; Shiell, Reading Acts, 117–22; Dorandi, “Publication,” 185; cf., e.g., Diog. Laert. 1.122; Corn. Nep. 25 (Atticus), 14.1; Cic. Att. 2.1; 12.44; Sen. E. Controv. 1.pref. 19; Sen. Y. Ep. Lucil. 95.2; Statius Silv. 2.pref.; Iambl. V.P. 21.98–99; other sources in Keener, Matthew, 297. For public readings of histories, see Nicolai, “Place of History,” 23.
Cic. Fam. 7.1.3. Readers and reciters in wealthy households were frequently educated slaves (see Duff and Spawforth, “Anagnōstēs,” noting Cic. Att. 1.12.4; Corn. Nep. 25 (Atticus), 13.3; 14.1; Aul. Gel. 3.19).
For readings at banquets, see, e.g., Pliny Ep. 9.17.3.
See Alexander, “Book Production,” 86, 90; Dewey, “Oral-Aural Event,” 145–47; Winterbottom, “Recitatio”; cf. Iambl. V.P. 21.99; Starr, “Reading Aloud.” On the limitations of literacy in antiquity, see the oft-cited Harris, Literacy. Lecturers sometimes brought copious research and hoped that their hearers would encourage them to continue reading (Sen. Y. Ep. Lucil. 95.2).
E.g., Phaedrus 4.prol. 17–19. The wealthy might also acquire copies for their personal libraries, which became a symbol of status (see Kunst, “Privatbibliotheken”; Suda, s.v. “Epaphroditos,” in Sherk, Empire, §178 I, p. 238; Cic. Fam. 13.77.3; Quint. fratr. 3.4.5; Att. 1.4, 7; such personal libraries appear even in Egypt’s nome capitals [Lewis, Life, 60]).
Cic. Att. 12.44. Cf. Gamble, “Literacy,” 646.
Other writers recited their works at public festivals (e.g., Dio Chrys. Or. 8.9; 27.5), but with less hope of sponsorship for continued circulation (many did not want to hear them; cf. 27.6).
Cic. Att. 2.1; Statius Silv. 2.pref.
Citroni, “Patronage, literary,” 1125. A noble might even dedicate a work to the emperor, apologizing for its unworthiness (Pliny E. N.H. pref. 6, 12).
On public literacy, see, e.g., Lewis, Life, 61–62, 81–82. Scholars typically estimate literacy at ca. 10 percent, and not higher than 20 percent (Meeks, Moral World, 62; Botha, “Literacy”; Aune, Dictionary of Rhetoric, 276; Gamble, “Literacy,” 644–45; Scholer, “Writing,” 1283; see esp. discussion in Harris, Literacy, who is often cited by the other scholars and allows for regional variation; in Jewish Palestine, see Hezser, Literacy, 473, 496, as cited in Harrison, Authorities, 20). By “literacy” such estimates often mean the ability to read and understand texts of medium complexity.
See Horsfall, “Statistics” (cited in Harrison, Authorities, 20); Curchin, “Literacy”; Millard, “Literacy.” In Egypt most metropolites (citizens of Greek-speaking nome capitals), in contrast to typical villagers, could read and write (Lewis, Life, 61–62). Note also the inscriptions (esp. CIL 4.4345, 4356, 4397, 4706–85; 7.12.18–20) cited by Franklin, “Literacy,” 89–98, and noted in Harrison, Authorities, 20–21 (also citing Woolf, “Literacy,” 48–49).
See, e.g., Suet. Vergil 33; Shiell, Reading Acts, 115. Winterbottom, “Recitatio,” 1296, cites Pliny Ep. 5.12.1–2 but guesses (probably rightly) that it often generated flattery instead (Hor. Ars 428–31).
Sen. E. Controv. 1.pref. 19. Since Seneca knew which was the true author, he probably is accusing the other of showing off, not of plagiarism.
Dorandi, “Copy,” 776. Some question whether the friendship between Cicero and Atticus was as ideal as has often been supposed (Marchetti, “Words”).
Richards, Letter Writing, 55, noting one scribe’s quip in a manuscript note that “good writing involved the cooperation of right hand, pen and knee”; for the diverse kinds of secretaries available, see 64–80.
Heyworth and Wilson, “Editions, second,” 809.
Ibid., 812; cf. idem, “Variants.” Political changes sometimes required revisions (“Editions, second,” 811).
Cic. Fin. 3.3.10; 5.5.12. For papyrus and other writing substances, see Maehler, “Books.”
Dion. Hal. Thuc. 24, noting that Thucydides continued revising his eight books over a twenty-seven-year period. Aune, “Publication,” 389, notes two revisions of a work by Tertullian (Marc. 1.1).
Cf. Symm. Ep. 1.31. Augustus apparently wanted such a rough draft of the Aeneid, or at least a part of the work (Suet. Vergil 31), and Virgil provided him private readings of Aen. 2, 4, and 6 (Suet. Vergil 32). Earlier scholars’ hypothesis of a proto-Luke (e.g., Streeter, Gospels, 201–22; C. Williams, Acts, 12–13) fits the pattern of ancient authors going through several drafts, although the suggestion that Luke subsequently discovered and added Mark suggests a more thorough revision than is likely. Morton in Morton and MacGregor, Structure, 35, proposed a proto-Acts to match proto-Luke.
This claim is probably a fiction; the writer of Rhet. Alex. almost certainly hoped that this detailed work would circulate beyond the reader he flatters.
Aune, “Publication,” 389 (citing Quint. Inst. 1.pref. 7–8).
Elsewhere, however, Pliny complains that speeches do not “read” well, and hence he agrees to read his speech to his friends only on a friend’s recommendation (Ep. 2.19.1–4). The Roman practice of public recitals of aspiring poets and of literature was common in the early empire and rooted in earlier precedents (Bloomer, “Declamation,” 297).
A situation Pliny starkly contrasts with an earlier era (Ep. 1.13.3–4). He also complains about audiences too arrogant to applaud appropriately (6.17.1–6).
Thus Regulus failed to even invite Pliny to his attack on Rusticus (Pliny Ep. 1.5.4; cf. 1.5.2).
Pliny Ep. 5.3.8; 7.17.1–3, 7; 8.21.4.
E.g., Pliny Ep. 1.2.1.
E.g., Pliny Ep. 1.8.1; 2.5.2 (although here he does not send the part he is still revising, 2.5.1); 3.13.1, 5.
E.g., Pliny Ep. 1.2.5; 1.8.3. His hesitation might be feigned, inviting compliments—which he certainly doles out to others (e.g., 3.15.3).
Poems in Pliny Ep. 3.15.1–2.
Pliny Ep. 7.20.1 (to Tacitus, who was doing him the same favor, 7.20.2).
Pliny Ep. 5.10.3 (LCL, 1:367); cf. 5.10.1–2.
Pliny Ep. 4.7.2 (it was a memoir of the life of Regulus’s son, who died as a child).
This observation refers to its virtually universal geographic range in the ancient Mediterranean world; the questions of whether such works addressed particular “communities” () or whether Luke self-consciously treated his work as a continuation of Israel’s story () are discussed elsewhere in this commentary.