Persius opens his fourth satire with an obscene double entendre and a couple of historical presents:(1)
‘Rem populi tractas?’ (barbatum haec crede magistrum
dicere, sorbitio tollit quem dura cicutae)
‘quo fretus? dic hoc, magni pupille Pericli.
2 dura αGL : dira PVXΦ, Isid. Orig. xvii. 9. 71
The tense of tollit in line 2 is a puzzle. It will not do to ignore the problem, to declare ‘tollit = sustulit’ (Gildersleeve), or to pile up examples of the historical present in various authors (Némethy, Bo). Obviously a Latin passage may contain one or more present tense verbs referring to events in the past as if they were happening now, and grammars provide numerous examples of Latin writers shifting in and out of the historical present, often in mid-sentence, with far more freedom than writers of English could ever hope to enjoy. The question here is whether one sentence may contain present-tense verbs that refer to two entirely different times in the past, in this case times that are 35-40 years apart. Persius’ Socrates speaks of (and to) Alcibiades (born ca. 450 B.C.) as barely pubescent, which puts the dramatic date of Persius’ poetic dialogue in the early 430s: so much for dicere. As for tollit, Socrates died in 399, several years after Alcibiades. Persius, writing around A.D. 60, looks back nearly 500 years, then roughly 460, in the same parenthesis. If parallels exist for such a change of times within a single sentence of historical present, they need to be adduced.
Fortunately, there is no need to wait, since a better solution is hiding, as it were, in plain sight. Commentators often quote the (“faintly reminiscent” – Harvey) verbal parallel in Horace, Sat. 2.1.156, where hemlock is mentioned and the same verb is future: sed mala tollet anum uitiato melle cicuta. Would not a future tollet work to separate the two past times in Persius? ‘Imagine, reader, that you are listening to Socrates criticizing young Alcibiades, Socrates who will eventually die from hemlock poisoning.’ The time of tollet, though past to us, is future relative to the time of dicere. This seems perfectly natural to me in English and (so far as I can judge without any native speakers to consult) in Latin. Is there any reason it would not have worked for Persius?
In a brief survey of grammars on my shelves, I have found no discussion of what we might call the ‘relative historical future’. I have, however, run across two clear examples in adjacent passages of Juvenal, describing the lives and future-in-the-past deaths of Hannibal and Alexander. I boldface the historical presents, bold-underline the ‘relative historical’ futures and future perfect (10.159-72):
exitus ergo quis est? o gloria! uincitur idem
[nempe et in exilium praeceps fugit atque ibi magnus] 160
mirandusque cliens sedet ad praetoria regis,
donec Bithyno libeat uigilare tyranno.
finem animae, quae res humanas miscuit olim,
non gladii, non saxa dabunt nec tela, sed ille
Cannarum uindex et tanti sanguinis ultor 165
anulus. i, demens, et saeuas curre per Alpes
ut pueris placeas et declamatio fias.
Vnus Pellaeo iuueni non sufficit orbis,
aestuat infelix angusto limite mundi
ut Gyarae clausus scopulis paruaque Seripho; 170
cum tamen a figulis munitam intrauerit urbem,
sarcophago contentus erit.
It seems to me that Juvenal is doing very much what Persius had done before him, if tollet in 4.2 is correct. Note that Juvenal also uses present subjunctives libeat (162), placeas, and fias (both 167), and the perfect miscuit (163), whose tenses are clearly relative to the historical present, the subjunctives looking ahead to the near future (donec), the perfect back to a time further in the past (olim). I have been unable to find a Juvenal commentator who even notes the tenses of dabunt, intrauerit, and erit: apparently they assume (quite plausibly) that Latin works the same as other Indo-European languages in this respect.
Corruption of tollet to tollit in Persius 4.2 would have been easy: a scribe who realized that the scene was set far in the past (for Persius and even more so for himself) might easily have assumed that a future verb was impossible and that a historical present would do, without considering that present tollit would be (as I have argued) incompatible with the present dicere. In any case, the difference is a single letter.
(1) Both Persius and Juvenal are quoted from Braund’s Loeb (2004).