1. An Overview of the Secretary Hypothesis

Randolph Richards’ 1991 monograph renewed a debate over the role of the secretary in Paul's letters.Footnote ¹ Particular attention was paid to the issue of an author's writing in his/her own hand;Footnote ² this became a means of demonstrating Paul's consistent use of a secretary (Gal 6.11; 1 Cor 16.21; Phlm 19, cf. 2 Thess 3.17; Col 4.18).Footnote ³ The input of a secretary varied along a spectrum: transcriber, contributor and composer.Footnote ⁴ Richards' primary informant into the mechanics of composition was Cicero.Footnote ⁵

The key accent has been the secretary's contribution to, not merely the transcription of, the letter. Others, such as the co-named senders mentioned at the beginning of some letters (Sosthenes, Silas, Timothy), have joined the secretary (or have become the secretary), in a community of authorship.Footnote ⁶ This explains variations, even jarring interruptions,Footnote ⁷ in vocabulary and style. It is claimed that such factors lose validity when assessing the authenticity of Pauline lettersFootnote ⁸ and in adjudicating interpolations into those letters.Footnote ⁹ Richards therefore operates with a thirteen-letter corpus of Paul's authentic letters,Footnote ¹⁰ making, in Philip Towner's words, ‘the amanuensis hypothesis the default explanation’.Footnote ¹¹

Richards’ thesis has proved highly influential and has been applied beyond the Pauline corpus.Footnote ¹² Objections are rare.Footnote ¹³ The elements are basically repeated by Ian Elmer, although he allows for a greater impact of the Pauline community in the assemblage, shaping and content of Paul's letters for the collection after the apostle's death.Footnote ¹⁴ The personification of this ‘communification’ of the Pauline epistolary enterprise is Tertius. Rom 16.22 has assumed a landmark importance in the Pauline critical enterprise.

It is argued that, unlike most available secretaries, Tertius was not hired from the marketplace.Footnote ¹⁵ He was after all ‘in the Lord’ (ἐν κυρίῳ),Footnote ¹⁶ even if a member of Gaius’ household or the imperial bureaucracy.Footnote ¹⁷ He shaped Paul's thoughts into a consistent letter, though late additions produced some unevenness, especially in chapter 16.Footnote ¹⁸ Not only was Tertius responsible for inserting non-Pauline interpolations – such as Septuagint quotations and pre-formed traditions – where Paul had indicated,Footnote ¹⁹ but he contributed the long greetings list.Footnote ²⁰ This latter initiative was understandable because Tertius was on a trip from Rome, knew more of the membership of the churches there and so was better placed to do the unusual – provide a distinguishing feature of Romans.Footnote ²¹ It is assumed that Tertius had freedom of movement, carrying his tools of trade with him.Footnote ²² He therefore becomes a ‘professional secretary’,Footnote ²³ apparently permitted to make, and self-financing, the journey. Tertius formulated the recommendation for Phoebe that opens chapter 16, working in credentials that Paul suggested.Footnote ²⁴ Tertius supplied crucial information about the state of the churches in Rome that helped to guide the letter's arguments.Footnote ²⁵ Accordingly, he may even have been part of Paul's team,Footnote ²⁶ though, apparently, not significant enough to gain mention elsewhere.

A complex of issues arises: Tertius’ identity, his relationship to Paul, the method and purpose in the construction of ancient writings (whether literary or quasi-literary), and, perhaps most importantly, whether our own concentration has become misplaced in the assessment of the personnel key to the realisation of Paul's intent in Romans.

2. The Identity of Tertius

Speculation about the identity of Tertius has been complicated by the designation ‘professional secretary’, a ‘skilled amanuensis’. Such classifications create an anachronistic impression of a secretary to a company chair, a free agent providing his expertise, or a fanciful status elevation of Tertius almost to that of a royal or civic γραμματεύς. Such an elite position is targeted by Paul in 1 Cor 1.20.Footnote ²⁷ Even allowing that here in the Corinthian correspondence Paul adopts a heightened rhetorical flourish, the antipathy still sits awkwardly with the construction of an elevated standing for Tertius such as to draw him into the company of a co-worker or co-sender, even a co-author (even though nowhere so described by Paul). No rhetorical possibility is entertained that the naming of co-senders in a letter is about raising the profile of the co-named ‘as a fully approved emissary’.Footnote ²⁸ As will be seen, Tertius is probably considerably lower among secretarial types than the interpretation of γραμματεύς as ‘professional’ conveys. As will be seen, the office of γραμματεύς varies considerably in status and position, much as ‘town clerk’ in modern local government is a position far removed from a ‘clerk’ in an accountant's office.

Tertius’ name and greeting are entirely self-initiated and without parallel in the morphology of greetings: not even Paul uses the form ἀσπάζομαι.Footnote ²⁹ The argument that an author stood behind the final product is special pleading,Footnote ³⁰ somehow to imply Paul's knowledge and approval before the letter was sent. In any case, this initiative subverts the argument about Tertius’ co-authorship (more on this below), especially given that the other supposed secretary references in the Pauline letters (Gal 6.11; 1 Cor 16.21; 2 Thess 3.17; Col 4.18; Phlm 19) are quite different from that provided here in Romans 16, precisely because the initiative is Paul's.

The reliance on Cicero's letters for the assessment of Paul's letters generates the impression that Tertius is another Tiro,Footnote ³¹ as adept as his master in the requisite grammar, syntax and vocabulary of a communication.Footnote ³² The accent on ability overlooks two crucial aspects. The first is that Cicero had a team of secretarial assistants. Tiro was certainly his particular assistant – indeed one of his posthumous editors.Footnote ³³ But when the calligraphic aesthetics of the final product were required,Footnote ³⁴ Tiro was not the chosen;Footnote ³⁵ nor, it seems, was Philotimus, one of Cicero's copyists (librarii).Footnote ³⁶ Cicero esteemed beautiful writing,Footnote ³⁷ as did his public: ‘ancients valued beautiful handwriting’.Footnote ³⁸ Tertius may therefore have been simply the writer of beautiful letters (καλλίγραφος).Footnote ³⁹ The second aspect is that Tiro was a slave (a verna, manumitted in 53 bce), as were all Cicero's secretaries. This status is crucial for the assessment of the role of a secretary in the claim – and acknowledgement of the claim – upon authorship.

This is where Tertius’ name becomes important. Elmer had ventured, on the one hand, that ‘[h]is name is Latin, rather than Greek, and is common amongst slaves and freedmen of the period.’ On the other hand, he can assert, ‘the name Tertius is not a Latin name often attested’.Footnote ⁴⁰ With 695 hits for the nominative form ‘Tertius’ in the Clauss-Slaby Epigraphik-Datenbank,Footnote ⁴¹ the first option should be followed. Elmer also suggested that Tertius and Quartus might be ranking names, perhaps in the household of Gaius.Footnote ⁴² He does not want to expunge servility; but the number-naming of slaves is rare.Footnote ⁴³ The greater frequency of such number-naming, whether as praenomina or cognomina,Footnote ⁴⁴ occurs for citizens.Footnote ⁴⁵ The cognomen is particularly common.Footnote ⁴⁶ Quartus as ‘the (not our) brother’ might indicate a lineage (free) or legal (servile) designation.Footnote ⁴⁷ It may even turn the whole of v. 23 into an extension of Tertius’ insertion begun in v. 22.Footnote ⁴⁸ Tertius can also be used as a praenomen, not just a cognomen,Footnote ⁴⁹ which would sway the interpretation towards citizenship, provided the tria nomina were known. When occurring in isolation, the pendulum swings towards servility. Either way, there is nothing in the name that requires a Roman provenance.Footnote ⁵⁰ This leaves us then with weighing the probabilities.

Edwin Judge has provided a series of analyses of the names of those around Paul, with the intent of establishing the social profile of Pauline connections in early Christianity.Footnote ⁵¹ He takes Tertius as a cognomen. Although the cognomen would ‘mostly imply citizenship’, in the case of Tertius (but not Quartus) he reckons that the name might as easily suggest Greek ethnicity and/or servilityFootnote ⁵² – that is, the praenomen of the master becoming the cognomen, indeed the only name, of the individual.Footnote ⁵³ Judge wanted to explore the adoption of Latin names into Greek environments. Τέρτιος scores highly in this regard. The name is found in servile or servile-origin cases in Achaia and surrounding regions.Footnote ⁵⁴ More generally, Iiro Kajanto finds that numbering cognomina, though no longer indicating birth sequence, were frequently given to slaves.Footnote ⁵⁵

The probabilities look evenly balanced even though virtually all secondary commentators lean strongly towards servility.Footnote ⁵⁶ Certainly, if Tertius is a slave, then Quartus is most unlikely to be his fellow name-numbered slave and certainly not his familial relative. The situation might be different however if Tertius is a Roman citizen. The decision therefore cannot be made simply on the basis of onomastics.

The key lies with the implied occupation of Tertius, that is, secretary. This general category actually masks a range of occupations that were usually quite distinct in a household of reasonable wealth, and it obscures the range of functions possible in the designation of γραμματεύς. A γραμματεύς in Corinth might refer to a leading citizen who held high office, a position of considerable prestige, one responsible, inter alia, for key functions of the city gymnasium.Footnote ⁵⁷ But the term might indicate lesser standing, often with a qualifying epithet of those whom the γραμματεύς specifically serves – an association, a village, the military. One unnamed γραμματεύς from Hellenistic Corinth was the (free) clerk to the δικασταί.Footnote ⁵⁸ Γραμματεύς can cover multiple yet separate writing functions, from the recording of land registrations to the preparation of letters.Footnote ⁵⁹ S/he may supervise a scribal staff, often servile, who had quite narrow functions.Footnote ⁶⁰ A γραμματεύς might also be among a band of slaves responsible for secretarial duties in a household. In P.Oxy. xliv.3197, dated 111 ce, the testamentary division of slave property designates some slaves by secretarial task: one is a γραμματεύς, two are προχειροφόροι (amanuenses),Footnote ⁶¹ five are νωτάριοι.Footnote ⁶² The terminology is significant as an indication of the highly specific breakdown of tasks to slaves.Footnote ⁶³ But the proportions are also telling. The ratio of νωτάριοι to γραμματεύς suggests that the latter was probably the one responsible for the final writing up or approval of whatever document had been dictated, even if this were not the only input on the production line that a document might receive. Indeed, the actual papyrus document indicates that considerable ability (‘competent professional cursive’) went into its execution (up to l. 19).Footnote ⁶⁴ The hands of the three beneficiaries by contrast are noticeably inferior (ll. 20–2).Footnote ⁶⁵ One of the slave notaries probably wrote up the document; the finish suggests that it was formally executed from a draft. All, however, are slaves, even if, as frequently, there is a hierarchy.Footnote ⁶⁶ Some form of hierarchy in secretarial positions is also evident even where servility is unclear.Footnote ⁶⁷ One of the responsibilities for these fellow-secretaries may have been to witness as well as execute documents.Footnote ⁶⁸ This suggests that when we look to the various secretarial functions, the default position is that they will be executed by slaves, whether public, private or even for hire,Footnote ⁶⁹ and that hierarchical layering governs the dispersal and execution of their functions. Tertius is to be located in this context.

Whatever function we may want to confer on Tertius from this evidence, he designates himself the writer of the letter (ὁ γράψας τὴν ἐπιστολήν (ἐν κυρίῳ)), that is, putting himself at the end of the production line, not in the prior stages. This increases the probability of Tertius being a slave, in part because there is evidence that his creative ability in relation to expression, as distinct from letter-forming, is limited – the ambiguity caused by the syntactical placement of ἐν κυρίῳ does not inspire confidence that Tertius was responsible for chapter 16 or any other part of the letter.Footnote ⁷⁰ Further, and this is a point not acknowledged by those who emphasise his servility, it is most unlikely that he would be a traveller from Rome. Without further, and explicit, qualification we would find ourselves thrust on the horns of the Onesimus dilemma, confronting Paul and/or a householder with a fugitive slave. More likely, Tertius resided in Corinth or perhaps nearby Cenchreae, as a slave in the household of Stephanas, Chloe, Phoebe or some other.

3. The Presence of Tertius in the Text

In ancient manuscripts and print editions of Romans, Tertius is part of the text along with every other person in the greetings. That is, what we now designate v. 22, follows immediately upon the end of v. 21 and precedes v. 23, seamlessly included in the lines of Romans that yield the stichoi enumeration. On this basis, Tertius was always there, in the body of Paul's letter. The presence of variants to the text of v. 22, as early as ⁴⁶,Footnote ⁷¹ demonstrates that the verse was included in the main body from an early period – at least from the late second century – and subject to familiar scribal activity.

The most cited example of the intrusion of a third person into a text is Alexis, as reported by Cicero.Footnote ⁷² This is different from the recognition of a second or third hand, especially since suchlike is usually the author adding his/her recognisable writing to the end of the letter (as in Gal 6.11; Phlm 19; 1 Cor 16.21).Footnote ⁷³ Often, the verso reveals a different hand(s) providing the sender, and recipient's name and address.Footnote ⁷⁴

All commentators have approached v. 22 as belonging to the main text, even when Paul is given the credit for writing one or other subscript,Footnote ⁷⁵ or the whole of chapter 16.Footnote ⁷⁶ The problem generated when v. 23(–4) is taken as coming from Paul rather than tied to Tertius’ salutation is that Paul's greetings appear to be interrupted. One might expect the ἀσπάζεται of v. 21 and v. 23 to flow without hiatus, without disruption by Tertius’ first-person ἀσπάζομαι. The move between the first and third person forms is common enough,Footnote ⁷⁷ especially when there is an opening greeting,Footnote ⁷⁸ but not with a change of hand. One third-century letter moves from ἀσπάσασθε to ἀσπάζεται in the original hand and is then followed by another hand, writing ἀσπάζου Ἰσιδώραν καὶ τὰ τέκνα αὐτῆς before the farewell formula.Footnote ⁷⁹ Another letter, dated to 118 ce, has the author taking up the pen to write the formulaic farewell (and date) and a third hand joining his greetings – Ἱερακίων ἀσπάζομαί σε, γλυκύτατε – at the foot of the page after the letter ending.Footnote ⁸⁰ Sometimes the first person is used of someone other than the author, but apparently at the same time as the letter was being written by a scribe.Footnote ⁸¹ Clearly, there is no single way in which the addition might have been crafted. But what is also clear is that, even when different hands are noticeable in the body of papyrus letters, a self-identified scribe, different from the author, is not among them.

The structure of chapter 16 is uneven and, towards the end, uncertain. Verses 25–7 are vagrant in the manuscript tradition.Footnote ⁸² While the end of chapter 16 is their usual fixed address, they have been known to turn up after 14.23, 15.33. Sometimes duplicated, split,Footnote ⁸³ shortened, the verses have attracted their own variant readings or been dropped altogether. Stability is not the most notable feature of chapter 16's text.

4. A Para-textual Tertius?

On the basis of the arrangement of greetings and conclusions found in ancient letters, neatness is not always their most compelling feature either, even when a cultivated hand has produced the substance of the letter. It is only the reproduction of these texts, in the aftermath of the final form of a letter, that provides the orderly arrangement. This raises the question of whether the greeting from Tertius, now our v. 22, was as neatly contained in the original letter as our modern texts and ancient manuscripts construct.

We know from numerous instances in the New Testament how material in the margin, often dubbed a ‘scribal gloss’, can find its way into the text in the course of transmission.Footnote ⁸⁴ Marginal material can originate in different ways. Codex Claromontanus (D^p), given that there was space on the line after the existing text of Rom 8.1, simply made an addition to the text in smaller letters. Most scribal glosses began life as a small-lettered horizontal comment adjacent to the relevant section of the main text. It was an ancient practice.Footnote ⁸⁵

Here the phenomenon observed by Margit Homann is particularly relevant. Following a ‘literary’ mention in a letter of Cicero,Footnote ⁸⁶ she has called it versiculus transversus, that is, writing perpendicular and marginal to the main text, executed by turning the papyrus 90°. The 221 Latin and Greek papyri and ostraca that she has found displaying the feature span the first to fifth centuries (most frequently in the second to fourth centuries), with provenances from Oxyrhynchus to the Wadi Fawakhir. All but five use the left-hand margin.Footnote ⁸⁷ A majority (98) have only one line of writing; 75 have two rows; 1 example provides seven lines. There is no apparent restriction on the content in the margin. Homann has found continuations of the substance of the main text, signatures, prayers, addresses, return addresses, thanksgivings, invitations, general closing formulae and, most important for this study, greetings, sometimes in continuation of greetings begun in the main column of writing.Footnote ⁸⁸ As variable as is the range of material contained in the versiculus transversus so also are the types of letters to which it is attached.

A good example – curiously missing from Homann's listFootnote ⁸⁹ – is the papyrus BGU ii.423, a letter from a new army recruit, Apion, to his father Epimachos. The conclusion follows the usual pattern, containing final greetings (ll. 18–20) and a few additional matters by way of postscript (ll. 21–4) – all written in the usual horizontal fashion across the recto. But then, there is an additional greeting in the same hand written vertically along the left margin: ἀσπάζεταί σε Σερῆνος ὁ τοῦ Ἀγαθοῦ [Δα]ίμονος [καὶ . . . .]ς ὁ τοῦ [. . .]ρου̣ καὶ Τούρβων ὁ τοῦ Γαλλωνίου καὶ Δ[. . .]νᾶς ὁ τ[οῦ . . .]σεν[. . .][. . . .] .[ . . .] .[ -ca.?-], ‘Serenos the son of Agathodaimon greets you and [. . .], and Tourbon the son of Gallonios, and D[. . .]nas the s[on of][. . . .’ The greeting clearly comes as an afterthought and, with space gone from the horizontal lines, was added vertically. The mode of greeting shifts from the second person singular imperative, ἄσπασαι, in the vertical text body – that is, a greeting coming from the writer of the letter through the recipient. In the margin, it becomes the third person singular indicative, ἀσπάζεται (‘[Serenos] greets’) – that is, a greeting that the author of the letter is passing on from another.Footnote ⁹⁰ This letter does not have the first person, ἀσπάζομαι, as in Rom 16.22, but this is common in other letters, with or without inclusions of other forms, including the usual one of chapter 16, ασπάσασθε.Footnote ⁹¹ Of course, the number of lines that could be written transverse varies according to the size of the lettering and width of the margin, whether of a single-sheet letter or the last column of a larger sheet glued with previous sheets to form a scroll (as would, most likely, have been the case for Romans).

Tertius’ greeting is ten words long – forty-nine lettersFootnote ⁹² – easily accommodated vertically into a one-centimetre margin. One second-century letter, PSI viii.943, contains a greeting of seven words, thirty-one letters, written into one transverse line of a one-centimetre left margin: ἀσπάζονταί σε πάντες οἱ ἐν οἴκῳ πολλά. A third-century letter's margins easily accommodated two lines of text (P.Oxy. vi.937), the first line having ten words (fifty letters) with a second line dispensing the farewell, ἐρρῶσθαί σε εὔχομαι. If, for some reason, these original letters came to be copied, as happened with Romans, one can readily understand how the marginal greeting could enter the body of the letter, probably inserted into the other greetings at some point if they were present. Of course, this assumes that in such a copying, care would be taken to reserve space for reassignment into the body of the letter.

We have very few examples of secretaries adding their greetings to a letter from another. (The list, to my knowledge, is Cicero, Att. 5.20, P.Giss.Apoll. 18, P.Oxy. xlix.3505 and perhaps P.Oxy. xlii.3057.) The crux is where such an entry might have been made. No manuscripts give any indication that Tertius was originally in the margins, but New Testament manuscripts with material now adjudged to come from marginal glosses usually provide no indication as to the original placement of the material so incorporated. Given the notice from Cicero in the mid-first century bce and the witness provided by papyri from the first or first-to-second centuries ce onwards,Footnote ⁹³ the versiculus transversus was a practice employed by secretaries around the time of the writing of Romans. My argument is that a reasonable case can be made that the margin was the place in the original letter to the Romans where Tertius wrote his greeting. An early copyist working from the original pulled Tertius into the main body of greetings either without much thought as to the relative placement or to ensure that Paul's own words held the final place in the letter (that is v. 23, perhaps with a benediction (v. 24)).Footnote ⁹⁴

5. Tertius’ Relationship to Paul

Apart from v. 22, no other information is available that might fill out the question of Tertius’ relationship to Paul. Some suggest that this unique intrusion ‘must … signal a close relationship to Paul and a significant contribution to the writing of this letter’.Footnote ⁹⁵ One wonders, if this was the case, why there was no corporate greeting, ἀσπαζόμεθα, as we find in some papyri even in the context of an individual's letter.Footnote ⁹⁶

One phrase in Tertius’ interruption to the text stands out: ἐν κυρίῳ.Footnote ⁹⁷ Scholars have selected two key aspects of the wording. On the one hand, the phrase, in isolation, is placed in tandem with identical occurrences throughout chapter 16 (that is, at vv. 2, 8, 11, 12 (bis), 13). James Dunn asserts that Tertius is following Paul's pattern;Footnote ⁹⁸ if Paul himself handwrote the greetings of chapter 16, that would make Paul's pattern very clear!Footnote ⁹⁹ Richards, on the other hand, argues, on the basis of the brevity of the greetings in other letters of Paul, that ‘Paul instructed Tertius to greet the leaders of the church in Rome’.Footnote ¹⁰⁰ He appears to suggest that not merely the naming but the general pattern of naming was Tertius’ contribution. I have already argued that the use of the first-person singular possessive pronoun countermands this argument. But what is also neglected at both poles of these discussions is that the phrase is almost uniquely Pauline in the New Testament (in its discrete form occurring forty-two times; elsewhere only in Rev 14.13). Given that no one credits Tertius with being the secretary for Paul's other letters, it is clear that the phrase ἐν κυρίῳ is Paul's own expression.

But this raises the second issue. It is sometimes acknowledged that the meaning of the phrase for Tertius’ whole greeting is ambiguous.Footnote ¹⁰¹ It might be taken, from its proximity to ὁ γράψας τὴν ἐπιστολήν, to modify the manner of Tertius’ writing, marking Tertius’ contribution as Christian service. Alternately, it might be taken as an adverbial phrase qualifying ἀσπάζομαι. This is the preference of most,Footnote ¹⁰² even though, on closer investigation, this is not its usage in previous verses. The pronounced syntactical awkwardness in v. 22 therefore undercuts any suggestion that Tertius was responsible for the refinement of Paul's language through the letter, let alone becoming a contributor to the argument. But it also implies that Tertius is imitating a Pauline catchphrase, even if somewhat crude in execution or an afterthought tacked on to his own main intent, namely, salutation. Further, the general lack of ambiguity in the previous instances of chapter 16 strengthens the notion that Paul did not see Tertius’ intrusion. Given the repeated accent on corrections that are made to drafts, especially of important writings (whether letter or other genre), Paul's knowing retention seems less likely.

The effort to reconstruct Tertius’ relationship to Paul is constrained by the limited evidence. However, efforts to bring Tertius into a contributing (as distinct from copying) role in Paul's letter to the Romans, or at least the greetings list of the letter, falter on two scores. First, the phrase ἐν κυρίῳ is distinctly Paul's language – an evocation of common interest and relationshipFootnote ¹⁰³ – which seems to be somewhat ineptly imitated in Tertius’ addition. Secondly, and consequentially, the possibility should be entertained that Paul (like Cicero's friend, Atticus) did not know of, and therefore had no occasion to approve of or correct, Tertius’ self-revealing supplement. This demands consideration of the construction of the letter to the Romans.

6. The Construction of Ancient Writings

The presenting difficulty is that the ‘non-literary naiveté’ of thousands of Egyptian letters do not match the sustained argumentation of Paul of Tarsus.Footnote ¹⁰⁴ There are some exceptions – P.Ammon 1.3, for example, approximately matches the length of Galatians.Footnote ¹⁰⁵ Paul's letters are demandingly long, with Romans leading the burden of concentration, all 7,111 words. The emperor Claudius’ rescript to the Alexandrians (P.Lond. vi.1912) contracts to 784 words. When Pliny the Younger half-apologised to his correspondent Quadratus for the length of a letter, the number of words came to 983. Normally, Pliny recommended brief letters (and poems) – to avoid irritation – though he allowed that friendship demanded longer correspondence, as also the importance of a subject.Footnote ¹⁰⁶

Such comparisons invite consideration of the mechanics of construction. Most Pauline commentators assume a simple direct transaction between Paul and an increasingly able Tertius, now often equipped with shorthand.Footnote ¹⁰⁷ Pliny fortunately provides a number of insights into the composing process – his own and that of others.Footnote ¹⁰⁸ He regarded himself as a lesser light than the more prolix and witty Cicero,Footnote ¹⁰⁹ but for all that, his contributions may be the more pertinent.Footnote ¹¹⁰ His compositions may be letters, but may also include other works, even poetry. A tight genre-restriction that confines us only to letter composition fails to recognise that, for the ancients, while there is a recognition of genre distinctions, this does not overly vary the mechanics of composition of a writing of some length. Pliny admits as much in his famous letter on the eruption of Vesuvius.Footnote ¹¹¹ A model extracted from Pliny will not provide the exact method behind the construction of Romans because of the variables but it will provide at least a sense of the probabilities and afford us a better means of assessing the contribution of Tertius.

There were a number of steps between the initial conception and the final product released for an audience. Multiple wax tablets are frequently mentioned. They were Pliny's regular companions at a hunt. He also sought out ‘paper’ (charta) to use.Footnote ¹¹² So it seems that at least for a member of the elite, access to the requisite materials for writing, even at a preliminary level, was not the problem (cf. 2 Tim 4.13).

This immediacy of tactile contact was not Pliny's preferred method, at least when he was not mundanely scribbling off ‘unliterary productions’ (scribo plurimas sed inlitteratissimas litteras).Footnote ¹¹³ Rather, he preferred the shuttered dark of his room, where the exactitude of every word could be weighed – ‘I read every detail so that I might correct everything.’Footnote ¹¹⁴ When ready, he summoned a notarius (perhaps skilled in tachygraphy, but this is not clear) for dictation. He then returned to his room and repeated the whole procedure for his next writing assignment.Footnote ¹¹⁵

The foundation begun, Pliny seems often to have worked on drafts, including, sometimes, of his letters. This usually involved additions as well as refinements, often informed by interactions with established literature by incorporation or adaptation.Footnote ¹¹⁶ But crucial to the process was engagement with friends: ‘This is my practice, that before I convey my work to the public, I submit it to the judgement of my friends.’Footnote ¹¹⁷

There is no mention of secretaries to whom Pliny turns for critical assessment. Rather secretaries serve the process. In one letter Pliny speaks of gathering a group of friends together for the appraisal of one of his drafts. Secretaries supplied copies of the writing on which friends were to add their notes (adnotanda) as the work was read.Footnote ¹¹⁸ The process might be repeated.Footnote ¹¹⁹ Ultimately, he retained the authority of his own judgementFootnote ¹²⁰ just as he was wary of overrevising.Footnote ¹²¹ Whatever the involvement of others, there is no doubt that just as style was a key element in recognising an individual authorFootnote ¹²² – whether Cicero, Demosthenes or whomever – Pliny also wanted to ensure his own future reputation by claiming sole authorship of a well-crafted piece of writing. Quality inscribed the person, not merely the text (cf. 2 Cor 10.10).Footnote ¹²³

This introduces an aspect of the production line that is frequently overlooked, namely, the crucial importance of the work passing oral muster. Without a reading of the work, it is clear that Pliny felt that the revision was incomplete – historians, tragedians and poets did the same.Footnote ¹²⁴ Sometimes, he sought out one who had a reputation as a good reader for the performance (even a freedman of his own household)Footnote ¹²⁵ hovering over a copy to make corrections.Footnote ¹²⁶ Some masters kept a lector in the household,Footnote ¹²⁷ not merely to hear the latest letter or literary purchase, but to listen to their own writing.Footnote ¹²⁸ Again, a reader, like a secretary, was not conferred co-authorship. But he could alert a writer to oral infelicities as well as brilliance.