¹ On Aristo, whose name repeatedly crops up in the Digest, see Kunkel, W.Herkunft und soziale Stellung der römischen Juristen (1952)Google Scholar, 141–4, 318–21; Sherwin-White, A. N.The Letters of Pliny. A Historical and Social Commentary (1966), 136–7Google Scholar; Syme, R. ‘Correspondents of Pliny’, Historia 34 (1985), 324–59Google Scholar, at 352 (= R.P. 5.469); Birley, A. R.Onomasticon to the Younger Pliny (2000)CrossRefGoogle Scholar, 92.

² The Fasti Ostienses date Dexter's death to 24 June 105 ( Vidman, L.Fasti Ostienses, edendos, etc. (1982)Google Scholar, 46). Sherwin-White, op. cit. (n. 1), 38–9, takes a ‘book-date’ for Book 8 of a.d. 107/8 and maintains that ‘Ep. 14 is certainly held over from an earlier date’ (ibid., 49). Appropriately, Dexter's only living appearance in Pliny's letters is to advocate leniency (5.13.4–6).

³ Talbert, R. J. A.The Senate of Imperial Rome (1984)Google Scholar, 281: ‘a rambling, wordy letter’; ibid., 282: ‘irritatingly he never makes clear whether release did then carry the day against exile.’ I do not intend to open up the psephological questions raised by this letter (for which, see Farquharson, R.Theory of Voting (1969)Google Scholar; Budziszewski, J. ‘Persuading Caesar: a new interpretation of Farquharson's problem’, Public Choice 51 (1986), 129–40CrossRefGoogle Scholar), but it seems clear enough to me that relegation won. Given an opening scenario of let us say (i) acquittal 45 per cent – (ii) relegation 25 per cent – (iii) death 30 per cent (since (ii) and (iii) can together defeat (i), but (i) is stronger than (iii) — s. 14, s. 24), the transfer of all the supporters of (iii) to (ii) reported in s. 25 would guarantee success for (ii). I have numbered the sententiae in the same order as Pliny (s. 12); when he concludes that ‘the third was rejected and the second won’ (s. 26), it would be perverse not to assume the same order. Hence Pliny's closing words: ‘(the third) which, since it could not defeat both, chose by which it would be beaten’ (‘quae, cum ambas superare non posset, elegit ab utra uinceretur’), i.e. since (iii) could not defeat both (i) and (ii), it chose to be defeated by (ii) by being incorporated into it. The fact that Pliny's own sententia was not carried if anything increases the force of his selfless quest for justice (in his terms) and humane desire to block the death penalty at any cost. His ability to make rhetorical capital out of a disappointing outcome is equally clear from 2.12.

⁴ Sherwin-White, op. cit. (n. 1), 461.

⁵ Pliny flags this up at s. 16: ‘allow me to account for my position before you as if we were there [sc. in the Senate] …’ (‘permitte mihi sic apud te tamquam ibi … rationem iudicii mei reddere’). Compare Henderson, J.Pliny's Statue. The Letters, Self-Portraiture and Classical Art (2002)Google Scholar, 126, on the Classicus trial (3.9): ‘handling [events] in the story, and handling them as the story, fold into each other, and must double up as one’. As Mayer, R. ‘Pliny and Gloria Dicendi’, Arethusa 36.2 (2003), 227–34CrossRefGoogle Scholar, at 232, notes, the ‘triumph of pure advocacy’ in this letter probably never had the form of a full-scale speech.

⁶ Almost precisely a third (382 of 1,122 words = 34 per cent). Pliny authorizes such attention to minutiae: in 4.11.16 he warns Minicianus that he will count not just the pages, but the words and even syllables of his next letter.

⁷ cf. Beutel, F.Vergangenheit als Politik. Neue Aspekte im Werk des jüngeren Plinius (2000)Google Scholar, 179–81, 258–62, who rightly emphasizes the political weight of the passage (taken, however, in isolation).

⁸ Scholarship has become increasingly attentive to the artistry of the collection; see especially Marchesi, I.The Art of Pliny's Letters. A Poetics of Allusion in the Private Correspondence (2008)Google Scholar.

⁹ He makes this notoriously clear (for Tacitus’ benefit) in his attack on breuitas in 1.20.

¹⁰ Pliny's five longest letters are: 5.6 (1,521 words), 3.9 (1,216), 8.14 (1,122), 2.17 (1,086), 9.13 (981); in other words, they comprise the two villa letters, the Classicus trial, the quasi-trial where Pliny ‘avenges’ Helvidius Priscus, and 8.14. Carlon, J. M.Pliny's Women. Constructing Virtue and Creating Identity in the Roman World (2009)CrossRefGoogle Scholar, 227–39, gives word-counts for every letter in the collection, not all quite right. Mine, like hers, include uale but not the inscriptiones.

¹¹ 5.6.42: ‘primum ego officium scriptoris existimo, titulum suum legat atque identidem interroget se quid coeperit scribere, sciatque si materiae immoratur non esse longum, longissimum si aliquid accersit atque attrahit’ (‘I consider it a writer's first duty to read his title and ask himself time and time again what it is he has set out to write; and to know that if he dwells on his subject he is not being lengthy, but if he gratuitously drags in something else, he is being too lengthy’). On the literary politics of size in 5.6.41–4, see also Squire, M. J.The Iliad in a Nutshell: Visualising Epic on the Tabulae Iliacae (forthcoming)Google Scholar, ch. 7. For Pliny's usual rule of ‘one letter, one theme’, see Sherwin-White, op. cit. (n. 1), 3–4; Gamberini, F.Stylistic Theory and Practice in the Younger Pliny (1983)Google Scholar, 153–61.

¹² Murgia, C. E. ‘Pliny's Letters and the Dialogus’, HSCP 89 (1985), 171–206Google Scholar, at 189–201, has convincingly shown that the publication of Book 9 was accompanied by an edited, or re-edited, version of a nine-book collection. In his wake, debate has focused especially on the two ‘hunting’ letters to Tacitus, 1.6 and 9.10 (see recently Marchesi, op. cit. (n. 8), 118–35, and Edwards, R. ‘Hunting for boars with Tacitus and Pliny’, ClAnt 27 (2008), 35–58Google Scholar); the responsion of 1.5 and 9.13 has also been remarked (Murgia, op. cit., 199: ‘[9.13] balances and in effect answers Ep. 1.5’; cf. Sherwin-White, op. cit. (n. 1), 93; Hoffer, S. E.The Anxieties of Pliny the Younger (1999)Google Scholar, 88). Whether Book 10 was edited posthumously, or by Pliny himself (as argued by Woolf, G. D. ‘Pliny's Province’ in Bekker-Nielsen, T. (ed.), Rome and the Black Sea Region. Domination, Romanisation, Resistance (2006), 93–108Google Scholar, and Stadter, P. E. ‘Pliny and the ideology of empire: the correspondence with Trajan’, Prometheus 32 (2006), 61–76Google Scholar), it is clearly presented as a detached adjunct.

¹³ It is worth remarking that 2.11 and 2.12 together make a similar length to 8.14, and almost precisely the same length as the villa letter 2.17 (1,084, 1,122 and 1,086 words respectively): further reason to suppose that letter lengths are consequent. On the counterweighting of 2.11–12 and 2.17 within their book, see Gibson, R. and Morello, R.An Introduction to Pliny the Younger (forthcoming)Google Scholar, ch. 7.

¹⁴ One might think also of 2.14, both for its legal theme (as Roy Gibson suggests) and for the shared subject of tirocinium fori (common ground noted by Beutel, op. cit. (n. 7), 258) — an institution in decline according to 2.14.2–4, or, in the version of 8.14.5–8, crushed by Domitian. Such structural considerations as these must remain fluid, given Pliny's characteristic avoidance of patterns which are too perfect in his ‘endless tease of disguised patterning and forsworn composition’ (Henderson, op. cit. (n. 5), xi). For Pliny, as for Augustan poets, ‘interest in balance and arrangement does not require a strict or mechanical overall scheme’, but equally ‘interest in ποικιλία does not imply a thoroughgoing lack of deliberate, formal structural design’ ( Johnson, W.‘The Posidippus papyrus: bookroll and reader’ in Gutzwiller, K. (ed.), The New Posidippus. A Hellenistic Poetry-Book (2005), 70–80Google Scholar, at 79 n. 31).

¹⁵ Pliny is Tacitus’ most resonant contemporary sounding-board: cf. most recently Marchesi, op. cit. (n. 8), 97–206; also the debate on Dialogus (above, n. 12; Woodman, A. J. ‘Tacitus and the contemporary scene’ in idem (ed.), Cambridge Companion to Tacitus (2010), 31–43CrossRefGoogle Scholar); Whitton, C. L. ‘“Let us tread our path together”: Tacitus and the Younger Pliny’ in Pagán, V. E. (ed.), Blackwell Companion to Tacitus (forthcoming)Google Scholar. On Tacitus’ reception more broadly, see Haverfield, F. ‘Tacitus during the late Roman period and the Middle Ages’, JRS 6 (1916), 196–201Google Scholar; Zecchini, G. ‘La fortuna di Tacito e l’Historia Augusta’ in Bonamente, G. and Duval, N. (eds), Historiae Augustae Colloquium Parisinum (1991), 337–50Google Scholar; and Giua, M. A. ‘Tacito e i suoi destinatari: storia per i contemporanei, storia per i posteri’ in Casanova, A. and Desideri, P. (eds), Evento, racconto, scrittura nell’antichità classica (2003), 247–68Google Scholar.

¹⁶ This line doubles as a textual cue, a reminder that the reader has met Aristo before (cf. 1.22.2, ‘quam peritus ille et priuati iuris et publici!’).

¹⁷ Pliny is explicit about his process of eulogizing Trajan by denigrating Domitian, a technique he exploits most relentlessly in the Panegyricus, but also regularly in the Epistles. Note Pan. 53.5, ‘licet nobis et in praeteritum de malis imperatoribus cotidie uindicari et futuros [sc., no doubt, et praesentes] sub exemplo praemonere …’. See especially Ramage, E. S. ‘Juvenal and the establishment. Denigration of predecessor in the “Satires”’, ANRW 2.33.1 (1989), 640–707Google Scholar, at 652–3; Hoffer, op. cit. (n. 12), especially 61–6; Beutel, op. cit. (n. 7), passim; Freudenburg, K.Satires of Rome. Threatening Poses from Lucilius to Juvenal (2001)CrossRefGoogle Scholar, 215–34. Trajanic eulogy has its place in Book 8 too (8.4 on the second Dacian War).

¹⁸ ‘Thesaurus’: cf. 1.22.2, ‘mihi certe quotiens aliquid abditum quaero, ille thesaurus est’, apt preparation for 8.14. For a theoretical exposition of the now familiar triangulation of writer, addressee and secondary readership, see Beutel, op. cit. (n. 7), 139–56. The surprise of Sherwin-White, op. cit. (n. 1), 461, that Pliny bothers to give information ‘as well known to Aristo as to Pliny’ demonstrates the problem with treating these letters as document, not literature.

¹⁹ The confident argumentation of ss. 12–23 (only superficially repressed in s. 24, ‘sed quid ego similis docenti?’) makes clear that Pliny does not expect his professions of self-doubt to be taken at face value.

²⁰ If Aristo is indeed an equestrian (Kunkel, op. cit. (n. 1), 319 n. 665; Sherwin-White, op. cit. (n. 1), 461; Syme, op. cit. (n. 1), 352: ‘a small municipal man’), Pliny is in typical style presenting a world where senator and eques work together for the common good. The same may be said of the different expertises of the two, with Pliny as practitioner, Aristo as the jurisprudent with the theoretical scientia (s. 10). Yet it is not so much legality (as we conceive it) that is on display as Pliny's oratory, and his trademark humanitas (n. 114).

²¹ On such self-revelation, see Harrison, S. J. ‘Poetry, philosophy, and letter-writing in Horace Epistles I’ in Innes, D., Hine, H. and Pelling, C. (eds), Ethics and Rhetoric: Classical Essays for Donald Russell on his Seventy-Fifth Birthday (1995), 47–61Google Scholar, at 50–1. On the didactic nature of epistles, see also Woolf, G. ‘The city of letters’ in Edwards, C. and Woolf, G. (eds), Rome and the Cosmopolis (2003), 203–21Google Scholar, at 213. Vogt-Spira, G. ‘Die Selbstinszenierung des jüngeren Plinius im Diskurs der literarischen Imitatio’ in Castagna, L. and Lefèvre, E. (eds), Plinius der Jüngere und seine Zeit (2003), 51–61Google Scholar, at 51, sees in the written Pliny the homoios prescribed by Aristotle as the ideal tragic figure.

²² Guillemin, A.-M.Pline et la vie littéraire de son temps (1929)Google Scholar, 77, cites this letter as especial ‘preuve qu’il connaissait mal le droit civil et le droit public’. So too for Bütler, H.-P.Die geistige Welt des jüngeren Plinius (1970)Google Scholar, 139 n. 38, it ‘zeigt Plinius von einer Ahnungslosigkeit, die bei einem Mann, welcher selbst dem Senat präsidiert hat, mehr als befremdet’. Hoffer, op. cit. (n. 12), 25, has it right: ‘Pliny is not in despair or even indecision: he already knows the correct answer’.

²³ His presentation in ss. 12–23 is not so much pedantic as sophistic; note in particular how he veers from juristic proofs to opportunistic moralizing in s. 21, where his only argument is that mercy is intrinsically fairer, complete with an ironical comparison of the alternative to a gladiatorial show (‘scilicet ut in spectaculis …’). Clearer still, in s. 24 he says that he won his argument ‘nescio an iure, certe aequitate’.

²⁴ Sherwin-White, op. cit. (n. 1), 461: ‘historical excursus’; Beutel, op. cit. (n. 7), 259: ‘historischer Excursus’. Gamberini, op. cit. (n. 11), 159, calls it the ‘most conspicuous’ digression of the Epistles.

²⁵ ss. 1–11 are structured A–B–C–C–B–A around a central opposition of former times and Pliny's generation: A1 (s. 1: Aristo, and Pliny's putative mistake, ‘errauerim’) – B1 (ss. 2–3: problem of slow recovery from past ignorance) – C1 (ss. 4–6: former times) – C2 (ss. 7–9: our generation) – B2 (s. 10: problem of the ‘breue tempus’ they have had to catch up) – A2 (ss. 10–11: Aristo, and Pliny's putative mistake, ‘error’).

²⁶ I take the term from Flower, M. A. and Marincola, J.Herodotus Histories Book IX (2002)Google Scholar, 6–7, though it is of course equally at home in epic. This suggestion seems less odd when the linearity of the collection is considered (below, Section V), not to mention its historiographical element. On the subtle interweaving of historiography into Pliny's collection see especially Ash, R. ‘“Aliud est enim epistulam, aliud historiam … scribere” (Epistles 6.16.22): Pliny the historian?’, Arethusa 36.2 (2003), 211–25CrossRefGoogle Scholar.

²⁷ To take just one much discussed example, Tacitus’ digression at An. 4.32–3 stands in evident and provocative juxtaposition with the Cordus trial in An. 4.34–5 (see Moles, J. ‘Cry freedom: Tacitus Annals 4.32–35’, Histos (1998)Google Scholar, especially s. 4; Sailor, D.Writing and Empire in Tacitus (2008)CrossRefGoogle Scholar, 250–313); for another useful consideration of historical digression, see Wiedemann, T. ‘Sallust's “Jugurtha”: concord, discord and the digressions’, G&R 40 (1993), 48–57Google Scholar.

²⁸ Ep. 8.14.2–12 and the Agricola have been paired in several recent discussions, including Liebeschuetz, W. ‘The theme of liberty in the Agricola of Tacitus’, CQ 16 (1966), 126–39CrossRefGoogle Scholar, at 133; Bütler, op. cit. (n. 22), 139; Beutel, op. cit. (n. 7), 180 n. 500; R. Gazich, ‘Retorica dell’esemplarità nelle lettere di Plinio’ in Castagna and Lefèvre, op. cit. (n. 21), 123–41, at 127–31; Haynes, H. ‘Survival and memory in the Agricola’, Arethusa 39.2 (2006), 149–70CrossRefGoogle Scholar, at 158–60. Further references below.

²⁹ A bolder (if schematic) argument is that Tacitus’ model general and Domitian-antitype Agricola becomes for Pliny a paradigm for his model emperor Trajan. Possible points of contact between Ag. and Pan. are catalogued in Mesk, J. ‘Zur Quellenanalyse des Plinianischen Panegyricus’, WS 33 (1911), 71–100Google Scholar, at 91–4; Durry, M.Pline le Jeune. Panégyrique de Trajan (1938)Google Scholar, 61; Bruère, R. T. ‘Tacitus and Pliny's Panegyricus’, CP 49 (1954), 161–79Google Scholar, at 162–4 with n. 22.

³⁰ On 2.1 and its possible engagement with Ag. see Syme, R.Tacitus (1958)Google Scholar, 121; Marchesi, op. cit. (n. 8), 189–99. Pliny begins 6.16 (to Tacitus) with a courtesy nod to Ag. 1.1 (‘tradere posteris’ ~ ‘posteris tradere’), as Berry, D. H. ‘Letters from an advocate: Pliny's “Vesuvius” narratives (Epistles 6.16, 6.20)’, PLLS 13 (2008), 297–313Google Scholar, at 299, has also noted. Like Berry, I am unconvinced by the suggestion of Tzounakas, S. ‘Neque enim historiam componebam: Pliny's first epistle and his attitude towards historiography’, MH 64 (2007), 42–54Google Scholar, that 1.1.2 recalls Ag. 42.4. Another Agricolan reflex may occur in 8.20.1, ‘ut proximorum incuriosi longinqua sectemur’: this use of incuriosus with the genitive is a novelty in Ag. 1.1, ‘incuriosa suorum aetas’ (cf. Heubner 1984 ad loc.). The keen scholarly debate over Pliny's intertextual conversation with the Dial. (n. 12) need not detain us here, except as further evidence of the close engagement of these two authors. On Histories, see below, n. 54.

³¹ Most memorably deployed in Cicero's Philippics (see Manuwald 2007 ad Phil. 3.29, and Arena, V. ‘Invocation to liberty and invective of dominatus at the end of the Roman Republic’, BICS 50 (2007), 49–74Google Scholar, who sees it in Stoic terms), in the speeches in Sallust's Histories, and retrojected to the early Republic in Livy.

³² Principate: cf. Roller, M. B.Constructing Autocracy. Aristocrats and Emperors in Julio-Claudian Rome (2001)CrossRefGoogle Scholar, 213–87. Pan.: e.g. 66.5, ‘obsaepta diutina seruitute ora reseramus’; cf. Pan. 2.3, 2.5, 7.6, 8.1, 11.5, 45.1, 45.3, 52.7, 55.2, 55.7, 63.6, 68.2, 72.6, 85.2, 88.1 for explicit seruitus and dominatio (and cognates). Martial's Saturnalian Epigrams 11 (a.d. 96) plays out Nerva's accession as liberation (especially Epig. 11.7), but he largely avoids the lexis of seruitus (dominus twice, however, in Epig. 10.72, a.d. 98).

³³ Liebeschuetz, op. cit. (n. 28), 138–9, sees the ‘parallel’ between Britain and Rome as ‘unintentional’ but ‘real’. It is hard to imagine that Tacitus would not have appreciated the force of the juxtaposition. In the crudest lexical terms, note the agglomeration of seruitus and cognates alone in Calgacus’ speech (eight times in Ag. 30–2); and, for the Senate, seruitus (Ag. 2.3, 3.3) and seruientium (Ag. 40.3). For a full account, see Lavan, M. ‘Slavishness in Britain and Rome in Tacitus’ Agricola’, CQ (forthcoming)Google Scholar, and compare the remarks of Roberts, M. ‘The revolt of Boudicca (Tacitus, Annals 14. 29–39) and the assertion of libertas in Neronian Rome’, AJP 109 (1988), 118–32Google Scholar, at 131–2, on the ‘model of imperial oppression’ applied to both Britain and Rome in An. 14.

³⁴ In 3.5.5 Nero's late principate is characterized as seruitus. The word recurs only at 8.24.8 (of a province: see Section VI) and 10.66.1 (of slaves).

³⁵ By this stage ‘reducta libertas’ (uel sim.) seems to be Plinian shorthand for the change of regime: Pan. 58.3, ‘redditae libertatis’, 78.3, ‘libertatem reuoces ac reducas’ and ‘reciperata libertas’, Ep. 9.13.4, ‘redditae libertatis’; cf. the well-known inscription to Nerva's ‘restoration of freedom’ (‘libertati … restitutae’, ILS 274 = CIL 6.472), and the coin-legend ‘libertas publica’ (BMC 3.xxxviii–ix, 1–30; Shotter, D. C. A. ‘The principate of Nerva: some observations on the coin evidence’, Historia 32 (1983), 215–26Google Scholar).

³⁶ Heubner 1984 ad loc. A flame simile is central to the intertextual knot between Dial. 36.1 and Ep. 1.6 discussed (inter alios) by Murgia, op. cit. (n. 12), 174–5, Marchesi, op. cit. (n. 8), 122–8, and Edwards, op. cit. (n. 12), 41–2 (but denied by Brink, C. O. ‘Can Tacitus’ Dialogue be dated?’, HSCP 96 (1994), 251–80Google Scholar, at 260–2, and by Woodman, op. cit. (n. 15), 32–3).

³⁷ On mixed metaphor, a recognized poetic device, see Norden 1903 ad Aen. 6.236. Neither Lucretius (3.896, ‘tacita pectus dulcedine tangent’) nor Seneca (Ag. 496, ‘miserisque lucis tanta dulcedo est malae’) approaches the daring of Pliny's formulation.

³⁸ Heubner 1984 ad loc.: ‘Zum Gedanken …’.

³⁹ Pliny's pair of adjectives is attested elsewhere, in Livy 1.19.4 (the plebs in ancient times, ‘multitudinem imperitam et illis saeculis rudem’), Tac., D. 19.2 and Sen., Ep. 72.9 (the philosophically ignorant, ‘imperitis ac rudibus’), and I would not wish to exclude the influence especially of the latter. Yet the close similarity of context, and the surrounding allusions, make allusion to Tacitus the most likely of the three.

⁴⁰ On Pliny's productive use of allusion, see especially Schenk, P. ‘Formen von Intertextualität im Briefkorpus des jüngeren Plinius’, Philologus 143 (1999), 114–34Google Scholar, and Marchesi, op. cit. (n. 8), passim.

⁴¹ For the gesture, cf. ps.-Sen., Oct. 288, ‘nos quoque nostri sumus immemores’, considered ‘striking’ by Boyle 2008 ad loc. (Ferri 2003 ad loc. compares Ag.).

⁴² Ag. 1.1. Here too the contrast is between exemplary olden times (in an unspecified past) and more recent decline. Again, other parallels are available (in particular, Caes., B.G. 6.11.4 and b.c. 3.92.5; cf. also Pliny, N.H. 18.168, ‘antiquitus factitatum’), but none has any obvious contextual relevance. Both marshal the (common) trope of passing down knowledge: s. 4, ‘tradenda minoribus’, Ag. 1.1, ‘posteris tradere’; recall that Pliny has already used ‘posteris tradere’ in the proem, as it were, of his quasi-historical Ep. 6.16 to Tacitus (above, n. 30).

⁴³ The only comparison drawn by Sherwin-White, op. cit. (n. 1), 462; also by Seelentag, G.Taten und Tugenden Traians. Herrschaftsdarstellung im Principat (2004)Google Scholar, 286 n. 83. Cf. also Pan. 14.5, ‘alienisque uirtutibus tunc quoque inuidus imperator’, and Ep. 2.1.3 (on Verginius Rufus), ‘Caesares quibus suspectus atque etiam inuisus uirtutibus fuerat euasit’. Once again, Pliny recontextualizes in 8.14, putting uirtus in the singular and giving it a specifically military reference. The referent of Ag. 1.4 is problematic: see Sailor, op. cit. (n. 27), 53–72, especially 58–9, for an excellent reading.

⁴⁴ In now familiar technique, Pliny repeats ‘inertia’ but replaces one prepositional phrase (‘pro sapientia’) with another (‘in pretio’). In Ag. 6 Nero is the proto-Domitian, prelude to the monograph's principal tyrant (compare the Neronian seruitus of Ep. 3.5.5 (above, n. 34) anticipating 8.14; and cf. Ag. 45.2, Mart. 11.33, Juv. 4.38, 4.152). Compare too the gerund and jingle of s. 5 ‘imperare parendo’ (‘give orders while obeying’) with Agricola's perfectly balanced qualities at Ag. 8.3, ‘uirtute in obsequendo, uerecundia in praedicando’ (‘valour while obeying, modesty in public pronouncements’).

⁴⁵ Syme, op. cit. (n. 30), 212 n. 2.

⁴⁶ For the sentence ‘ingenia nostra … contusa sunt’, cf. Ag. 3.1, ‘sic ingenia studiaque oppresseris facilius quam reuocaueris’. Haynes, op. cit. (n. 28), 158–9 identifies a shared ‘metaphor of the body’ (Ag. 3.1, ‘ut corpora nostra …’; left implicit by Pliny).

⁴⁷ Sherwin-White, op. cit. (n. 1), 463, observes that ‘multos per annos’ (s. 3) is a stretch, since Pliny had ‘in fact’ been a senator for only six years under Domitian, but the associative plural gives Pliny room enough for manoeuvre. Is he thinking of Tacitus’ lingering evocation of those fifteen long years (Ag. 3.2, ‘per quindecim annos, grande mortalis aeui spatium’)?

⁴⁸ For ‘uidimus’ and the rhetoric of painful autopsy, cf. Ag. 45.1, ‘non uidit Agricola …’, 45.2, ‘praecipua sub Domitiano miseriarum pars erat uidere et aspici’ (the trope is not new in Tacitus: see especially Cic., De or. 3.8, and Nisbet and Hubbard 1970 ad Hor., C. 1.2.12). ‘tulimus’ equivocates between ‘we endured’ and ‘we allowed’ — not unlike the patientia of Ag. (the tough karteria of Agricola's men, Ag. 33.2, but passive — slavish — compliance by the Senate, Ag. 2.3).

⁴⁹ On the term ‘associative plural’ and its use by Tacitus, see Sinclair, P.Tacitus the Sententious Historian. A Sociology of Rhetoric in Annales 1–6 (1995)Google Scholar, 53–5; specifically on Ag. proem, see Sailor, op. cit. (n. 27), 70–1. ‘adflixit’ is Reitzenstein's conjecture for a much debated crux.

⁵⁰ Pliny's strenuous efforts to align himself with the ‘martyrs’ are well known (e.g. 1.5, 3.11, 7.33, 9.13), and indeed lead Hoffer, op. cit. (n. 12), 24–5, to believe that in 8.14.9 too, in contrast to Tacitus, ‘Pliny rejects the charge of collaboration’ (‘participes malorum’ argues against). The project of Pan. is to align Trajan and all senators as common victims (e.g. Pan. 44.1, ‘uixisti nobiscum, periclitatus es, timuisti …’, 72.6, ‘nosti necessitudinem seruitutis’).

⁵¹ Sailor, op. cit. (n. 27), 60–2, makes the attractive suggestion that Tacitus is executing a literary damnatio memoriae, and notes further the near-absence of Domitian's name from Pan. Within Ag., its absence from the proem (also) intensifies the effect of its glowering presence in Ag. 39–45.

⁵² Three times in Ep. 1.5, for instance. See the index nominum in Mynors’ Oxford text for a full list.

⁵³ e.g. Aesch., Ag. 36–7; Phaedrus, Fables 3.prol.34–5, ‘seruitus … quae uolebat non audebat dicere’. Tacitus’ own formulation is indebted to Cicero, especially Pro Marcello, as Narducci, E. ‘Tacito e la Pro Marcello di Cicerone. Nota a Historiae 1.1’, Prometheus 32 (2007), 231–2Google Scholar, has shown. It is the similarity of structure between Pliny's and Tacitus’ sentences which draws attention (‘quod uelles … quod nolles’ ~ ‘quae uelis … quae sentias’).

⁵⁴ ‘Felicitas’ was of course a slogan of the Nervan restoration (cf. 10.12.2; Nerva's edict in 10.58.7; Ag. 3.1), not for the first time (see Eden 1984 ad Sen., Apoc. 1.1). Note the uariatio in the shift from ‘licet’ to ‘libet’, from ‘quae … quae’ to ‘quid … quod’ and from chiasmus (‘sentire … et … dicere’) to asyndetic bicolon (‘scire … exercere …’). Where Tacitus characteristically avoids clausulae, Pliny (equally characteristically) generates a pair of cretic-spondees. Conversely, Haynes, op. cit. (n. 28), 159, reads H. 1.1.4 as ‘echoing’ Pliny, with a shift from ‘libet’ to ‘licet’ serving as ironic Tacitean comment on (lack of) free speech. Other considerations aside, this is unlikely chronology: Pliny's Vesuvius letters (6.16 and 6.20) show that Tacitus’ project was well underway already by the time of Book 6, although 5.8 is flimsy evidence for publication (or not) in a.d. 105, as the diametrically opposed interpretations of Syme, op. cit. (n. 30), 118, and Sherwin-White, op. cit. (n. 1), 335, suggest. For the proverbial brevity of happiness, see e.g. Sen., Ag. 928, ‘o nulla longi temporis felicitas’.

⁵⁵ A limit rightly stressed by Sailor, op. cit. (n. 27), 66 n. 40.

⁵⁶ Unnoticed by commentators, the final words of H. 1.1.4, ‘sentias dicere licet’, evoke the last three words of the traditional formula with which the presiding consul summoned the Senate: ‘senatores quibusque in senatu sententiam dicere licet’ (Gell., N.A. 3.18.6–8; Festus 339.16 Lindsay; cf. Talbert, op. cit. (n. 3), 187 n. 23, and add Livy 36.3.3).

⁵⁷ H. proem recalls Ag. generally and in some specifics, e.g. Ag. 3.1, ‘felicitatem temporum’ ~ H. 1.1.4, ‘rara temporum felicitate’; cf. also Ag. 45.1, ‘consularium caedes … feminarum exilia’ ~ H. 1.3.1, ‘secutae maritos in exilia coniuges’ (so Ogilvie and Richmond 1967 ad loc.) as well as H. 1.2.2, ‘plenum exiliis mare, infecti caedibus scopuli’ (the only times exilia and caedes, plural, are paired in Latin). Sailor, op. cit. (n. 27), 119–38, teases out the rich implications of slavishness in H. 1.1.1, ‘dominantes’.

⁵⁸ Tacitus receives three letters in Book 6, two in Book 7, and two in Book 9. He is a subject of 9.23 and, as most believe, of 9.27 (e.g. Syme, op. cit. (n. 30), 120, Sherwin-White, op. cit. (n. 1), 509–10).

⁵⁹ ‘Tongue-tied’ rather than ‘tongueless’ because ‘elinguis habet linguam, sed usu eius caret’ (De differentiis in Keil, H.Grammatici Latini, vol. 7 (1880)Google Scholar, 529.7).

⁶⁰ The first adjective pair mutus – elinguis is widely imitated from Livy onwards (see TLL V.2 390.84–391.22). tacitus – fractus is unparalleled.

⁶¹ For a contextualized reading of Cicero's speech see Nicholson, J.Cicero's Return from Exile: the Orations Post Reditum (1992)Google Scholar; on his shaping of the recent past in this period, Riggsby, A. M. ‘The Post Reditum speeches’ in May, J. M. (ed.), Brill's Companion to Cicero. Oratory and Rhetoric (2002), 159–95CrossRefGoogle Scholar, at 160–72 (though with little on Red. sen.).

⁶² According to Red. sen. 12, Gabinius surpassed any tyrannus in forbidding public displays of grief at Cicero's exile; s. 13–16 paint him as a libidinous drunkard (cf. Sest. 18, and see Ash, R.Ordering Anarchy. Armies and Leaders in Tacitus’ Histories (1999)CrossRefGoogle Scholar, 96–105, on such stock ‘tyrant’ invective); ss. 18 describes the ‘uacuo non modo a bonis sed etiam a liberis atque inani foro’; in s. 19 the pro-Ciceronian tribune Annius ‘seruitutem depulit ciuitati’. For the empty forum as index of tyranny compare Luc. 5.31–2 (Caesar), Pl., Pan. 48.3 (Domitian), Tac., An. 4.70.2 (Tiberius).

⁶³ Belying, then, Pliny's lament that his life is far too dull by comparison with Cicero's (9.2.2–4) — though that letter is of course inviting precisely the comparison he claims to reject. On Pliny's affiliation to Cicero, clear both from declarations and from his epistolary project as a whole, see Weische, A. ‘Plinius d. J. und Cicero. Untersuchungen zur römischen Epistolographie in Republik und Kaiserzeit’, ANRW 2.33.1 (1989), 375–86Google Scholar; Riggsby, A. M. ‘Pliny on Cicero and oratory: self-fashioning in the public eye’, AJP 116 (1995), 123–35Google Scholar; Marchesi, op. cit. (n. 8), 207–40, 252–7.

⁶⁴ Note further the shared focus on the visual, ‘prospeximus’ ~ ‘uidebatis’. TLL V.2 390.81–3 rightly draws the connection, as does Trisoglio, F.Opere di Plinio Cecilio Secondo (1973)Google Scholar, 820 n. 218 (without comment). It may be a ‘metafora degna di Tacito’ (Gazich, op. cit. (n. 28), 129), but Tacitus did not write it.

⁶⁵ The latter is clear enough from Ag. 2–3 ‘uocem’, ‘adempto per inquisitiones etiam loquendi audiendique commercio’, ‘uoce’, ‘tacere’, ‘incondita ac rudi uoce’. For the crushing silence of tyranny, see also e.g. Luc. 1.247, ps.-Sen., Oct. 350–1, Stat., Theb. 1.169–70, 2.481.

⁶⁶ Set alongside Ep. 8.14.9 in TLL IV 806.34–5.

⁶⁷ Ovid follows Lucr. 5.1018, ‘ingenium fregere’; perhaps also, as Luck 1977 ad loc. suggests (inter alios), Cic., Ad fam. 4.8.1, itself concerned with exile. This self-diagnosed ‘blunted intellect’ can be distinguished from the invective ‘hebes ingenium’ (i.e. ‘blockhead’) of Cic., Tusc. 5.45, N.D. 2.17, Phil. 10.17.

⁶⁸ See also Sen., Ep. 24.16, ‘hebetabitur [sc. animus]’ (by solitude); Pl., Pan. 70.5, ‘sincera rectaque ingenia … hebetabat … reputatio’ (= the thought that good provincial governance would earn no rewards under Domitian).

⁶⁹ Both Ovid and Seneca, like Tacitus (n. 55) but unlike Pliny, are using ingenium with emphasis on literary talent. Pierini, R. Degl’InnocentiTra Ovidio e Seneca (1990)Google Scholar, 103–66, has amply shown that Seneca has Ovid's exile poetry in mind (see especially 112–22 on Polyb. 18.9).

⁷⁰ ‘haec, utcumque potui, longo iam situ obsoleto et hebetato animo composui’ (Polyb. 18.9).

⁷¹ Compare the combination ‘longis hebetata malis’ with Pliny's ‘mala … quibus … hebetata … sunt’.

⁷² cf. also Thy. 237 (Atreus), ‘per regna trepidus exul erraui mea’.

⁷³ In Ovid's case this is self-evident, whatever one's view of the ‘reality’ of his exile; for Thyestes, see Pierini, R. Degl’Innocenti ‘Il tema dell’esilio nelle tragedie di Seneca: Autobiografia, meditazione filosofica, modelli letterari nel Thyestes e nell’Oedipus’, QCTC 8 (1990), 71–83Google Scholar. With the fractum ingenium of all three authors, compare too Cicero's ‘tacitam et fractam ciuitatem’ in Red. sen. (quoted above), where he develops a conceit that the whole ciuitas shared his exile (cf. Narducci, E. ‘Perception of exile in Cicero. The philosophical interpretation of a real experience’, AJP 118 (1997), 55–73Google Scholar, at 66–9; Cohen, S. T. ‘Cicero's Roman exile’ in Gaertner, J. F. (ed.), Writing Exile. The Discourse of Displacement in Greco-Roman Antiquity and Beyond (2007)Google Scholar, 109–28, at 111).

⁷⁴ Ovid and Seneca jostle for prominence in Pliny's very first letter, which alludes not just to Ex Ponto 3.9.53 (‘postmodo collectas utcumque sine ordine iunxi’) as observed by Syme, R. ‘The dating of Pliny's latest letters’, CQ 35 (1985), 176–85CrossRefGoogle Scholar, at 176 (= R.P. 5.478) (the similarity is also noted by Froesch, H. H.Ovids Epistulae ex Ponto I–III als Gedichtsammlung (1968)Google Scholar, 51), but also, I suggest, to the opening of Seneca's Epistulae morales: Pl., Ep. 1.1, ‘collegi non seruato temporis ordine …’ ~ Sen., Ep. 1.1, ‘ita fac, mi Lucili: uindica te tibi, et tempus quod adhuc aut auferebatur aut subripiebatur aut excidebat collige et serua’ (Tony Woodman suggests a further intertext with Tac., D. 1.3, ‘seruato ordine’, though he believes D. to be the later text: cf. Woodman, op. cit. (n. 15)). Seneca is named only once in the Epistles (5.3.5), but that is no index of his allusive import. For some instances in Book 8 see below, n. 116. See also Henderson, op. cit. (n. 5), 11–13, 27–30, 84–5, 122; Marchesi, op. cit. (n. 8), 233–6. Pliny has followed his master Quintilian's precept, that in Seneca ‘there is much to admire, but you need to be selective’ (Inst. 10.1.131, ‘multa etiam admiranda sunt, eligere modo curae sit’). Though Seneca's tragedies may seem a less obvious allusive target than his letters, they were well known, to judge from allusions and references in Martial, Quintilian, and especially Statius ( Vessey, D.Statius and the Thebaid (1973)Google Scholar, 72–8; Davis, P. J.Seneca: Thyestes (2003)Google Scholar, 81–6). Their influence on Tacitus, once flatly denied (e.g. Henry, D. and Walker, B. ‘Tacitus and Seneca’, G&R 10 (1963), 98–110Google Scholar, at 107), is proposed by Santoro-L’Hoir, F.Tragedy, Rhetoric, and the Historiography of Tacitus’ Annales (2006)CrossRefGoogle Scholar, 51–6, 102–8, 204–20.

⁷⁵ For a recent bibliographical overview, see J. F. Gaertner, ‘The discourse of displacement in Greco-Roman antiquity’, in idem, op. cit. (n. 73), 1–20. On the reappropriation of exile in contemporary Greek literature as an ‘intense (if stylized) engagement with highly topical issues of power and identity’, see the stimulating discussion of Whitmarsh, T. J. G.Greek Literature and the Roman Empire. The Politics of Imitation (2001)Google Scholar, 133–80 (quotation from 179).

⁷⁶ Though see also Sen., Ag. 991–2, ‘inops egens inclusa, … exul, inuisa omnibus’: Aegisthus describes Electra's imminent imprisonment (cf. inclusa) at the hands of a tyrant — himself — as exile (cf. exul). This is no lazy recycling from Med. 21, ‘exul pauens inuisus’ (as implied by Tarrant 1976 ad loc.), but an index of the power and flexibility of exile as a theme.

⁷⁷ Doblhofer, E.Exil und Emigration: Zum Erlebnis der Heimatferne in der römischen Literatur (1987)Google Scholar, 227 (cf. 221–41), building on the notion of Cicero as émigré de l’intérieur set out by Herescu, N. J. ‘Les trois exils de Cicéron’ Atti del I Congresso internazionale di studi ciceroniani I (1959), 137–56Google Scholar. In Ad fam. 4.8.2, for instance, Cicero advises Marcellus that he may as well ‘be in exile’ in the comfort of his own home (‘hunc tamen aptissimum esse etiam ad exulandum locum’).

⁷⁸ On the strategic value of ‘exile on main street’ as apologia, see also Sailor, op. cit. (n. 27), 32–3.

⁷⁹ ‘The voice of Cicero in Tacitus’ Agricola’, forthcoming article.

⁸⁰ Mayor 18772 ad Juv. 3.9 (at 1.173–82) is an exhaustive compilation of ancient testimonies. For the continuing view that ‘the matter was … in general much less regarded than the manner’ (ibid., at 1.180), see e.g. Dupont, F. ‘Recitatio and the reorganization of the space of public discourse’ in Habinek, T. and Schiesaro, A. (eds), The Roman Cultural Revolution (1997), 44–59Google Scholar, at 50, and Johnson, W. A. ‘Toward a sociology of reading in Classical Antiquity’, AJP 121 (2000), 593–627Google Scholar, at 614–15.

⁸¹ cf. Altman, J.Epistolarity: Approaches to a Form (1982)Google Scholar, 88: ‘the epistolary form is unique in making the reader (narratee) almost as important an agent in the narrative as the writer (narrator)’.

⁸² cf. Altman, op. cit. (n. 81), 167–84. Of course, linearity is only one of several ways of joining the dots in books of letters, as of poetry (hence Henderson's, op. cit. (n. 5), xi (with n. 5), useful concept of a shifting ‘kaleidoscope’), but it is an important one.

⁸³ Pliny's linearity has been little explored, despite Mayer's call to arms (op. cit. (n. 5), 232). See, however, Ludolph, M.Epistolographie und Selbstdarstellung. Untersuchungen zu den ‘Paradebriefen’ Plinius des Jüngeren (1997)Google Scholar, 92–8, a structural analysis of the sequence Ep. 1.1–8; Berry, op. cit. (n. 30), 302–3 on 6.15–16; and Marchesi, op. cit. (n. 8), especially 12–52. The letter collection was presumably presented on a bookroll, itself an inducement to linear reading — and to roll backwards as well as forwards (see Sickle, J. van ‘The book roll and some conventions of the poetic book’, Arethusa 13 (1980), 5–42Google Scholar, at 5–6). Gibson and Morello, op. cit. (n. 13) offer much stimulating discussion of larger sequential patterns in Pliny.

⁸⁴ For another name-pun, see the close of Book 8 (‘… quod esse maximum debet’, 8.24.10, addressed to Maximus). Such wordplay can be productive and meaningful for Pliny: an important instance is the collection's diurnal cycle from bright Septicius Clarus (1.1) to the dusk of Pedanius Fuscus (9.40), seen by Barchiesi, A. ‘The search for the perfect book: a PS to the new Posidippus’ in Gutzwiller, K. (ed.), The New Posidippus: A Hellenistic Poetry Book (2005), 320–42Google Scholar, at 331–2, and Marchesi, op. cit. (n. 8), 249–50. Tony Woodman suggests (inter alia) a pun in the commendation of silentium to Tacitus in 1.6. For more instances see Gibson and Morello, op. cit. (n. 13), ch. 6. In 8.13 Genialis’ name matches the feel of the letter, and both contrast with the sombre tone of the excoriation of Domitian to come.

⁸⁵ Bernstein, N. W. ‘Each man's father served as his teacher: constructing relatedness in Pliny's letters’, CA 27 (2008), 202–30Google Scholar, at 217. In another contribution (idem ‘Cui parens non erat maximus quisque et uetustissimus pro parente: parental surrogates in imperial Roman literature’ in Hübner, S. R. and Ratzan, D. M. (eds), Growing Up Fatherless in Antiquity (2009), 241–56CrossRefGoogle Scholar, at 251–5), he investigates a political reading of the ‘surrogate’ paternity in 8.13 and 8.14, set with the praise of extra-familial succession in the Panegyricus. In the speech, though, Pliny is careful not to reject natural paternity, explicitly introducing Trajan pater (Pan. 89.2–3: contra Bernstein, 254) and hedging his bets on the next imperial succession (note the remarkable prayer at Pan. 94.5 that Jupiter preferably grant Trajan an heir ‘whom he has produced, nurtured, and made to resemble an adoptive son’ (‘quem genuerit quem formauerit similemque fecerit adoptato’).

⁸⁶ Lucr., D.R.N. 3.3–4; Callimachus, Aetia 1 fr. 1.26 Pfeiffer; Hor., Epp. 1.19.21, 2.2.80, A.P. 286–7. Where Callimachus and Horace (in varying degrees) refuse to follow other poets’ tracks, Lucretius proclaims devotion to his ‘philosophical model, “father” Epicurus’ ( Hardie, P. R.The Epic Successors of Virgil: a Study in the Dynamics of a Tradition (1993)Google Scholar, 110).

⁸⁷ Cic., De or. 1.105: cf. Gazich, op. cit. (n. 28), 132 with n. 30.

⁸⁸ Aen. 1.94–6 (cf. Od. 5.306–7). I owe this suggestion to James Diggle. For open citations of Virgil see the index in Mynors’ Oxford text, and on Pliny's productive use of them, see Krasser, H. ‘extremos pudeat rediisse – Plinius im Wettstreit mit der Vergangenheit. Zu Vergilzitaten beim jüngeren Plinius’, A&A 39 (1993), 144–54Google Scholar; Marchesi, op. cit. (n. 8), 27–39, 140–2, 154–5, 175–89, suggests more subtle Virgilian play.

⁸⁹ Hardie, op. cit. (n. 86), 98–105, has shown the vigour of the trope of ‘allusion that alludes to itself’ in Roman epic (taking the term from Ricks, C. ‘Allusion: the poet as heir’ in Brissenden, R. F. and Eade, J. C. (eds), Studies in the Eighteenth Century III (1976), 209–40Google Scholar): here Pliny borrows both verbally and tropically. See too Krasser, op. cit. (n. 88), 145–6, on 5.8.3. Another famous (and for Pliny recent) instance of such self-conscious uestigia is Stat., Theb. 12.817, discussed usefully by Marchesi, op. cit. (n. 8), 142; there the familial model is wife–husband (cf. Aen. 2.711 and perhaps Georg. 4.487).

⁹⁰ Aen. 5.320; cf. Krasser, op. cit. (n. 88), 152–4, assuaging the doubts of Neuhausen, K. A. ‘Plinius proximus Tacito. Bemerkungen zu einem Topos der römischen Literaturkritik’, RhM 111 (1968), 333–57Google Scholar; Marchesi, op. cit. (n. 8), 135–43. By the usual estimates, Tacitus (born a.d. 56–8) was three to six years older than Pliny (born a.d. 61 or 62). Pliny tempers this emphasis on disparity earlier in the letter (7.20.3).

⁹¹ Marchesi, op. cit. (n. 8), 102–13 (on 8.7) and 135–43, exposes the fallacy of taking Pliny's modesty in his letters to Tacitus literally.

⁹² ibid., 37–9, supporting the light verbal echo with arguments from context.

⁹³ To construct such a relationship is by no means to abrogate agonistic intent. On the dynamics of filiation in epic, Hardie, op. cit. (n. 86), 88–119, is fundamental.

⁹⁴ Pliny rarely makes such an obvious juxtaposition of extremes of length (and tone), but compare 1.20–1 and 2.1–2. 9.13 is bracketed by short notes (which are much commoner in Book 9 as a whole), but 9.14, even at first glance, is anything but light.

⁹⁵ Terentius Junior has already completed an equestrian and senatorial career when Pliny first mentions him in 7.25 (Sherwin-White, op. cit. (n. 1), 434). His other letter, 9.12, is concerned with paternity (note also 7.25.3, ‘ut bonum patrem familiae’).

⁹⁶ The pair beomes common in Book 9: see 9.16, 9.20, 9.28.1–2; cf. 8.21.6, ‘hunc adhuc musteum librum’. It is tempting to speculate that there is a metaphor at play here, as there is with hunting, a ‘not uncommon image for the process of artistic creation, inuentio’ ( Hunter, R. L.The Shadow of Callimachus (2006)CrossRefGoogle Scholar, 31; on hunting in Pliny, see Edwards, op. cit. (n. 12), 43–4).

⁹⁷ One cannot literally take ‘tot uoluminibus’ to refer to the single letter 8.14, but that does not exclude a possible hint. The ultimate combination of epistolary and meta-epistolary reference is 9.40.1, ‘scribis pergratas tibi fuisse litteras meas …’. See also Beutel, op. cit. (n. 7), 157–64, on ‘metabriefliche Aussagen’.

⁹⁸ cf. 7.25, especially 7.25.4, ‘quantum ille legit, quantum tenet!’.

⁹⁹ For possible allusion to Cic., Ad fam. 7.18.2, see Marchesi, op. cit. (n. 8), 228–9. Pliny refers, strictly, to anything he writes henceforth (scripserimus is future perfect: cf. R. Kühner and C. Stegmann Ausführliche Grammatik der Lateinischen Sprache. Zweiter Teil: Satzlehre (1912²), 2.197–9).

¹⁰⁰ Sherwin-White, op. cit. (n. 1), 4, 461. The latter comment both contradicts his own observations about Pliny's fast and loose approach to jurisprudence (ibid., 464, on ss. 19 and 23) and does an injustice to the stylistic panache of ss. 12–23 (for a rare appreciation, see Trisoglio, op. cit. (n. 63), 824 and 828; ibid., 822 suggests an intertext in s. 13 with Luc. 1.98, ‘concordia discors’).

¹⁰¹ A position argued most fully by Heath, M.Unity in Greek Poetics (1989)Google Scholar.

¹⁰² Approaching Pliny's books thematically is a new sport; contrast the vigorous debates over the function of the book in Augustan poetry and, closer to Pliny, Martial's epigrams. See, however, Hoffer, op. cit. (n. 12); Henderson, op. cit. (n. 5); Fitzgerald, W. ‘The letter's the thing (in Pliny, Book 7)’ in Morello, R. and Morrison, A. D. (eds), Ancient Letters. Classical and Late Antique Epistolography (2007), 191–210CrossRefGoogle Scholar; Gibson and Morello, op. cit. (n. 13) on Books 6 and 7. Beard, M. ‘Ciceronian correspondences: making a book out of letters’ in Wiseman, T. P. (ed.), Classics in Progress (2002), 103–44Google Scholar, was a crucial standard-bearer. Bernstein, op. cit. (n. 85, 2008), 220–1, tentatively proposes a theme of paternity and succession woven into Book 8; this suggestion need not be mutually exclusive with mine here and indeed, given the prominence of the paternity theme in 8.14.4–6 (ibid., 208–10), further supports my contention that the centrality of this letter to its book is more than numerical.

¹⁰³ Roller, op. cit. (n. 32) and Fitzgerald, W.Slavery and the Roman Literary Imagination (2000)CrossRefGoogle Scholar are fundamental; see recently Edwards, C. ‘Free yourself! Slavery, freedom and the self in Seneca's Letters’ in Bartsch, S. and Wray, D. (eds), Seneca and the Self (2009), 139–59Google Scholar. On slavery in Greek political discourse, and the impossiblity of detaching ‘literal’ from metaphorical usage, see briefly DuBois, P.Slaves and Other Objects (2003)Google Scholar, 117–30.

¹⁰⁴ I follow Pliny himself, and Latin usage, in minimizing at will the legal distinction between slaves and freedmen (n. 122).

¹⁰⁵ Note especially the bridge between Britain and Rome: the Britons are astonished that Roman soldiers are frightened of the freedman Polyclitus, being people ‘among whom the flame of freedom even then still burned, and where the power of freedmen was not yet known’ (‘apud quos flagrante etiam tum libertate nondum cognita libertinorum potentia erat’, An. 14.39.2). The two Roman affairs are introduced with provocative, alliterative juxtaposition: ‘in the same year notable crimes were committed at Rome, one by a senator's audacity, another a slave's’ (‘eodem anno Romae insignia scelera, alterum senatoris, seruili alterum audacia, admissa sunt’, An. 14.40.1). On the interplay, see Roberts, op. cit. (n. 33) and Lavan, M.Rome and the Provinces (forthcoming)Google Scholar, ch. 3.

¹⁰⁶ There are differences too: Cassius is severe, Pliny clement; Tacitus’ debate focuses on the slaves (though freedmen are mentioned at An. 14.45.2, where they escape relegation), Pliny's on the freedmen (though slaves are mentioned at 8.14.15).

¹⁰⁷ Ginsburg, J. ‘In maiores certamina: past and present in the Annales’ in Luce, T. J. and Woodman, A. J. (eds), Tacitus and the Tacitean Tradition (1993), 86–103Google Scholar, at 96–102 (with earlier bibliography).

¹⁰⁸ Pliny is no less mordant than Tacitus on the social inversion of earlier principates (nn. 124–5). The theme is likewise recurrent in the Domitianic world of Juvenal's first book (e.g. Sat. 1.26–8, 1.101–5, 2.58–60, 3.131–2, 5.72–5); note especially the imperial slave shouting ‘run!’ to the amici principis in 4.75–6 (a bite missed by recent commentators). In all three writers, then, such ‘status anxiety’ (Roller, op. cit. (n. 32), 264–72) finds expression, with varying nuance, in periodized terms.

¹⁰⁹ I make a necessarily speculative further suggestion, that Pliny's connection of Tacitus with this senatorial debate of a.d. 105 might respond to a particular episode in the Histories. Tacitus’ proemial promise of ‘corrupti in dominos serui, in patronos liberti’ (H. 1.2.3) as one of the climactic crimes of the Flavian age suggests a notorious case involving freedmen in the 90s of which we are ignorant. Given further Tacitus’ thematization in H. as in An. of senatorial ‘enslavement’ (above, n. 57), it is conceivable, if hardly certain, that Pliny's diptych is influenced by a lost portion of H. (or vice versa).

¹¹⁰ In imposing seruitus on Britons, Agricola is doing the job of a Roman general — but the obvious links between Calgacus’ rhetoric and the position of the ‘Stoic martyrs’ inevitably blur any tidy boundaries. The pinpoint dilemma is ‘obsequium ac modestiam’ (Ag. 30.3 and 42.4), which I will not open up here; see Whitmarsh, T. J. G. ‘“This in-between book”: language, politics and genre in the Agricola’ in McGing, B. and Mossman, J. M. (eds), The Limits of Biography (2007), 305–33Google Scholar.

¹¹¹ Likewise, of course, Annals 14.

¹¹² The responsion of s. 2, ‘priorum temporum seruitus’ and s. 12, ‘referebatur de libertis’, beginning ‘digression’ and ‘main letter’ respectively, pushes the parallel. Is there allegory at some level? It was uncertain whether Dexter had been killed ‘in crime or through loyalty’ (s. 12, ‘incertum … scelere an obsequio perempti’): what about the ‘crime’ and/or ‘loyalty’ of Domitian's familia, the Senate, in his death — and their complicity in allowing him to live until then? I am grateful to William Fitzgerald for provoking these (necessarily open-ended) questions.

¹¹³ Pliny does not specify his status, though ‘quidam ex meis’ (‘some of my people’, 8.1.1) confirms that he is slave or freedman; the latter seems more likely, given that Zosimus, another afflicted reader, is ‘libertus meus’ (5.19.2).

¹¹⁴ All three letters add to Pliny's repertoire of humanitas towards slaves and freedmen (programmatically established in 1.4), as further represented by his advocating leniency in the debate of 8.14. He explicitly defends it, just after 8.14, in 8.16.3–4; meanwhile, recollection of the thematically similar 3.14 (another Letter 14, and another letter which unites two segments in part through the theme of slaves), where Pliny approves the punishment of Larcius Macedo's familia for his murder, protects him from appearing as a soft touch. On Pliny and slaves, see at length the statistics of Gonzalès, A.Pline le jeune: esclaves et affranchis à Rome (2003)Google Scholar. Roy Gibson suggestively proposes that the ‘melancholy’ of the illness letters forms part of the gradual ‘darkening’ of the collection (towards the dusk of 9.40: above, n. 84).

¹¹⁵ Pliny jokes that he is ‘still prolonging the Saturnalia’ (‘ego adhuc Saturnalia extendo’, 8.7.1), and Marchesi, op. cit. (n. 8), 108–9, sees humour lurking in ‘magister’, evoking both ‘teacher’ and ‘slave-master’. Given that a teacher was very often a slave, the wit may be more complex: the suggestion in 8.7.2 that in ‘playing the magister’ (‘sumam … personam magistri’) Pliny will ‘exercise a right’ (‘exseram … ius’), act ‘more freely’ (‘liberius’), and (in an admittedly common phrase) ‘send/release’ no books (‘missurus sum’) does not lack paradox.

¹¹⁶ On this productive Senecan metaphor (e.g. Ep. 47.17), see Edwards, op. cit. (n. 103). Pliny has Seneca's famous forty-seventh letter in mind in 8.16 too (Cova, P. V. ‘Arte allusiva e stilizzazione retorica nelle lettere di Plinio’, Aevum 46 (1972), 16–36Google Scholar, at 24–7). Slavery also intrudes, rather differently, into 8.18, with Domitius’ complaint that he must lick his slaves’ fingers daily (8.18.9: i.e. be helped to eat).

¹¹⁷ Whereas in 8.1.3 Encolpius enjoys treatment from ‘attentive doctors’ (‘medici diligentes’), as William Fitzgerald points out.

¹¹⁸ In the same sentence, Pliny urges him not to appear ‘inexperienced and unfamiliar’ (‘rudis et incognitus’) to the locals. Do we see once again Tacitus’ language of the literary voice (Ag. 3.3, ‘incondita ac rudi uoce’; cf. 8.14.3, ‘rudes nos et imperitos’, Section III above) being appropriated for governance, this time provincial? Tacitus aside, the correspondence with 8.14.3 highlights the complementarity of the two letters’ exemplary projects.

¹¹⁹ See Lavan, op. cit. (n. 105), chs 2 and 3 (including discussion of this letter).

¹²⁰ Zucker, F. ‘Plinius epist. VIII, 24: Ein Denkmal antiker Humanität’, Philologus 84 (1928), 209–32CrossRefGoogle Scholar, at 219–32; Sherwin-White, op. cit. (n. 1), 477; Cova, op. cit. (n. 116), 28–35; Marchesi, op. cit. (n. 8), 223–4.

¹²¹ Roller, op. cit. (n. 32), 271–2. On this letter, see d’Escurac, H. Pavis ‘Pline le Jeune et l’affranchi Pallas (Ep. 7.29; 8.6)’, Index: Quaderni Camerti di studi romanistici 13 (1985), 313–25Google Scholar; Beutel, op. cit. (n. 7), 247; Gonzalès, op. cit. (n. 114), 125, 133–6.

¹²² Pallas was of course a libertus, but seruus makes for eloquent paronomasia. Pliny later calls him mancipium (8.6.14), a word Tacitus applies to freedmen with similar disdain at H. 2.57.2 and 3.47.1; as 8.6.16 (‘quod dari liberto promitti seruis uidebant’) reminds us, today's freedman is yesterday's slave. Pliny and Tacitus exploit a form of disparagement common already in Cicero (see Treggiari, S.Roman Freedmen During the Late Republic (1969)Google Scholar, 265–6). Naturally, Pliny's views of freedmen are context-specific: Trajan's are a different matter altogether, as Trajan himself is allowed to clarify at 6.31.9: ‘he [sc. the imperial freedman Eurythmus] is no Polyclitus, and I am no Nero’ (‘nec ille Polyclitus nec ego Nero’); note the courteous references at Epp. 10.27–8, 63, 67, 84–5, and cf. below, n. 125.

¹²³ For Pliny obsequium is usually a desiderandum (e.g. 8.14.7), but in emphasizing here misplaced compliance (8.6.11 ‘obsequi’, 8.6.12 ‘obsequereter … obsequio … obsecuturi … obsequi’), he draws close to the debased, servile obsequium which pervades Tacitus’ narrative of successive principates (e.g. H. 1.15.4). On this complex term, see Vielberg, M.Pflichten, Werte, Ideale. Eine Untersuchung zu den Wertvorstellungen des Tacitus (1987)Google Scholar, 131–4; Beutel, op. cit. (n. 7), 95–8; M. Pani ‘Sulla nozione di “obsequium” in Tacito e Plinio il Giovane’ in idem, Potere e valori a Roma fra Augusto e Traiano (2003²), 159–80.

¹²⁴ The inversion continues in 8.6.12, with Pallas giving orders to his patron Caesar (‘imperat enim libertus patrono’) in familiar anti-Claudian invective (cf. Sen., Apoc. 6.2, 15.2; Juv. 14.331; Tac., An. 11.35.1 (e.g.); Suet., Claud. 28–9; Dio 60.2.4–5). Does Pliny have Seneca's ‘Saturnalicius princeps’ (Apoc. 8.2) in mind when he fills the following letter with Saturnalian jokes (n. 115)? The inversion features also in Pan. 88.1 (with reference not least to Claudius and Pallas): ‘plerique principes, cum essent ciuium domini, libertorum erant serui’, etc. (for the wordplay, cf. Sall., H. 1.55.2, Vell. 2.73.2, Tac., H. 1.36.3).

¹²⁵ In the Pan., Pliny praises Trajan for putting freedmen back in their place (Pan. 88.2), restoring the social hierarchy more broadly (Pan. 42.2, ‘reddita est amicis fides, liberis pietas, obsequium seruis …’), and correcting, analogously, power relations between Rome and the provinces (Pan. 31.3, ‘seruiat [sc. Aegyptus]’, 11.5–12.1, 32.2). Earlier in Book 8, he showed the correct place for slaves, out of politics and in the home: ‘seruis res publica quaedam et quasi ciuitas domus est’ (8.16.2, ‘as far as slaves are concerned, a house is a sort of republic and quasi-state’; cf. Sen., Ep. 47.14, with Cova, op. cit. (n. 116), 26).

¹²⁶ cf. 8.6.17, where he apologizes for (i.e. advertises) his generically exceptional aggression (‘indignationem quibusdam in locis fortasse ultra epistulae modum extulerim’). The Pallas decree does indeed (later) draw a bite from Tacitus at the Senate's praise of a rich ‘libertinus’ for his ‘antiqua parsimonia’ (An. 12.53.3).

¹²⁷ cf. 1.6, 7.20, 9.14, each following a letter advertising Pliny's links to the victims of a.d. 93. I consider this more closely in Whitton, op. cit. (n. 15).

¹²⁸ Trajan left Rome for his second Dacian war on 4 June 105, not returning until mid-107 (F.Ost.: see Vidman, op. cit. (n. 2), 46–7). Dexter died on 24 June (ibid.).

¹²⁹ For eloquentissimus, see 2.1.6 and cf. 2.11.17. The praise is rare, and strong.

¹³⁰ cf. Haynes, op. cit. (n. 28), 156 n. 27, on Ag. as both ‘performative’ and ‘constative’.

¹³¹ As we have seen, the apology of ss. 2–12 provides dramatic justification for the consultation of Aristo. Contra: Haynes, op. cit. (n. 28), 160. Beutel, op. cit. (n. 7), 258–62, reads this letter as an ironic exhortation to Pliny's fellow senators, since he finds ‘breue tempus’ (s. 12) a stretch for the eight years since Domitian's death. But temporal terms are notoriously flexible, and the rhetoric of s. 12 depends on the contrast of ‘breue tempus’ with the phrase ‘multos per annos’, applied to Domitian's principate in s. 9 (n. 47). As in 9.13 (another belated return to the heady days immediately after Domitian's accession, and a climax of the Epistles as a whole), Pliny works to keep the accession rhetoric ringing even as the collection approaches its close and Trajan approaches his decennalia.