Tacitus, Annales 4.3:

pellit domo Seianus uxorem Apicatam, ex qua tres liberos genuerat, ne paelici suspectaretur.

The translation on Perseus (Church and Brodribb) gives:

Sejanus, to avert his mistress's jealousy, divorced his wife Apicata, by whom he had had three children.

How does ne paelici suspectaretur mean to avert his mistress's jealousy?

Ne suspectaretur is easy: "lest he should be suspected". Paelici is a dative of paelex "mistress". But I'm not seeing what a dative could be doing in this clause.

Apparently, too, there's an alternate reading of this phrase: L&S under suspecto quote this as ne pellici suspectaretur, with a passive infinitive of pellicio "allure, entice". It would then mean "lest he be suspected of being enticed". This at least works grammatically, but I'm not sure what it's supposed to mean.

What's going on with this phrase?


This can be read as a dativus auctoris. It should then be translated thus:

lest he should be suspected by his mistress / be suspect to his mistress

Common in gerundive constructions (hostis nobis vincendus est), the dativus auctoris is also occasionally used with other passive verbs; it is then most common with past participles (mihi cognitum est: "it is known by me" or "it is known to me"). It indicates the person by whom something is done. Especially with verbs expressing a judgement, it is somewhat similar to the ordinary dative of videtur mihi "it seems to me" etc. Lewis & Short say the verb suspectaretur in this particular quotation is passive, i.e. not from the deponent verb suspector.

According to Kühner–Stegmann¹, this dativus auctoris was only used in the perfect and with a personal pronoun in older Latin (pace their own Accius example, see below); but Cicero and various later authors used it with other passive forms and substantive nouns, too. K.–S. also mention Tacitus specifically.

In poetry, it is (especially) used by the Augustan poets, even non metri causa. They say the construction is originally Latin, but was extended under Greek influence (after all, the agent in a passive construction is normally expressed by a dative in Greek).

Here follow a few examples taken from K.–S. and Gildersleeve:

Accius, Tragoediae 284:

Ergo me Argos referam, nam hic sum gnobilis
ne cui cognoscar noto.

"Then I will go back to Argos, for here I am notable, lest I be recognised by someone I know."

Cicero, Tusculanae Disputationes 5.68.4:

"Sumatur...nobis quidam praestans vir..."

"Let us assume an excellent man..."

Tacitus, Annales 2.50.13:

... ut exemplo maiorum propinquis suis ultra ducentesimum lapidem removeretur suasit.

" ...he suggested that, according to ancient custom, she should be removed by her next of kin beyond the two-hundredth milestone."

Tacitus, Annales 12.1.1:

Aelia Paetina...Narcisso fovebatur.

"Aelia Paetina was favoured by Narcissus."

The past participle or adjective suspectus is often used with a dative, which may also have influenced Tacitus to use the dativus auctoris with a passive imperfect (as in the quotation in question). Two examples with suspectus:

Sallustius, Bellum Iugurthinum 70.1.1:

Bomilcar...suspectus regi et ipse eum suspiciens novas res cupere...

"Bomilcar, suspected by the King [or: suspect to the King], and himself suspecting the King, desired a change of power..."

Tacitus, Annales 14.44.10:

Suspecta maioribus nostris fuerunt ingenia servorum...

"To our ancestors, the temper of their slaves was suspect..."

¹) Satzlehre vol. I (1912), §76 8d, p. 324:

enter image description here

  • I think that the following disjunction ("or") can be a bit misleading when giving the translation of mihi cognitum est (cf. your translations, the typical ones, "it is known by me" or "it is known to me"). I think that Latin speakers did distinguish between agent by-phrases and dative "agents". E.g., cf. the parallelism between Caesar equitatum coactum habebat (Caes. BG 1.15) and Caesari equitatus coactus erat. Crucially, the parallelism with habere does not hold for ablative agents: cf. A Caesare equitatus coactus erat.
    – Mitomino
    Jun 9 '19 at 20:50
  • Similarly, Mihi deliberatum est and A me deliberatum est have different (linguistic) meanings. The former is nicely translated in Romance by using the verb HAVE (e.g., Cat. 'tenir'): cf. "Ho tinc decidit" (lit. 'I have this decided') and "Això {ha estat/fou} decidit per mi" (lit. 'This {has been/was} decided by me'), respectively.
    – Mitomino
    Jun 9 '19 at 22:38

It can be useful to consult the relevant commentary made by A. J. Woodman (2018). The Annals of Tacitus (Book IV). Series: Cambridge Classical Texts and Commentaries (vol. 58). Cambridge, UK: Cambridge University Press.

Here is the relevant quote on page 87:

"It is unclear whether paelici is dative of agent (cf. 2.50.3n) or dative of the person concerned (as in alicui suspectus esse); either way, the use of the verb to mean 'to suspect, be suspicious of' seems to have been introduced by Tacitus (see OLD 2)".

According to Woodman, notice then that TWO interpretations are possible here, whereby the two translations provided by Cerberus (cf. lest he should be suspected by his mistress / be suspect to his mistress) correspond to TWO different (descriptive) linguistic analyses: i.e., according to Woodman, only the former translation corresponds to the dative of agent reading.

The expression ne paelici suspectaretur is also briefly commented on in the following book: Garrett G. Fagan & Paul Murgatroyd (2010). From Augustus to Nero: An Intermediate Latin Reader. Cambridge, UK: Cambridge University Press.

Here is the relevant quote on page 60:

"paelici is dative of agent. Sejanus did not want his mistress Livilla to have suspicions about his feelings for her, so he got rid of his wife".

  • In general, agent and experiencer can be distinguished. As to whether the distinction can be made in this particular praedicate frame (suspectaretur), I'm open to it, but I am not yet convinced either way; the fact that the phrase can be translated in different ways is not proof to me. // I think the Aktionsart is very important and must be considered. // Thematic roles interest me very much. // I wasn't sure how the two types of constructions in your comments corresponded to the two thematic roles in the example.
    – Cerberus
    Apr 3 '20 at 3:00
  • P.S. I found this page in an open tab in my browser from long ago (I tend to hoard tabs...); it seems I never got around to replying. I always read your posts and comments with great interest, even though I'm not always good at replying!
    – Cerberus
    Apr 3 '20 at 3:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.