I will phrase my question through an example. Consider this sentence in English:

I do not know whether you wrote where you are.

This has one governing clause ("I do not know") and two indirect questions, first subordinate to the governing clause, the second subordinate the first one. In Latin the predicate of a subordinate conjunctive clause follows consecutio temporum (sequence of tenses), and the correct tense depends on that of the governing clause.

The first question is governed by the present tense "I know" (scio) and describes an event before that of the governing clause, so I should use the perfect tense for "wrote" (scripseris). But which of the verbs scio and scripseris is considered to govern the second question? If scio, I should use the present tense for "are" (sis). If scripseris, then imperfect (esses). I am not sure, so I am left with two possible translations:

Nescio, an scripseris necne, ubi sis.
Nescio, an scripseris necne, ubi esses.

Are both of these grammatical and good style in classical Latin? If yes, is there a difference between the two? (The starting point is the meaning of the opening example sentence in English.)

In general, if a conjunctive clause is subordinate to another conjunctive clause, what determines the tense in the "lowest" clause? Does it only depends on the tense of the nearest dominant clause or that of the clause at the top of the hierarchy? I would prefer to see examples from classical literature where such nested clauses appear.

Clauses subordinate to a nominal form were treated here, but the source used there does not discuss the present question.

3 Answers 3


This is what Adolf V. Streng (Latinan kielioppi, 5th edition, 1936) says in §161.2:

Finnish: Toisen tahi kolmannen asteen konjunktiivinen sivulause mukautuu predikaattinsa tempuksen puolesta sitä lähinnä hallitsevan sivulauseen mukaan.

Free translation: A conjunctive subordinate clause of second or third order adapts to the closest governing clause with respect to its tense.

Examples (no citation given):

  1. Nescio, quid causae fuerit, cur nullas ad me litteras dares.
  2. Quid tibi placeat, pergratum erit, si ad me scripseris.
  3. Gratias ago Pisoni, qui non, quid efficere posset in re publica, cogitavit, sed quid ipse facere deberet.

By higher order subordinate clauses he means clauses subordinate to a subordinate clause — exactly what the original post was about.

If Streng is to be believed, the tense of a subordinate clause depends on the immediate governing clause according to consecutio temporum. This also answers the example situation:

Wrong: Nescio, an scripseris necne, ubi sis.
Right: Nescio, an scripseris necne, ubi esses.

  • llmavirta: Couldn't this whole thing have been simplified: "scripseris", perfect (subjunctive) tense, without "have", is an historic (secondary) tense; therefore, by the sequence-of-tense rule, it must link to another secondary tense e.g. "esses" and not "sis" (primary)?
    – tony
    Commented Jan 31, 2022 at 14:31
  • @tony Whether that's the way it is was the whole question. If the main clause is present but a subordinate clause is past, does the tense of the subsubordinate clause depend on the closest dominant clause or the one at the top of the hierarchy? Indeed the natural expectation is that it is the hierarchically closest clause that matters, but it is not obvious that it must go so.
    – Joonas Ilmavirta
    Commented Jan 31, 2022 at 14:36

If we have a subordinate clause depending on superordinate conjunctive clause, we must consider the tense of the conjunctive:

(A) present or perfect "logic" (assimilable to a present), it should be used the tenses prescripted by the consecutio temporum of the primary times, like in Sen. ep. 32,1:

sic vive tamquam quid facias auditurus sim

(B) conjunctive imperfect, pluperfect or perfect "past" (optative or concessive), and then we should use the tenses prescripted by the consecutio temporum of the historical times, like in Cic. Att. 1,5,2:

testis erit tibi ipsa, quantae mihi curae fuerit, ut Quinti fratris animus in eam esset is, qui esse deberet

You could find information on the tenses prescripted by the consecutio temporum of the subordinate clauses in almost all latin syntax reference works, for example:

  • E. Woodcock, A New Latin Syntax, Bristol 1959, pp. 98-108;
  • A. Ernout - F. Thomas, Syntaxe latine, Paris 1964, pp. 547-554;
  • J.B. Hofmann - A. Szantyr, Lateinische Syntax und Stilistik, München 1972, pp- 393-402;
  • A. Traina - T. Bertotti, Sintassi normativa della lingua latina, Bologna 1985, pp. 345-354.

If yes, is there a difference between the two?

The answers here seem to imply that the difference is that one is grammatical and the other isn't, without explaining why. Instead of just saying "this is wrong, this is right", we can try and gain a better insight by taking a speaker's point of view. This will allow us to arrive at a functional explanation, which is that the two sentences in your example (with my corrections) mean two different things:

  • Nesciō utrum scrīpserīs necne, (aut) ubi sīs "I don't know whether you wrote to me or not, or where you are [at the moment]" - here the two questions both depend on nesciō;
  • Nesciō utrum scrīpserīs necne, ubi essēs "I don't know whether you wrote to me or not(,) where you would have been" - this is difficult to parse in either language due to the placement of necne "or not", since ubi essēs could be understood as coordination (above) or subordination (below). A better phrasing is:
  • Nesciō utrum scrīpserīs ubi essēs necne "I don't know whether you wrote to me regarding your whereabouts or you didn't".

As you can see, the tense of the dependent clause serves to determine which main verb it's subordinate to - the present nesciō or the past scrīpserīs - yielding different interpretations.

I replaced an with utrum because an normally introduces the second and less immediately apparent option, thus annōn = rhetorical necne. Besides this, nesciŏ·an also functions as an adverbial meaning "perhaps, I'd say".

Notice that the often confusing distinction between an and aut serves a parallel purpose: nesciō Gāius an Jūlius "I don't know whether it was Gaius or Julius" is the disjunctive, exclusive "or", nesciō Gāius aut Jūlius "I don't know, it was Gaius or Julius" is the conjunctive, inclusive "or". In the latter case Gaius and Julius could be the same person, for which vel is the canonical choice of conjunction.

  • The question is only about nested subordination, where ubi sis/esses is subordinate to a subordinate clause. I agree that the phrasing is less than optimal, but the question is about the choice of tense in such subordination so I tried to strip my examples of other structure. The main question was: Which one of scio and scripseris is considered to govern sis/esses in such nested subordination from the point of view of choice of tense? On that question you seem to answer with the other answers.
    – Joonas Ilmavirta
    Commented Jan 25, 2022 at 20:25
  • @JoonasIlmavirta Yes, but the other answers seem to suggest that only one is grammatical, referring to a prescriptive consecutio temporum. I tried to highlight that both constructions are grammatical, that the last clause can be subordinated to either, and that the choice of tense determines meaning. In this I think the other answers are as unsatisfactory as answering "which is correct mālum dat or mālō dat" without mentioning that either is possible but expresses different meaning. I think grammar makes no sense when divorced from meaning and explaining meaning makes grammar self-evident. Commented Jan 25, 2022 at 20:32
  • In general I agree, but here the meaning was fixed in the question. Perhaps my wording wasn't clear enough. The other meaning with the other form is an interesting side remark.
    – Joonas Ilmavirta
    Commented Jan 25, 2022 at 21:26
  • @JoonasIlmavirta Don't you agree that your question is unlikely to be asked by someone who realises that the choice of tense is not arbitrary, prescriptive etc., and the governing verb isn't "considered" to be one or the other, but is chosen by the speaker based on the meaning they're trying to express? Just like asking the mālum/mālō question is predicated on not realising that the two express two different meanings. This is what I mean when I say that knowing the meaning is knowing the grammar, and conversely omitting reference to meaning cruicially obfuscates the whole issue Commented Jan 25, 2022 at 22:37

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.