In the film "Spartacus" (1960) Marcus Publius Glabrus, having just lost six cohorts of the garrison of Rome, in an ill-starred attempt to crush the slave-uprising in its incipient stages, is summoned to the Senate for a debriefing. Senator Gracchus rises to his feet:

"If we had punished every commander, who made a fool of himself, we'd have no-one left above the rank of centurion."

Translating this into both direct & indirect speech: it's a conditional sentence; impossible conditions (it did not happen: we did not punish every commander). In advanced texts e.g. Allen & Greenough "impossible conditions" is called "contrary-to-fact" i.e. "counterfactual".

FIRST CLAUSE: (The protasis: a statement of the condition.)

A counterfactual condition, in the past tense, requiring a pluperfect subjunctive for the verb.


A relative "qui" clause. In direct speech the verb is in the indicative.

THIRD CLAUSE: (The apodosis: the result of the condition.)

The present-tense consequence of action/ inaction, in the past, requiring the imperfect subjunctive.


"si puniissemus (punivissemus) omnem imperatorem, qui se ludificatus est, neminem (relinquentem) maioris ordinis centurione haberemus."


Gracchus told the Senators that if they had punished every commander, who had made a fool of himself, they would have no-one left above the rank of centurion."

Rules for changing a condition contrary-to-fact into indirect speech are given in (A & G) section 589; 3(b): p.383; (https://dcc.dickinson.edu/grammar/latin/conditions in indirect discourse)


The protasis always remains unchanged in tense.


The relative clause will require the accusative-infinitive construction for indirect speech.


(A & G): "The apodosis, if active, takes a peculiar infinitive form, made by combining the participle in -urus with fuisse."


Gracchus Senatoribus narravit si punissisent (punivissent) omnem imperatorem, quem se ludificatum esse, neminem (relinquentem) maioris ordinis centurione habituros fuisse.

Are the two translations correct?

  • 2
    In the indirect discourse version, I'm not sure that the verb in the relative clause should be infinitive. Doesn't that happen only when the relative is functioning as a so-called connecting relative (= et/sed is)? I would think subjunctive is needed in this instance.
    – cnread
    Jun 2, 2021 at 17:01
  • 2
    This is an excellent question which touches upon several interesting corners of grammar; minor point (which could distract from the central questions): punissisent should be punivissent. I also doubt that se ludificari is a good translation for "make a fool of oneself," it sounds Latinglish to me, but I could be wrong. Jun 2, 2021 at 18:14
  • 1
    I'd think the relative clause would be subjunctive even in the direct version (relative clause of characteristic).
    – TKR
    Jun 3, 2021 at 0:54
  • 2
    @TKR No, because he is not talking about a characteristic ("the sort of commander who makes a fool of himself"), but about those commanders who, however characteristic or not, did so in the actual campaign. That said, I would argue in favour of the subjunctive too, due to attraction of mode. Jun 3, 2021 at 10:02
  • 2
    @tony That is a rather literal translation though for an idiom which really means "to act stupidly in a disgraceful way. For lack of a better idea, I'd say: stultitia se dedecorare. Jun 3, 2021 at 11:28

1 Answer 1


A number of points:

  • I do not think se ludificari is an appropriate translation for “make a fool of oneself.” As I wrote in the comments, I would prefer: ⋯ qui stultitia se dedecoravit (who disgraced himself through foolishness).

  • Relative clauses in counterfactual condicional clauses often take on the subjunctive of the governing clause, which is called attractio modi. The pluperfect subjunctive is dedecoravisset, which is quite a mouthful, though it can at least be shortened to dedecorasset.

  • Relinquere does the one who leaves. The one who is left is therefore relictus. There is also a related, more or less synonymous word named reliquus. Instead of neminem relictum/reliquum haberemus you could also say neminem retineremus, but I am not sure that is necessarily better.

  • I am also unsure about maior ordo centurione. I think a higher rank should at least be superior instead of maior. I also changed the word order with Hor. C. 3,30 in mind ;-)

So I would translate the direct-speech version thus:

Si puniissemus omnem imperatorem, qui stultitia se dedecorasset, neminem ordinis centurione superioris reliquum haberemus.

For the second part:

  • Narravit seems an unusual choice, I would have gone with a simple dixit.

  • Punissisent should be punivissent (or puniissent).

  • Relative clauses in an AcI context do not become AcI themselves. They retain the finite verb, but stand in the subjunctive mood, following the c.t.

So we end up with:

Gracchus Senatoribus dixit si punivissent omnem imperatorem, qui stultitia se dedecorasset, neminem ordinis centurione superioris reliquum habituros fuisse.

  • Does puniissemus have an extra I?
    – Joonas Ilmavirta
    Jun 4, 2021 at 15:25
  • @Joonas llmavirta: According to Wiki's conjugation tables; I made the mistake of a single "i" in the original posting--if it is a mistake?
    – tony
    Jun 4, 2021 at 15:34
  • @tony I see, I read too hastily. I'm not sure if the I at the end of the perfect stem is retained in all forms. Contracting to a single long I is appealing, but I'm not sure if it's done.
    – Joonas Ilmavirta
    Jun 4, 2021 at 15:59
  • @Joonas llmavirta: Concepts e.g. "relative clause of characteristic" & "attraction of mode" are completely new to me. The latter I am still asking Seb about--it appears to be the legitimising of a mistake--do you have a thought on this one?
    – tony
    Jun 5, 2021 at 8:51
  • @tony Those concepts are well worth studying if you want to dig deeper into Latin grammar. It makes some sense to call all language evolution "legitimization of mistakes", and I see nothing deeply different in the various attraction phenomena (of mode and case) in Latin. Things don't always go as you'd expect, but language is not exactly a logic puzzle.
    – Joonas Ilmavirta
    Jun 5, 2021 at 19:44

Your Answer

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

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