¹ Unless noted otherwise, the Greek and the Latin texts cited follow the respective Oxford Classical Texts.

² Lucretian echoes are noted by Boyle, A.J., Seneca's Troades (Leeds, 1994), 220Google Scholar, ad 1009–12, and Keulen, A.J., L. Annaeus Seneca Troades – Introduction, Text, and Commentary (Leiden, 2001), 475Google Scholar.

³ See below, n. 18.

⁴ See Fantham, E., Seneca's Troades. A Literary Introduction with Text, Translation, and Commentary (Princeton, 1982), 364–5Google Scholar for the tradition of the image ‘smoke of Troy’, which goes back to Hom. Il. 21.522–3.

⁵ See e.g. the parallels noted by Fantham (n. 4), 262–71, and Boyle (n. 2), 172 (‘[t]his ode's view and imaging of death are decidedly Epicurean’). I will provide a fuller discussion of this ode's Lucretian presences elsewhere.

⁶ The OLD s.v. divides the lemma into thirteen subdivisions.

⁷ Fantham (n. 4), 365, ad loc.; Keulen (n. 2), 490, ad loc. understands ‘distinguishing mark or feature’.

⁸ See OLD s.v. signum 4, and LSJ s.v. σημεῖον 3, esp. 3b. Hacking, I., The Emergence of Probability. A Philosophical Study of Early Ideas about Probability, Induction and Statistical Inference (Cambridge, 2006 ²), 31–48 CrossRefGoogle Scholar argues that the sign was transformed into the concept of (inductive) evidence in the late Renaissance.

⁹ See Allen, J., Inference from Signs. Ancient Debates about the Nature of Evidence (Oxford, 2001)Google Scholar, with Burnyeat, M.F., ‘The origins of non-deductive inference’, in Barnes, J. et al. (edd.), Science and Speculation. Studies in Hellenistic Theory and Practice (Cambridge, 1982), 193–238 Google Scholar (esp. on the origins of the debate) and D. Sedley, ‘On signs’, in J. Barnes et al. (this note), 239–72 (on the Epicurean treatise De Signis by Philodemus).

¹⁰ Allen (n. 9), 10–11. See also his chapter on ‘Epicurean sign-inference in Philodemus’, 194–241, esp. 234–41. On the (primitive) mode of analogy and its role in early Greek intellectual enquiry, see Lloyd, G.E.R., Polarity and Analogy. Two Types of Argumentation in Early Greek Thought (Cambridge, 1966)Google Scholar. A famous example is the Epicurean belief that the size of the sun, moon and stars is no greater than we see it: see Lucr. 5.564–91, with Bailey, C., Titi Lucreti Cari De Rerum Natura Libri Sex. Edited with Prolegomena, Critical Apparatus, Translation, and Commentary (Oxford, 1947), 1406–10Google Scholar, and Schiesaro, A., Simulacrum et imago. Gli argomenti analogici nel De rerum natura (Pisa, 1990), 67–8Google Scholar.

¹¹ See Soph. OT 916 and frr. 574 and 811 Nauck² (passages noted by Allen [n. 9], 1–2 with n. 2).

¹² Fraenkel, E., Aeschylus Agamemnon (Oxford, 1950), 377Google Scholar, ad loc. notes that ‘it is by the smoke (and by that alone) that the conquered city is even now discernible’ (my emphasis).

¹³ Despite Tarrant, R.J., Seneca Agamemnon (Cambridge, 1976), 10Google Scholar and Tarrant, R.J., ‘Senecan drama and its antecedents’, HSPh 82 (1978), 213–63, at 215Google Scholar, it is not impossible that Seneca knew Aeschylus’ Agamemnon: cf. already C.J. Herington's observations in his review of Tarrant's commentary in Phoenix 32 (1978), 270–5CrossRefGoogle ScholarPubMed, at 273–4, and see also Degiovanni, L., ‘Sui modelli nell’Agamemnon di Seneca: tre note testuali e interpretative’, SCO 50 (2004), 373–95Google Scholar, Lavery, J., ‘Some Aeschylean influences on Seneca's Agamemnon ’, MD 53 (2004), 183–94Google Scholar, and Hutchinson, G.O., Greek to Latin. Frameworks and Contexts for Intertextuality (Oxford, 2013), 21–2CrossRefGoogle Scholar, esp. n. 28. My pointing to the Aeschylean instance of εὔσημος as a precursor of the Senecan use of signum does of course not necessarily require direct use of Aeschylus’ play as a source. Note that wording and emphasis are very different in Virgil's use of the commonplace of smoke rising from captured Troy: saxa uides, mixtoque undantem puluere fumum (Aen. 2.609) and Ilium et omnis humo fumat Neptunia Troia (Aen. 3.3).

¹⁴ See Striker, G., ‘Academics fighting Academics’, in Inwood, B., Mansfeld, J. (edd.), Assent and Argument. Studies in Cicero's Academic Books (Leiden, 1997), 257–76CrossRefGoogle Scholar, at 262 n. 7, with Cicero's free switching between nota and signum in Luc. 33–6.

¹⁵ See also Staley, G.A., Seneca and the Idea of Tragedy (Oxford, 2010), 134–6Google Scholar, who regards ‘signification’ as one of the key components of Stoic poetics and Seneca's idea of tragedy in particular.

¹⁶ Relevant occurrences of the term include its use in Aristotelian logic, and then Stoic and Epicurean philosophy as the basis of inference: cf. e.g. Epicurus, Ep. Pyth. 97 (p. 43 Usener),  τὰ φαινόμενα, ἃ δεῖ σημεῖα ἀποδέχεσθαι (on this passage, see further Allen [n. 9], 201–2), and Phld. Sign. 27 (p. 68 De Lacy), τὸ κοινόν, ὅτι τὰ φανερὰ τῶν ἀδήλων ἐστὶ σημεῖα.

¹⁷ On this, see Burnyeat (n. 9), 211.

¹⁸ The close proximity of signum and uidere in the Senecan line (hoc signo … uidebunt, 1055) is important: on the role of visual perception in the process of σημείωσις, see Fowler, D.P., Lucretius on Atomic Motion. A Commentary on De Rerum Natura, Book Two, Lines 1–332 (Oxford, 2002), 186–95Google Scholar (on Lucr. 2.112–41). Cf. esp. Lucr. 2.125–8, where the motion of dust particles is a visible signum for the motion of atoms: hoc etiam magis haec animum te aduertere par est | corpora quae in solis radiis turbare uidentur, | quod tales turbae motus quoque materiai | significant clandestinos caecosque subesse. Cf. also Seneca's own description of the dust particles at QNat. 5.1.2 (cited by Fowler [this note], 195), retaining the element of visual perception: quod ex hoc intellegas licet: cum sol in aliquem clausum locum infusus est, uidemus corpuscula minima in aduersum ferri, alia sursum, alia deorsum, uarie concursantia. Note, however, that Seneca uses the phenomenon for different ends: unlike the Epicurean Lucretius, he does not want to prove the motion of atoms; rather, he uses it as evidence of the air never being motionless even if it is ‘still’.

* I would like to thank CQ’s anonymous reader and editorial team as well as Tobias Reinhardt for their very useful feedback.