As is well-known, the use of "datives of agent" in so-called "passive periphrastic" constructions (formed by the gerundive/verbal adjective with -nd- and the verb esse) like the one exemplified in (1) is very typical and frequent, whereas the use of "ablatives of agent" in these gerundival constructions is often said to be related to stylistic factors: e.g. to avoid ambiguity (e.g. see (2)) or to maintain a syntactic parallelism (e.g. see (3)). In (2) the use of a canonical dative of agent (vobis) instead of the ablative of agent (a vobis) could have caused a parsing problem since the intransitive verb consulere ‘take care for’ already takes a dative argument (quibus). As for (3), the use of the ablative of agent with a gerundive a iudicibus condemnandus is said to be motivated by the presence of the previous participial construction a populo praeteritus est.

(1) Haec vobis provincia est defendenda. (Cic. Manil. 14) 'You have to defend this province'.

(2) Aguntur bona multorum civium quibus est a vobis et ipsorum causa et rei publicae consulendum. (Cic. Manil. 6) ‘The property of many citizens is at stake, which you ought greatly to regard, both for your own sake, and for that of the republic.’ (C. D. Yonge, 1856, Perseus)

(3) Nec, si a populo praeteritus est quem non oportuit, a iudicibus condemnandus est qui praeteritus non est. (Cic. Planc. 8) ‘And it does not follow because a man has been passed over by the people who ought not to have been, that he who has not been passed over is to be condemned by the judges.’ (C. D. Yonge, 1891, Perseus site)

Still, the story is NOT so simple. As pointed by some scholars, there are gerundival constructions like the ones exemplified in (4) and (5), which are typically found in Cicero's works, where the use of the ablative of agent is not obviously accounted for by appealing to merely "stylistic" factors like the ones involved in (2) and (3).

(4) Sed tamen et Crassus a consulibus meam causam suscipiendam esse dicebat. (Cic. Sest. 41)

(5) Ergo haec et agenda sunt ab oratore, quae explicauit Antonius, et dicenda quodam modo (Cic. De Or. 3, 10, 37).

So, putting stylistic questions aside, I was wondering if the gerundival constructions that can be claimed to be truly passive are those ones that have an ablative of agent. That is to say, I was wondering if the example in (1) has an active interpretation, whereas the following example in (6) has a passive one:

(6) Haec a vobis provincia est defendenda. 'This province must be defended by you'.

Danesi, Johnson, and Barðdal (2017) also offer a critique of the traditional claim that a typical gerundive construction with esse like the one exemplified in (1) has a passive interpretation. In my opinion, these authors are correct when claiming that, despite what Latin grammatical tradition says about the so-called (in fact, miscalled) “passive periphrastic”, a typical and frequent gerundival construction like (1), which contains a dative of agent, does NOT involve a passive interpretation. However, their constructionist proposal omits the relevant empirical fact that this gerundival construction is indeed passive in a less typical usage, namely, when an ablative of agent (e.g. a vobis ) is used instead of a dative of agent (cf. the much more frequent (1) with the less frequent (but possible!) Haec a vobis provincia est defendenda). See also the examples (4) and (5), whose existence is omitted by Danesi et al. (2017).

As pointed out by Miller (2000: 340-341) and Danesi, Johnson, and Barðdal (2017: 143ff.), i.a., the alleged passive interpretation in the gerundival construction in (1) can be attributed to the fact that there is agreement between a patient subject (haec provincia ) and the gerundive, which is the typical agreement found with personal passives with perfect participles (Porta clausa est). In my opinion, a strong piece of evidence for the non-passive nature of gerundival constructions with a dative of agent can be claimed to come from the well-formedness of examples like (7). The existence of impersonal constructions like the one in (7), which contains a deponent unaccusative verb like mori ‘to die’, shows that these constructions are not (at least, necessarily) passive. As is well-known, passive constructions can be formed from transitive verbs (personal passives) or, in some languages like Latin or German, from so-called unergative verbs (impersonal passives) but cannot be formed from so-called unaccusative verbs (e.g. see Perlmutter [1978] for this classical generalization). But see this post for some qualifications.

(7) moriendum est enim omnibus ‘All must indeed die.’ (Cic. Tusc. 1, 9, 15)

Notice then that a prediction is that an invented example like (8), which contains the deponent unaccusative verb mori, should be incompatible with an ablative of agent like ab omnibus, which, unlike the dative of agent, does signal the existence of a passive construction (NB: the star * in (8) stands for "ungrammatical"). In contrast, an impersonal passive is predicted to be possible with an intransitive/unergative verb (e.g. see (2)). The predicted ill-formedness of (8) is probably related to the typical nonexistence of impersonal passives with perfect participles formed from deponent unaccusative verbs: *Mortuum est beate (cf. the well-formedness of impersonal passives with unergative verbs: Pugnatum est acriter). Again, see this post for some qualifications.

(8) *moriendum est enim ab omnibus. (cf. the well-formedness of exs. (4) and (5) above).

Is that prediction correct?

In contrast, I would expect that an example like pugnandum est enim a me could be possible (for example, in a context where an ambiguity could arise: e.g. mihi est pugnandum non imperatori sed patri: cf. pugnandum est a me non imperatori sed patri). Interestingly, note that another strategy to avoid this ambiguity is to use the following less classical construction with habere: Pugnandum habebam non imperatori sed patri (Sen. contr. 10, 2, 4).

Similarly, when dealing with gerundival constructions without the verb esse (e.g. see (9)), it is often claimed that a passive meaning is also involved.

(9) placet … contra gaudere nosmet omittendis doloribus... (Cic. fin. 1, 56)

However, the claim that the gerundive construction in (9) has a passive interpretation is not well-supported. For example, Pinkster (2015: 289/295; The Oxford Latin Syntax, vol. 1) concludes the following: “In <ex. (9)>, the ablative gerundival clause omittendis doloribus can only be understood in an active sense, and no obligation is involved (…) interpretation of a gerundive as passive or (more frequently) active depends on contextual and sometimes on extralinguistic information.” NB: Pinkster's (correct) observation/conclusion that gerundives are "more frequently active" should not go unnoticed since it goes against what is often found in many traditional grammars of Latin.

I agree with Pinkster’s conclusion that the gerundive in (9) has an active interpretation and that, unlike (1), no deontic meaning is involved in (9). However, I'm not sure about how to interpret his vague proposal/generalization that the passive or active interpretation depends on "contextual and/or extralinguistic information". For example, as pointed out above, I would say that a passive interpretation is clearly involved when an ablative of agent is found in a verbal context (e.g. cf. the less frequent construction in (6) Haec a vobis provincia est defendenda ‘This province must be defended by you’ with the more frequent one in (1)) or in a non-verbal context like the one discussed in this post: e.g. cf. De hac provincia a vobis defendenda (as discussed in this post, what is interesting of this type of examples is that a dative of agent could not be licensed in this non-verbal context. But see this answer to this post for some important qualifications).

  • in this older question:latin.stackexchange.com/q/15229/1982 and the linked Q., "nasciturus....", some of the above has been covered in the comments we exchanged e.g. gerundives in the oblique cases lose their passive & deontic properties--effectively becoming gerunds (active); though, they must still agree in number, case & gender. I won't repeat everything but I enjoyed reading them all again. I recall the blood, sweat & tears in all the thinking I had to do, at the time--the nature of your questions!!
    – tony
    Commented Jul 10, 2023 at 13:35


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.