1 Inv. No. 270-1867. The ivory is the first of the so-called Grado-Chair ivories, a sequence of fourteen stylistically coherent ivories which have been dated any time between the sixth and twelfth centuries. They were first discussed in Graeven, H., ‘Der heilige Markus in Rom und in der Pentapolis’, Römische Quartalschrift 13 (1899) 45–91Google Scholar, which argued that they should be dated to 610–41 ce. W. F. Volbach's theory that they are an eleventh-century Southern Italian collection related to the Salerno ivories has found favour with New Testament scholars but carbon-dating of the Mark plaque revealed that it should be dated to 440–670 ce with a 95.4% probability (with a 68.2% probability that it they were made between 550 and 650). See Volbach, W. F., Elfenbeinarbeiten der Spätantike und des frühen Mittelalters (Mainz: Von Zabern, 1976)Google Scholar, and discussion in Williamson, P., ‘On the Date of the Symmachi Panel and the So-Called Grado-Chair Ivories’, Through a Glass Brightly: Studies in Byzantine and Medieval Art and Archaeology Presented to David Buckton (ed. Entwistle, C.; Oxford: Oxbow, 2003) 47–50Google Scholar, at 48–9.

2 This understanding of the literary relationship between the two figures is captured in both the way the relief has traditionally been titled and scholarly descriptions of it.

3 For example, on the relationship of the composition of Mark to the First Jewish War with Rome, see Collins, A. Yarbro, Mark (Hermeneia; Minneapolis: Fortress, 2007) 14Google Scholar; Stein, R. H., Mark (BECNT; Grand Rapids: Baker Academic, 2008) 14–15Google Scholar.

4 See Bauckham, R., ‘For Whom Were Gospels Written?’, The Gospels for All Christians: Rethinking the Gospel Audiences (ed. Bauckham, R.; Grand Rapids: Eerdmans, 1998) 9–49Google Scholar; Mitchell's, M. M. ‘Patristic Counter-Evidence to the Claim that “The Gospels Were Written for All Christians”’, NTS 51 (2005) 39–79CrossRefGoogle Scholar; R. Bauckham, Jesus and the Eyewitnesses: The Gospels as Eyewitness Testimony (Grand Rapids: Eerdmans, 2006) and ‘Is There Patristic Counter-Evidence? A Response to Margaret Mitchell’, The Audience of the Gospels: The Original Function of the Gospels in Early Christianity (ed. E. W. Klink III; LNTS 353; London: T&T Clark, 2010) 68–110.

5 Generic analysis has tended towards the view that Mark is a species of ancient biography. This argument was first made by J. Weiss, Das älteste Evangelium: Ein Beitrag zum Verständnis des Markus-Evangeliums und der ältesten Evangelien (Göttingen: Vandenhoeck & Ruprecht, 1903) and revived by R. A. Burridge, What Are the Gospels? A Comparison with Graeco-Roman Biography (Cambridge: Cambridge University Press, 1992). On generic analysis, see now H. K. Bond, The First Biography of Jesus: Genre and Meaning in Mark's Gospel (Grand Rapids: Eerdmans, 2020).

6 Cf. the Mark of Phlm 24; Col 4.10; 2 Tim 4.11. Inasmuch as those treatments of the question want to probe Mark's identity further, it is on the basis of Acts and 1 Peter, rather than the scene envisioned by Papias, that this conversation proceeds. See discussion in Niederwimmer, ‘Johannes Markus’, 178–83; C. Clifton Black, Mark: Images of an Apostolic Interpreter (Edinburgh: T. & T. Clark, 2001) 67; J. Marcus, Mark 1–8 (Anchor Yale Commentary; New York: Doubleday, 2002) 19–21; Bauckham (Jesus and the Eyewitnesses, 206), who describes John Mark as coming ‘from a Jewish Diaspora family and … presumably educated in Jerusalem’. On the relationship between Mark and Paul, see M. M. Mitchell, ‘Epiphanic Evolutions in Earliest Christianity’, ICS 29 (2004) 183–201 and M. Kok, ‘Does Mark Narrate the Pauline Kerygma of Christ Crucified? Challenging an Emerging Consensus on Mark as a Pauline Gospel’, JSNT 37 (2014) 139–60.

7 Examinations of the presentation of Mark in early Christian literature have rightly stressed the way Mark is framed as an acolyte of Peter and the manner in which Peter serves to authenticate and authorise the Gospel of Mark, perhaps even ensuring its survival. Overlooked, however, is how the second-century framing of the Second Gospel as disordered and the presentation of Mark himself as the mere conveyer of Petrine tradition mutually reinforce one another. See, for example, Kok's excellent Gospel on the Margins, 185–227. A rare exception to this rule is A. Yadin-Israel, ‘“For Mark Was Peter's Tannaʾ”: Tradition and Transmission in Papias and the Early Rabbis’, JECS 23 (2015) 337–62.

8 This article takes its lead from an often-overlooked 1984 article by philologist Günther Zuntz, which examined the way that gentile readers of the Gospel of Mark would have understood the text. Among his many interesting insights, he noted that for gentiles the appellation ‘Jesus Christ’ in the opening titular sentence of the Gospel would either have meant ‘Jesus-ointment’ or something along the lines of ‘Jesus the painted one’. More probably, he argued, gentiles might have assumed that a typographical or pronunciation error had taken place and that the copyist or lector had intended to write ‘Chrestos’ rather than ‘Christos’ (cf. Tacitus, Ann. 15.44). On this assumption, the ‘reader’ would assume that the Gospel was about a Semitic slave (Jesus) who had been given the new name Chrestos. Zuntz's article was an elegant articulation of the ways in which descriptions of identity and status vary according to the cultural positionality of the reader. G. Zuntz, ‘Ein Heide las das Markusevangelium’, Markus-Philologie. Historische, literargeschichtliche und stilistische Untersuchungen zum zweiten Evangelium (ed. H. Cancik; WUNT 33; Tübingen: Mohr Siebeck, 1984) 205–22, at 205.

9 Even Matthew Larsen's thesis that Mark was incomplete and functioned as a ‘rough draft’ only grazes the question of how Mark himself is constructed. M. D. C. Larsen, Gospels before the Book (Oxford: Oxford University Press, 2018) and ‘Accidental Publication, Unfinished Texts and the Traditional Goals of New Testament Textual Criticism’, JSNT 39 (2017) 362–87. Moreover, Larsen is sceptical about the existence of ‘Mark’ at all, see his ‘Correcting Gospel: Putting the Titles of the Gospels in Historical Context’, Rethinking ‘Authority’ in Late Antiquity: Authorship, Law, and Transmission in Jewish and Christian Tradition (ed. M. D. Letteney and A. J. Berkovitz; London: Routledge, 2018) 78–103. See also the earlier important work of Kok, who uses similar language arguing that Clement views Mark as ‘rough notes’ in contrast to ‘the polished texts of Matthew and Luke’ (M. J. Kok, The Gospel on the Margins (Minneapolis: Fortress, 2015) 210–11).

10 For further discussion on literate slaves, see below. There is a tendency among scholars of early Christianity to overlook the presence of slaves in early Christian communities. On this, see R. A. Horsley, ‘The Slave Systems of Classical Antiquity and their Reluctant Recognition by Modern Scholars’, Semeia 83–4 (1998) 19–66. There is now, finally, a wealth of literature on slavery in general; in particular, see the important work of J. A. Harrill, The Manumission of Slaves in Early Christianity (Tübingen: Mohr Siebeck, 1995); J. Glancy, Slavery in Early Christianity (Oxford: Oxford University Press, 2002); J. A. Harrill, Slaves in the New Testament: Literary, Social, and Moral Dimensions (Minneapolis: Fortress, 2007); B. J. Brooten, ed., Beyond Slavery: Overcoming its Religious and Sexual Legacies (New York: Palgrave Macmillan, 2010); K. A. Shaner, Enslaved Leadership in Early Christianity (Oxford: Oxford University Press, 2018).

11 On Paul's use of secretaries, see E. R. Richards, Paul and First-Century Letter Writing: Secretaries, Composition and Collection (Downers Grove, IL: InterVarsity, 2004) and The Secretary in the Letters of Paul (WUNT ii.42; Tübingen: Mohr Siebeck: 2019). On Paul's literacy, see the discussion in C. Keith, ‘“In My Own Hand”: Grapho-Literacy and the Apostle Paul’, Bib 89 (2008) 39–58. On Origen's use of secretaries, see Jerome, On Illustrious Men 3.61 and Eusebius, Hist. eccl. 6.18.1–2, 23.1–2.

12 On the fragments of Papias, see J. Kürzinger, Papias von Hierapolis und die Evangelien des Neuen Testaments (Regensburg: F. Pustet, 1983); D. A. Baum, ‘Der Presbyter des Papias über einen “Hermeneuten” des Petrus: Zu Eusebius, Hist. eccl. 3,39,15’, TZ 56 (2000) 21–35; E. Norelli, Papia di Hierapolis, Esposizione degli oracoli del Signore. I frammenti (Milan: Paoline, 2005); D. R. MacDonald, Two Shipwrecked Gospels: The Logoi of Jesus and Papias's Exposition of Logia about the Lord (Atlanta: SBL, 2012); Stephen Carlson, Papias of Hierapolis, Expositions of Dominican Oracles: The Fragments, Testimonia, and Reception of a Second-Century Christian Commentator (Oxford Early Christian Texts; Oxford: Oxford University Press, forthcoming). For the rare view that this section of Papias is a forgery, see A. Loisy, Les évangiles synoptiques, vol. i (Paris: Ceffonds, 1907) 243.

13 On the title of Papias’ work and its relevance to his authorial project, see A. D. Baum, ‘Papias als Kommentator evangelischer Aussprüche Jesu: Erwägungen zur Art seines Werkes’, NT 38 (1996) 257–76. The dating of Papias’ writing to 130 CE, as Bauckham has discussed, rests on statements made by the fifth-century Philip of Side (Bauckham, Jesus and the Eyewitnesses, 13–14). This has led others to suggest that he wrote as early as 110 ce (or even earlier). See, for example, U. H. J. Körtner, Papias von Hierapolis (FRLANT 133; Göttingen, 1983) 89–94.

14 καὶ τοῦθ’ ὁ πρεσβύτερος ἔλεγεν· Μάρκος μὲν ἑρμηνευτὴς Πέτρου γενόμενος, ὅσα ἐμνημόνευσεν, ἀκριβῶς ἔγραψεν, οὐ μέντοι τάξει τὰ ὐπὸ τοῦ κυρίου ἢ λεχθέντα ἢ πραχθέντα. οὔτε γὰρ ἤκουσεν τοῦ κυρίου οὔτε παρηκολούθησεν αὐτῷ, ὕστερον δὲ, ὡς ἔφην, Πέτρῳ· ὃς πρὸς τὰς χρείας ἐποιεῖτο τὰς διδασκαλίας, ἀλλ’ οὐχ ὥσπερ σύνταξιν τῶν κυριακῶν ποιούμενος λογίων, ὥστε οὐδὲν ἥμαρτεν Μάρκος οὕτως ἔνια γράψας ὡς ἀπεμνημόσευσεν. ἐνὸς γὰρ ἐποιήσατο πρόνοιαν, τοῦ μηδὲν ὧν ἤκουσεν παραλιπεῖν ἢ ψεύσασθαί τι ἐν αὐτοῖς. (Eusebius, Hist. eccl. 3.39.15; trans. Ehrman, LCL 25, 103).

15 See Kok, Gospel on the Margins, 12.

16 So K. Niederwimmer, ‘Johannes Markus und die Frage nach dem Verfasser des zweiten Evangeliums’, ZNW 58 (1967) 172–88, at 185–8.

17 See F. H. Colson, ‘Tάξει in Papias (the Gospels and the Rhetorical Schools)’, JTS 14 (1912) 62–9. See also the summary statement of Kok: ‘The purpose of the apologetic was that Mark's fidelity to Peter compensates for the defects of the Gospel’ (Gospel on the Margins, 187). I agree with many others that the Prologue to Luke may house an implicit critique of Mark.

18 Rigg, who discerns no lack of order in Mark, argues that Papias has ‘garbled’ the tradition here and that the phrase should not be ‘οὐ μέντοι τάξει’ but rather ‘οὐ μέντοι ταχύς’ (‘not at all hastily’). This conjecture is based entirely on Rigg's own sense that Mark has structure. See H. A. Rigg, ‘Papias on Mark’, NT 1 (1956) 161–83, at 171.

19 Intriguingly there is at least one other candidate for the role. Clement of Alexandria writes that Basilides claims to have been taught by Glaucias, ‘the interpreter of Peter’ (Stromateis 7.106.4). The description prompted Niederwimmer to argue that Papias’ statements were motivated by anti-gnostic apologetic. See Niederwimmer, ‘Johannes Markus’, 186 and W. Bauer, Orthodoxy and Heresy in Earliest Christianity (ed. R. A. Kraft and G. Krodel; Philadelphia: Fortress, 1971) 184–9. Martin Hengel reversed the lines of influence, positing that Basilides developed his claim from reading Papias, in The Four Gospels and the One Gospel of Jesus Christ (trans. J. Bowden; London: SCM, 2000) 58.

20 For the view that Mark translated (spoken) Greek into Latin, see J. B. Lightfoot, The Apostolic Fathers, vol. i.2 (New York: MacMillan, 1889) 494. For the view that Peter grew up speaking both Aramaic and Greek, see M. Bockmuehl, The Remembered Peter in Ancient Reception and Modern Debate (WUNT i.262; Tübingen: Mohr Siebeck, 2010) 155.

21 Kürzinger adopts this perspective when he describes Mark as an expositor or ‘Mittelsmann’ (Papias, 16).

22 For this view, see T. Y. Mullins, ‘Papias on Mark's Gospel’, VC 14 (1960) 216–24 and Bauckham, Jesus and the Eyewitnesses, 210–11. In this final possibility, it is more probable that Mark acted as a scribe recording the oral words of Peter than as someone who translated an Aramaic or Hebrew document into Greek.

23 An enormous amount of material has been devoted to the question of Peter's literacy. On traditions of Peter's literacy in early Christianity, see S. A. Adams, ‘The Tradition of Peter's Literacy: Acts, 1 Peter, and Petrine Literature’, Peter in Early Christianity (ed. H. K. Bond and L. W. Hurtado; Grand Rapids: Eerdmans, 2015) 130–45. I depart from Adams, who argues that there was no such thing as an illiterate text broker in the ancient world; on the contrary, there were illiterate ‘scribes’ and ‘translators’ in the ancient world. See H. C. Youtie, ‘Ὑπογραφεύς: The Social Impact of Illiteracy in Graeco-Roman Egypt’, ZPE 17 (1975) 201–21. It is worth noting that Papias describes Matthew as having been written in Hebrew (or Aramaic?) and subsequently translated into Greek. Several scholars have seen here an implicit critique in the way that the pristinely ordered Matthew was translated into Greek as ‘each was able’ (ὡς ἧν δυνατὸς ἕκαστος). For example, see M. Hengel, Studies in the Gospel of Mark (Philadelphia: Fortress, 1985) 47–8 and Bauckham, Jesus and the Eyewitnesses, 225. What is missed here is the way in which the phrase ‘as far as possible’ is not an admission of intellectual impoverishment, but rather a standard legal disclaimer. See R. Mairs, ‘Hermēneis in the Documentary Record from Hellenistic and Roman Egypt: Interpreters, Translators and Mediators in Bilingual Society’, Journal of Ancient History 7.2 (2019) 1–53, at 8. The phrase appears in thirteen papyri dated to between 149 bce and 11 ce: UPZ ii.175 (146 bce, Thebes); UPZ ii.177 (136 bce, Thebes); P.Giss. i.36 (134 bce, Pathyris); P.Tebt. i.164 (112 bce, Kerkeosiris); BGU iii.1002 (55 bce, Hermopolis Magna); PSI v.549 (41 BCE, Oxyrhynchus); BGU xvi.2594 (8 bce, Chennis); CPR xv.1 (3 bce, Soknopaiou Nesos); CPR xv.2, CPR xv.3, CPR xv.4, SB i.5231, SB i.5275 (11 ce, Soknopaiou Neso).

24 See Cicero, De or. 1.155: ‘Afterwards I resolved . . . to translate freely Greek speeches of the most eminent orators. The result of reading these was that, rendering into Latin what I had read in Greek, I not only found myself using the best words . . . but also coining by analogy certain words such as would be new to our people, provided only they were appropriate’ (trans. Sutton and Rackham, LCL) and also Plautus, Asin. 11; Poen. 984; Jerome, Ep. 57. For criticisms of interpretari or literal translation, see Cicero, Fin. 3.15, 3.35; Jerome, Ep. 57. Cf. Jewish translational fidelity in Philo, Mos. 2.25–44. For discussion of translation theory and its colonialistic impulses, see C. Moatti, ‘Translation, Migration, and Communication in the Roman Empire: Three Aspects of Movement in History’, ClAnt 25 (2006) 109–40.

25 Only one hundred papyri, ostraca and inscriptions from Egypt contain references to the world of interpreters (hermēneis), leading many papyrologists to note their poor attestation in the documentary record. See, for example, Roger Bagnall's statement that P.Coll. Zen. ii.63 is ‘a good example of the reticence of the documents’, in R. S. Bagnall, Egypt in Late Antiquity (Princeton: Princeton University Press,1993) 233–5, at 233. See also R. S. Bagnall, C. Helms and A. M. F. W. Verhoogt, Documents from Berenike, vol. ii: Texts from the 1999–2001 seasons (Brussels: Fondation Égyptologique Reine Élisabeth, 2005) 27. For a survey of the epigraphic and inscriptional evidence, see C. Wiotte-Franz, Hermeneus und Interpres: Zum Dolmetscherwesen in der Antike (Saarbrucker Studien zu Archäologie und alten Geschichte 16; Saarbrücken: Saarbrücker Druckerei und Verlag, 2001). For the most recent analysis of the variety of uses of the technical terminology in the documentary records, see Mairs, ‘Hermēneis’, 1–53.

26 The 2007 publication of an inscription on a limestone pedestal from Colossae referring to ‘Markos, son of Markos, chief interpreter and translator for the Colossians’ attracted a great deal of attention when it first appeared. See A. H. Cadwallader, ‘A New Inscription, A Correction and a Confirmed Sighting’, EA 40 (2007) 112–18. Cadwallader suggested that Markos was the head of a formal bureau of interpreters. Chaniotis suggested that Markos was an interpreter of oracles (A. Chaniotis, ‘Epigraphic Bulletin for Greek Religion’, Kernos 23 (2010) 285–6), while Bornkamm argues that the title was honorific and that the specific translational needs were primarily legal. See L. Bormann, ‘Barbaren und Skythen im Lykostal? Epigraphischer Kommentar zu Kol 3:11’, Epigraphical Evidence Illustrating Paul's Letter to the Colossians (ed. J. Verheyden, M. Öhler and T. Corsten; WUNT 411; Tübingen: Mohr Siebeck, 2008) 193–6. Similar discussions involve the identity of the named hermēneis Apollonos, who appears in a cluster of Egyptian papyri (P.Cair.Zen. i.59065, P.Ryl. iv.563, PSI iv.409), in which he is called illiterate and appears to have acted as a broker.

27 See Apollonos in P.Cair.Zen. i.59065, P.Ryl. iv.563, PSI iv.409.

28 R. Mairs, ‘“Interpreting” at Vindolanda’, Britannia 43 (2012) 1–12.

29 See, for example, P.Oxy. ii.237, in which the Greek-named Dionysia, daughter of Chaeremon, requires an interpreter in order to participate in a legal dispute over a will. Eleven documentary papyri refer to the fact that one party or another spoke ‘through an interpreter’ (δι ̓ ἑρμηνέως) in a legal case. The earliest extant example is SB xviii.13156, which dates to the second century ce (at the earliest). The most remarked-upon examples are P.Oxy. ii.237 (above) and P.Coll. vii.175. A number of these examples involve the verbal translation of Greek letters into spoken Egyptian, discussed in B. Kelly, Petitions, Litigations, and Social Control in Roman Egypt (Oxford: Oxford University Press, 2011) 179–80. Women seem to be over-represented in these sources, so it is worth noting that the groups of monks in P.Oxy. lxiii.4397 ‘do not know letters’. Compare here the Sayings of the Desert Fathers, in which the Syriac monk Apa Poimen does ‘not know Greek and no interpreter could be found’. Text in M. Chaîne, Le manuscrit de la version copte en dialecte sahidique des ‘Apophthegmata Patrum’ (Cairo: Imprimerie de l'Institut Français d'Archéologie Orientale, 1960) no. 188. See also M. Choat and R. Yuen-Collingridge, ‘A Church with No Books and a Reader Who Cannot Write: The Strange Case of P.Oxy. 33.2673’, Bulletin of the American Society of Papyrologists 46 (2009) 109–38.

30 Mairs, ‘Hermēneis’, 8–9.

31 P.Kell. i.53 (4th cent.) provides a list of expenses for the production of at least one written document. Among those elements referenced are the relatively low costs of making a translation. Here we should understand that a slave was contracted from their owner for the work.

32 I draw here upon L. Venuti, The Translator's Invisibility: A History of Translation (London/New York: Routledge, 1995), in which Venuti describes the way that the power-laden act of translation renders itself invisible in order to obscure the ways in which it interprets and colonises other texts. In the case of ancient constructions of translation, it would seem that the translators themselves, as part of a class of enslaved or servile literary workers, are eliminated. This is not to say that ancient people did not worry about the slippage between translation/interpretation/misinterpretation. Not only is there considerable linguistic overlap between interpretation and translation, but original documents were sometimes produced in court order to mitigate fears of ‘misinterpretation/mistranslation’ in translated texts. See, for example, P.Oxy. xviii.2187.

33 On reading culture, see the groundbreaking W. A. Johnson, Reading and Reading Culture in the High Roman Empire: A Study of Elite Communities (Oxford: Oxford University Press, 2010).

34 See A. Palmer, ‘Egeria the Voyager’, Travel Fact and Travel Fiction: Studies on Fiction, Literary Tradition, Scholarly Discovery and Observation in Travel Writing (ed. Z. von Martels; Leiden: Brill, 1994) 39–53, at 49. Mark does not seem to have served the kind of quasi-touristic function of the interpreter who translated inscriptions on monuments for Herodotus.

35 O.Berenike ii.121, 113–17 ce: ‘On behalf of the fortune of Imperator Caesar Nerva Trajan Augustus Germanicus Dacicus and all his house, to Isis the very great goddess … (son of?) Papiris, interpreter and secretary (Παπείρεος ἑρμηνεὺς καὶ γρα[μματεύς) … under Marcus Rutilius Lupus, prefect of Egypt’ (trans. Bagnall, Helms and Verhoogt, Documents from Berenike, no. 121).

36 H. S. Gehman, The Interpreters of Foreign Languages among the Ancients: A Study Based on Greek and Latin Sources (Lancaster: Intelligensia, 1914) 17–18. As with other slave professions such as wet-nursing, there was a suspicion of foreigners and a preference for home-grown translators (see, for example, the story in Plutarch, Sertorius in which Sertorius decrees that children of good birth be trained as translators) but this discourse was probably a response to the practice. That interpreters, even those sent on diplomatic missions, were sometimes subjected to violence, even by those who dispatched them, is more understandable if the interpreters were of low status. On one occasion the Greek translator at the Persian embassy in Athens was put to death for daring to put the demands of barbarians into the Greek language (Plutarch, Themistocles 6.2). Among Roman examples, the names of most official diplomatic interpreters are lost to us. Seneca distinguishes between translators and authors writing when he remarks that those who carry around the work of others are ‘never auctores, but always interpretes, in the shadow of others’ (Ep. 33.8).

37 T. Y. Mullins, ‘Papias on Mark's Gospel’, VC 14 (1960) 216–24. Mullins’ argument is unpersuasive because it is highly unlikely that Peter was literate. His solution that the Gospel was part literary translation and part recollection does solve the problem of the strange use of ἔνια in the passage, which might be read as suggesting that only some of Mark was written in order or was dependent on Peter. Westcott interpreted it as meaning that Mark was a ‘memoir of “some events”’ in the life of Jesus (B. F. Westcott, A General Survey of the History of the Canon of the New Testament During the First Four Centuries (London: MacMillan, 1875) 75) while Zahn reasoned that Mark was dependent on Peter only for some portions of the Gospel (T. Zahn, Introduction to the New Testament, vol. ii (New York: Charles Scribner's Sons, 1909) 440–1).

38 As Kok puts it: ‘Grammatically, Papias is unclear about whether the aorist participle γενόμενος (having become) is to be read as prior to, or concurrent with, the aorist verb ἔγραψεν (he wrote) (Hist. Eccl. 3.39.15). Either the evangelist was Peter's interpreter before beginning to write what he (Mark) remembered of Peter's words or he acted as Peter's interpreter and scribe simultaneously, transcribing what he (Peter) remembered’ (Kok, Gospel on the Margins, 117).

39 Following Howley, I do not distinguish between the work of the secretarius or amanuensis from the notarius because the work of these literary slaves often overlapped: Cicero refers to Diphilus as the scriptor et lector of Crassus (Cicero, De or. 1.136). See discussion in J. Howley, ‘In Ancient Rome’, Further Reading (ed. M. Rubery and L. Price; Oxford: Oxford University Press, 2020) 15–27.

40 See R. Darnton, ‘What is the History of Books?’, Daedalus 111 (1982) 65–84; R. J. Starr, ‘Reading Aloud: Lectores and Roman Reading’, CJ 86 (1991) 337–43; T. Habinek, ‘Slavery and Class’, A Companion to Latin Literature (ed. S. Harrison; Oxford: Blackwell, 2005) 385–93; R. Winsbury, The Roman Book (London: Duckworth, 2009) 79–85; J. Howley, ‘In Ancient Rome’, 15–27.

41 See N. Horsfall, ‘Rome without Spectacles’, G&R 42 (1995) 49–56; M. McDonnell, ‘Writing, Copying, and Autograph Manuscripts in Rome’, CQ 46 (1996) 469–91. At a certain point in their lives diminished eyesight would have made most elites unable to read without the aid of a lector.

42 A rare exception to this rule might be the relationship between Epictetus, a freedman, and his student Arrian, a future consul, whose notes on Epictetus’ lectures form the basis for the Dissertationes and Encheiridion. Some have challenged the extent to which the Dissertationes is based on Epictetus’ actual teachings, but in either case Arrian claims a scribal relationship to Epictetus. See T. Wirth, ‘Arrians Erinnerungen an Epiktet’, MH 24 (1967) 149–89, at 172 and P. A. Brunt, ‘From Epictetus to Arrian’, Athenaeum 55 (1977) 19–48.

43 Attributing a particular speech to Lysias made collections of speeches more valuable to fourth-century booksellers. On the matter of authorship in Lysias, see T. N. Winter, ‘On the Corpus of Lysias’, CJ 69 (1973) 34–40. Rare counter-examples of collaborative authorship might include the Passion of Perpetua and Felicitas but in examples such as this the authors do not collaborate with one another but, rather, the redactor links his texts to those of his subjects.

44 E.g. Cicero, Fam. 16.3, 4, 7, 8, 9, 11. Cicero refers to co-operative letters in Att. 14.5.1 and there are, of course, Christian examples (e.g. Acts 15; 1 Clement; Polycarp, Phil.; Martyrdom of Polycarp). Murphy-O'Connor reports that ‘only 6 out of the 645 papyrus letters from Oxyrynchus, Tebtunis, and Zenon had a plurality of senders’ (J. Murphy-O'Connor, ‘Co-Authorship in the Corinthian Correspondence’, RB 100 (1993) 562–79, at 564).

45 See discussion in A. M. Riggsby, Caesar in Gaul and Rome: War in Words (Austin, TX: University of Texas Press, 2006) 134–5. In some instances, the self-description of a text as unpolished could easily have functioned as a form of faux humility and a means of accounting for the presence of errors in a text.

46 Larsen, Gospels before the Book, 11–36.

47 Compare, for example, the arrangement between Jesus and Thomas in the preface to the Gospel of Thomas in which Jesus ‘speaks’ and Thomas ‘writes’.

48 Bauckham, Jesus and the Eyewitnesses, 211. The extent to which a secretarius acted (in modern but not ancient constructions of authorship and textual production) as a collaborator is worth discussing, but this is not in Bauckham's purview.

49 There are rare references to elites taking dictation or editing the work of friends, but this seems to have happened in exceptional circumstances; see discussion in McDonnell, ‘Writing, Copying, and Autograph Manuscripts’, 477. The example he cites is Cicero recruiting senatorial scribes during the Catilinarian conspiracy, a situation in which secrecy was required.

50 Tiro is something of a celebrity, but the vast majority of scribal figures go unnamed and uncredited. For hints of other important slave-workers, see Isocrates, Panath. 231; Horace, Sat. 1.4.9–13; Quintilian, Inst. 10.3.

51 Cicero, Fam. 16.10.2 and 16.4.3.

52 We might compare here the treatment of Tertius who is named in Rom 16.12 as the one who wrote the letter to the Romans. Scholars go to great lengths to avoid describing him as enslaved. See, for example, Deissmann, who describes him as an ‘associate’ (A. Deissmann, St Paul: A Study in Social and Religious History (London and New York: Hodder and Stoughton, 1913) 225). Others call him a ‘scribe’, a term that, in the context of Judeo-Christian religion, confers intimations of both liberty and education. See, for example, the brilliant K. Haines-Eitzen, Guardians of Letters: Literacy, Power, and the Transmission of Early Christian Literature (Oxford: Oxford University Press, 2000) 79.

53 There is a debate about precisely whose memories are being recorded in this passage. Grammatically the person who is doing the remembering could be either Peter or Mark. I follow here C. Keith, The Gospel as Manuscript: An Early History of the Jesus Tradition as Material Artifact (Oxford: Oxford University Press, 2020) 52–3.

54 Goodspeed, for example, writes that ‘when Peter preached in his native Aramaic to little companies of Roman Christians, [Mark] stood at his side to translate his words… he had become so familiar with Peter's preaching, through his practice of translating it, that it was possible for him to remember and write down much that Peter had been wont to tell about his walks and talks with Jesus’ (E. J. Goodspeed, The Story of the New Testament (Chicago: University of Chicago Press, 1929) 49–50). Goodspeed's analysis reads both the habits of Jesus and Papias’ description of Peter and Mark's relationship through the lens of early twentieth-century intellectual discourse.

55 I am here discussing the construction of memory in the ancient world. Some recent scholarship on the Gospels has utilised the work of social memory theorists to discuss the manner in which gospel traditions preserve accurate eyewitness traditions. See particularly Bauckham, Jesus and the Eyewitnesses; M. Bockmuehl, Seeing the Word: Refocusing New Testament Study (STI; Grand Rapids: Baker Academic, 2006) 166–78. For an excellent overview of social memory theory and argument that the theory should not be used to discuss historical accuracy or inaccuracy, see C. Keith, ‘Social Memory Theory and Gospels Research: The First Decade. Part One’, Early Christianity 6 (2015) 354–76 and ‘Social Memory Theory and Gospels Research: The First Decade. Part Two’, Early Christianity 6 (2015) 517–42. It is not the intention of this project to establish the accuracy or inaccuracy of early Christian testimony, rather I am interested here in how ancient audiences thought about memory.

56 See, for example, Raffaella Cribiore's statement that ‘memory was the foundation of all knowledge in a world that could not rely on easily consulted books, tables of contents and indexes, library catalogues and electronic search tools’, in her ground-breaking Gymnastics of the Mind: Greek Education in Hellenistic and Roman Egypt (Princeton/Oxford: Princeton University Press, 2011) 166. See also T. Morgan, Literate Education in the Hellenistic and Roman Worlds (Cambridge: Cambridge University Press, 1998) 23–33, 45–6. For the view that Jesus had his disciples memorise his teachings, see B. Gerhardsson, Memory and Manuscript: Oral Tradition and Written Transmission in Rabbinic Judaism and Early Christianity (Grand Rapids: Eerdmans, 1998) 240–4 and Tradition and Transmission in Early Christianity (Lund: Gleerup, 1961) 22. Gerhardsson's analysis is based on rabbinic and Second Temple Jewish practices and assumes, despite considerable evidence to the contrary, that the disciples were well educated.

57 Seneca, Ep. 27.5. See discussion in Winsbury, The Roman Book, 79–85.

58 Pliny, Nat. 28.14. See discussion in O. Patterson, Slavery and Social Death (Cambridge, MA: Harvard University Press, 1982) 339.

59 On memory loss through aging, see Seneca, Con. 10. pr. 1.

60 Seneca, Ep. 27.5 and T. Wiedemann, Greek and Roman Slavery (London: Routledge, 1988) 26. See also Lucian's satirical diatribe The Ignorant Book Collector, which lampoons the efforts of the newly wealthy to acquire status by purchasing knowledge, and Petronius’ Trimalchio in the Satyricon.

61 In an interesting article articulating the relationship between Peter and Mark as that between Rabbi and tannaʾ, Yadin-Israel equates the status of the hermēneutes to that of the tannaʾ, those responsible for preserving traditions by committing to memory. Though I do not think that Papias’ text is as replete with rabbinic terminology as Yadin-Israel's article claims, the parallel is suggestive because the tannaʾ, while reliable repositories of tradition, are portrayed, as Roman literate slaves are, as failing to fully understand the things that they preserve. The parallels between the lower status tannaʾ and literate slaves seem ripe for exploration. See Yadin-Israel, ‘For Mark was Peter's Tannaʾ’, 349.

62 Hengel, Studies, 48–9; Bauckham, Jesus and the Eyewitnesses, 217–21.

63 On this, see A. Stewart-Sykes, ‘Taxei in Papias: Again’, JECS 3 (1995) 489–90 and examples in Thucydides, Hist. 1.97.2; Philostatus, Vit. Apoll. 1.2; Polybius, Histories 5.33).

64 Colson, ‘Tάξει in Papias’, 62–9; Kürzinger, Papias, 13–14. Kürzinger directs us to the technical use of σύνταξιν in Greek rhetoric as a term for structure and literary organisation (e.g. in Ps.-Aristotle, Rhet. Alex. 30–8).

65 Colson, ‘Tάξει in Papias’, 64. See also Larsen, Gospels before the Book, 87. Cf. Philostratus, Vit. Apoll. 1.2–3.

66 Larsen, ‘Accidental’, 377. Larsen's theory applies to the composition of the text itself as well as its reception and construction in the second and third centuries. For our purposes we will argue only the latter; Matthew and Luke certainly treat Mark in this way.

67 Tacitus, Ann. 4.53 (Agrippina, mother of Nero). See Riggsby, Caesar in Gaul and Rome, 134 and M. D. C. Larsen and M. Letteney, ‘Christians and the Codex: Generic Materiality and Early Gospel Traditions’, JECS 27 (2019) 383–415, at 398.

68 Larsen does not discuss legal transcripts in detail, but these kinds of documents were supposed to be especially reliable. On this, see R. A. Coles, Reports of Proceedings in Papyri (Papyrologica Bruxellensia 4; Brussels: Fondation Égyptologique Reine Élisabeth, 1966). On the function of legal documents in the ancient world, see Riggsby, Caesar in Gaul and Rome, 139–40. For the way in which early Christians deliberately appropriate the court record style in order to evoke a sense of proximity to events, see F. Millar, The Roman Republic and the Augustan Revolution (Chapel Hill, NC: University of North Carolina Press, 2002) 35.

69 Keith, Gospel as Manuscript, 52–3.

70 Larsen, Gospels before the Book, 90.

71 Larsen, Gospels before the Book, 90.

72 Martial 14.208.

73 Cicero, Fam. 16.10.2

74 Horace, Sat. 1.10.92.

75 Eusebius, Hist. eccl. 3.39.3–4. L. Alexander, ‘“The Living Voice”: Skepticism towards the Written Word in Early Christian and in Graeco-Roman Texts’, The Bible in Three Dimensions: Essays in Celebration of Forty Years of Biblical Studies in the University of Sheffield (ed. D. J. A. Clines, S. E. Fowl and S. E. Porter; Sheffield: Sheffield Academic, 1990) 221–47; Keith, Gospel as Manuscript, 4–6.

76 Here I mean with respect to the content of dictation and transcription. There are rare examples of enslaved and formerly enslaved authors in antiquity, e.g. Aesop, Phlegon, Hyginus and the narrators of Greek romance novels, but these were usually associated with collections of sayings, paradoxography and other genres of literature deemed déclassé.

77 This contrasts with documentary texts, for which slaves could be responsible. Howley describes it in the following way: ‘Tiro commits more obvious errors the further he moves from his patron Cicero … Tiro's failings … are not mere errors in interpretation, but some more fundamental flaw of character’ (J. A. Howley, Aulus Gellius and Roman Reading Culture: Text, Presence and Imperial Knowledge in the Noctes Atticae (Cambridge: Cambridge University Press, 2018) 182).

78 This example is taken from Larsen and Letteney, ‘Christians and the Codex’, 398.

79 On Mark's lack of authorial status, see Larsen and Letteney, ‘Christians and the Codex’, 398.

80 For later examples of Mark as Peter's interpreter not discussed here, see Tertullian, Marc. 4.5; Jerome, Com. Matt. praef. 2; Cosmas Indicopleustes, Top. 5.196.

81 Irenaeus, Haer. 3.1.1. The Greek text is preserved in Eusebius, Hist. eccl. 5.8.2–4. The text of Irenaeus is preserved in a fragmentary form in Greek and Latin. The texts used by Larsen and followed here are from A. Rousseau and L. Doutreleau, Irénée de Lyon, Contre Les Hérésies. Livre iii (Sources Chrétiennes 211; Paris: Les Éditions du Cerf, 1974) 22–5. For the view that Irenaeus himself did not write this statement, see B. Mutschler, ‘Was weiss Irenäus vom Johannesevangeliums aus der Perspektive seiner Rezeption bei Irenäus von Lyon?’, Kontexte des Johannesevangeliums: Das vierte Evangelium in religions- und traditionsgeschichtlicher Perspektive (ed. J. Frey and U. Schnelle; WUNT 175; Tübingen: Mohr Siebeck, 2004) 705–6. Thornton, the source of Mutschler's opinions, argues that this summary comes from a Roman source composed between 120 and 135 ce. For our purposes it is, however, still an example of early Christian interpretation. See C.-J. Thornton, Der Zeuge des Zeugen: Lukas als Historiker der Paulusreisen (WUNT 56; Tübingen: Mohr Siebeck, 1991) 8–67.

82 Larsen, Gospels before the Book, 19.

83 Irenaeus, Haer. 3.1.1: καὶ Λουκᾶς δέ, ὁ ἀκόλουθος Παύλου, τὸ ὑπ᾿ ἐκείνου κηρυσσόμενον Εὐαγγέλιον ἐν βίβλῳ κατέθετο may well be focused upon Marcion, whom Irenaeus squarely addresses later in this section, and Marcion's claim that he is preserving ‘Paul's Gospel’ (cf. 2 Tim 2.8).

84 So A. Yoshiko Reed, ‘ΕΥΑΓΓΕΛΙΟΝ: Orality, Textuality, and the Christian Truth in Irenaeus’ Adversus Haeresis’, VC 56 (2002) 11–46, at 25. See also M. M Mitchell, ‘Mark, the Long-Form Pauline Εὐαγγέλιον’, Modern and Ancient Criticism of the Gospels: Continuing the Debate on Gospel Genre(s) (ed. R. M. Calhoun, D. P. Moessner and T. Nicklas; Tübingen: Mohr Siebeck, forthcoming).

85 As is the case with Papias and Irenaeus, these stories are refracted through the writings of Eusebius. Clement of Alexandria, fr. 8.4–12 = Eusebius, Hist. eccl. 6.14.6–7 and Clement of Alexandria, fr. 9.4–20 = Eusebius, Hist. eccl. 2.15.1–2.

86 Clement of Alexandria, fr. 24 GCS 17.206 (PG 9.732). See discussion in M. L. Peppard, The Son of God in the Roman World (Oxford: Oxford University Press, 2011) 89–90. According to this text, which is preserved in Latin by the sixth-century monastic founder Cassiodorus, it was when Peter was preaching in Rome in front of imperial equites that some of these petitioned Mark to write.

87 This is explicit in Eusebius, Hist. eccl. 2.15.2. Sonship and emancipation are not necessarily mutually incompatible: the processes of adoption and manumission resembled one another.

88 The locative translation of προγεγράφθαι as ‘write before the public’ or ‘set forth publicly’ (rather than the more usual temporal interpretation ‘write first’) is from S. C. Carlson, ‘Clement of Alexandria on the “Order” of the Gospels’, NTS 47 (2001) 118–25, at 122–3. Audience request traditions are also found in Clement with respect to the composition of the Gospel of John (Eusebius, Hist. eccl. 6.14.7), in the Muratorian Canon and in gospel prologues. See Mitchell, ‘Patristic Counter-Evidence’, 50 n. 42.

89 M. Kennerly, Editorial Bodies: Perfection and Rejection in Ancient Rhetoric and Politics (Columbia, SC: University of South Carolina Press, 2018) 6.

90 On the distinction between private and public reading, see Johnson, Readers and Reading Culture, 55–9; Larsen, Gospels before the Book, 76–7; Keith, Gospel as Manuscript, 171–3.

91 For the Latin prologues to Mark, see J. Regul, Die antimarcionitischen Evangelienprologe (Vetus Latina 6; Freiburg: Herder, 1969) 29–30. For the argument that these prologues serve a unified purpose and character, see D. de Bruyne, ‘Les plus anciens prologues latines des évangiles’, RBén 40 (1928) 193–214. As Regul has elegantly demonstrated, these Latin prologues are not anti-Marcionite, Die antimarcionitischen Evangelienprologe, 266–67.

92 H. F. von Soden, Die Schriften des Neuen Testaments, vol. i (Berlin: Dunckler, 1902) 312.

93 See discussion in R. M. Grant, ‘The Oldest Gospel Prologues’, ATR 23 (1941) 231–45, at 236. See also Hippolytus, Elenchus 7.30, which uses the same phrase. Later traditions claim that Mark deliberately cut off one of his fingers in order to avoid being elevated either to the Jewish priesthood or to the Christian episcopate. A rich description of this tradition is provided by J. L. North, ‘MARKOS HO KOLOBODAKTYLOS: Hippolytus, Elenchus, vii.30’, JTS 29 (1977) 498–507. An excellent overview of various scholarly interpretations can be found in Kok, Gospel on the Margins, 220–6.

94 Grant, ‘The Oldest Gospel Prologues’, 236. Compare Harvey's suggestion that it was a nickname meaning ‘clumsy-handed’, Sancti Irenaei libros quinque adverus haereses, vol. ii (ed. W. W. Harvey; Cambridge: Typus Academicis, 1857) 4 n. 3.

95 E.g. Severian of Gabala, PG 63.541.

96 C. Marvin, ‘The Body of the Text: Literacy's Corporeal Constant’, Quarterly Journal of Speech 80 (1994) 129–49, at 132. See discussion in Kennerly, Editorial Bodies, 13–17. Cf. P.Oxy. lvi.3860.

97 The individual parts of prose are regularly described as fingers and limbs. See Aristotle, Rhet. 1408b–1409a; Cicero, Or. Brut. 149–233.

98 Horace, Sat. 1.10.64–72.

99 A. J. Malherbe, ‘A Physical Description of Paul’, HTR 79 (1986) 170–5.

100 Physiogn. 3.807b5–12. Cf. Soranus’ Gynaecology, which emphasises that the body of the infant must be massaged into proportionality.

101 I. Repath, ‘The Physiognomy of Adamantius the Sophist’, Seeing the Face, Seeing the Soul: Polemon's Physiognomy from Classical Antiquity to Medieval Islam (ed. S. Swain; Oxford: Oxford University Press, 2007) 487–547, at 520–1.

102 For a succinct and illuminating overview of scribal and secretarial practices in the Roman era, see Haines-Eitzen, Guardians of Letters, 27–34. It seems highly unlikely that Mark is being positioned as a ‘town-square’ scribe, since he appears to have been specifically linked to Peter.

103 Perhaps Mark is not credited with authorship but, rather, identified as the textual worker responsible for this accurate version of the gospel. Arguably, this may be the perspective of Irenaeus, who uses the formula ‘the Gospel according to … (τὸ κατά)’ to refer to the work of the evangelists. This formulation conveys, as Yoshiko Reed has so ably argued, ‘not that of a title per se (i.e. ‘The Gospel According to…’), but rather the more literal sense of a single Gospel according to various authors’. See Yoshiko Reed, ‘ΕΥΑΓΓΕΛΙΟΝ’, 20 n. 33. For the view that the titles of the Gospels date to the time of their earliest circulation, see S. J. Gathercole, ‘The Titles of the Gospels in the Earliest New Testament Manuscripts’, ZNW 104 (2013) 33–76 and, ‘The Alleged Anonymity of the Canonical Gospels’, JTS 69 (2018) 447–76.

104 See Kennerly, Editorial Bodies, 205.

105 See 1 Cor 3.4.

106 Aulus Gellius, Noct. att. 1.7.1; 13.21.16–17. See discussion in Howley, Aulus Gellius, 176–7. Tironian Ciceros were sufficiently popular in the second century for forgeries to be produced. See J. Zetzel, ‘Emendavi ad Tironem: Some Notes on Scholarship in the Second Century ad’, HSCPh 77 (1973) 225–43, at 231–2.