Elamur: **Mova** · [Parse](https://mova.elamur.ai/free/en/parse/) · [Dictionary](https://md.dictionary.elamur.ai/free/en/) · [Translate](https://md.translate.elamur.ai/free/en/) · [Research](https://md.research.elamur.ai/free/en/) · [MCP](https://mcp.elamur.ai/) · [Groupants](https://md.groupants.elamur.ai/free/en/)

[Mova](https://md.mova.elamur.ai/free/en/) · [this page in HTML](https://mova.elamur.ai/free/en/docs/languages/la/10-subordination/)

[home](https://md.mova.elamur.ai/free/en/) / [docs](https://md.mova.elamur.ai/free/en/docs/) / [languages](https://mova.elamur.ai/free/en/docs/languages/) / [la](https://md.mova.elamur.ai/free/en/docs/languages/la/) / 10. Subordination and the moods

# 10. Subordination and the moods

Classical Latin builds long periods: a main clause with participles, ablative absolutes and several subordinate clauses around it, each marked by its conjunction and its mood. The **subjunctive** is the mood of most subordinate clauses. It marks a clause as dependent, wished, purposed, possible or reported, rather than asserted by the speaker. **Indirect speech** turns statements into the accusative with the infinitive and questions and commands into the subjunctive. The **sequence of tenses** ties the tense of a subjunctive to the tense of the main verb. This chapter gives these in turn, then the conditional sentences, then how the system changed in late Latin.

## Contents

1. The subjunctive in main clauses
2. The sequence of tenses
3. Indirect statement: the accusative with the infinitive
4. Indirect questions and indirect commands
5. Purpose and result: *ut, nē, ut nōn*
6. *Cum* and the clauses of time
7. Causal and concessive clauses
8. Conditional sentences
9. Relative clauses with the subjunctive
10. Indirect discourse (*ōrātiō oblīqua*)
11. The change in late Latin: *quod, quia, quoniam*
12. How Mova handles it

---

## 1. The subjunctive in main clauses

In a main clause the subjunctive expresses (Bennett 1918 §§272–280):

- **a wish** (with or without *utinam*; *nē* when negative): *Valeās!* "May you fare well!", *Utinam adesset!* "If only he were here!";
- **an exhortation** in the first person plural: *Eāmus!* "Let us go!";
- **a prohibition** with *nē* + perfect subjunctive: *Nē timueris!* "Do not be afraid!" (or *nōlī* + infinitive: *nōlī timēre*);
- **a deliberative question**: *Quid faciam?* "What am I to do?";
- **possibility** (the potential): *Dīcat aliquis* "Someone may say"; *crēderēs* "you would have thought".

## 2. The sequence of tenses

A subjunctive in a subordinate clause takes its tense from the main verb (Bennett 1918 §§266–268):

| main verb | same time or later | earlier |
| --- | --- | --- |
| **primary** (present, future, future perfect, imperative; the perfect "have done") | present subjunctive | perfect subjunctive |
| **secondary** (imperfect, perfect "did", pluperfect, historic present) | imperfect subjunctive | pluperfect subjunctive |

> *Rogō quid faciās.* "I ask what you are doing." — *Rogō quid fēcerīs.* "I ask what you did." *Rogāvī quid facerēs.* "I asked what you were doing." — *Rogāvī quid fēcissēs.* "I asked what you had done."

Later action in an indirect question is the periphrasis with the future participle: *Rogō quid factūrus sīs* "I ask what you will do".

## 3. Indirect statement: the accusative with the infinitive

After verbs of saying, thinking, knowing and perceiving, a statement is reported as an **accusative and infinitive** (*accūsātīvus cum īnfīnītīvō*, AcI): its subject is in the accusative, its verb in the infinitive (Bennett 1918 §§330–331).

> *Dīcit **Caesarem venīre**.* "He says that Caesar is coming." *Dīxit **Caesarem vēnisse**.* "He said that Caesar had come." *Dīcit **Caesarem ventūrum esse**.* "He says that Caesar will come."

The infinitive's tense is relative to the main verb: the present for the same time, the perfect for earlier, the future for later. A reflexive pronoun (*sē, suus*) in the AcI refers to the subject of the verb of saying: *Dīcit **sē** venīre* "He says that *he* (himself) is coming"; *eum* would be another man. The same construction is the subject of impersonal expressions: *Constat **terram esse** rotundam*.

## 4. Indirect questions and indirect commands

**An indirect question** keeps its question word, and its verb goes into the subjunctive, following the sequence: *Nescio **ubi sit**.* "I don't know where he is." *Quaesīvit **num vēnissem**.* "He asked whether I had come." (Bennett 1918 §§295–296, 300)

**An indirect command** after verbs of ordering, asking, urging, persuading has *ut* (negative *nē*) and the subjunctive: *Imperāvit **ut** mīlitēs **venīrent**.* "He ordered the soldiers to come." *Iubeō* and *vetō* take the AcI instead: *Iussit mīlitēs venīre.*

**Verbs of fearing** take *nē* for what is feared to happen and *ut* (or *nē nōn*) for what is feared not to happen: *Timeō **nē veniat*** "I fear he may come"; *Timeō **ut veniat*** "I fear he may not come". (The French expletive *ne* after *craindre* continues this *nē*: [fr chapter 8](https://md.mova.elamur.ai/free/en/docs/languages/fr/27-negation/).)

## 5. Purpose and result: *ut, nē, ut nōn*

|  | positive | negative | mood |
| --- | --- | --- | --- |
| purpose | *ut* (*quō* with a comparative) | *nē* | subjunctive, by the sequence |
| result | *ut* (after *tam, ita, sīc, tantus, tālis*) | *ut nōn* | subjunctive, usually not by the sequence |

> *Edō **ut vīvam**.* "I eat in order to live." (purpose) *Tam fortis erat **ut nēmō eum vinceret**.* "He was so brave that nobody defeated him." (result)

A relative pronoun can introduce purpose: *lēgātōs mīsit **quī pācem peterent*** "he sent envoys to seek peace". Purpose is also expressed by *ad* + gerund or gerundive, the supine in *-um*, and *causā/grātiā* + genitive of the gerundive ([chapter 6](https://md.mova.elamur.ai/free/en/docs/languages/la/06-nonfinite/)) (Bennett 1918 §§282, 284).

## 6. *Cum* and the clauses of time

**Cum** "when, since, although" (Bennett 1918 §§287–293):

- with the indicative, it dates the main event (*cum temporāle*: *cum Caesar in Galliam vēnit* "at the time when …");
- with the subjunctive in the past, it gives the circumstances of a narrative event (*cum historicum*: ***Cum*** *Caesar in Galliam **vēnisset**, …* "When Caesar had come into Gaul, …"), and often a cause or a concession (*cum causāle, cum concessīvum*).

**Other conjunctions of time** take the indicative when they only date: *postquam* (perfect: *postquam vēnit* "after he came"), *ubi, ut* "when", *simul ac*, *dum* "while" (present indicative even in the past). *Antequam, priusquam* "before", and *dum, dōnec* "until", take the subjunctive when the action is expected or intended.

## 7. Causal and concessive clauses

**Cause.** *Quod, quia, quoniam* + indicative give the speaker's own reason; + subjunctive a reported reason (*quod … dīceret* "on the ground that he said"). *Cum* + subjunctive (§6) (Bennett 1918 §§285–286, 308–310).

**Concession.** *Quamquam* + indicative "although (in fact)"; *quamvīs, licet, cum* + subjunctive "however much, although"; *etsī, etiamsī* like conditions.

## 8. Conditional sentences

| type | *sī* clause (protasis) | main clause (apodosis) | example |
| --- | --- | --- | --- |
| simple, open | indicative, any tense | indicative | *Sī hoc dīcis, errās.* "If you say this, you are wrong." |
| future, more vivid | future (perfect) indicative | future indicative | *Sī hoc dīxeris, errābis.* |
| future, less vivid ("should … would") | present subjunctive | present subjunctive | *Sī hoc dīcās, errēs.* |
| unreal, present | imperfect subjunctive | imperfect subjunctive | *Sī hoc dīcerēs, errārēs.* "If you were saying this, you would be wrong." |
| unreal, past | pluperfect subjunctive | pluperfect subjunctive | *Sī hoc dīxissēs, errāvissēs.* |

Negative *nisi* "unless, if not" and *sī nōn* "if not (a single word)"; *sīn* "but if". The Romance conditional (French *-rais*) does not come from these subjunctives but from the late periphrasis infinitive + *habēbam* (*cantāre habēbam* → *chanterais*): [chapter 13](https://md.mova.elamur.ai/free/en/docs/languages/la/13-late-latin/) (Bennett 1918 §§301–305).

## 9. Relative clauses with the subjunctive

A relative clause takes the subjunctive when it expresses (Bennett 1918 §§283, 312):

- purpose (§5);
- result or characteristic: ***Sunt quī dīcant*** "There are people who say"; *nēmō est **quī nesciat*** "there is nobody who does not know";
- cause or concession: *Ō fortūnāte adulēscēns, **quī** tuae virtūtis Homērum praecōnem **invēnerīs*** "fortunate youth, *since* you found Homer as herald of your valour";
- a clause inside indirect speech (§10).

## 10. Indirect discourse (*ōrātiō oblīqua*)

A whole speech reported in indirect form, as in Caesar and Livy, transposes every clause (Bennett 1918 §§313–322):

- **main statements** go into the AcI;
- **main questions and commands** go into the subjunctive;
- **all subordinate clauses** go into the subjunctive, by the sequence of the verb of saying;
- **the persons** shift: *ego, nōs* → *sē, suus*; *tū, vōs* → *ille, is*.

> Direct: *Ego, sī vōs mē sequēminī, urbem capiam.* Indirect: *Dīxit **sē**, sī **illī sē sequerentur**, urbem **captūrum esse**.* "He said that if they followed him, he would take the city."

## 11. The change in late Latin: *quod, quia, quoniam*

In imperial and Christian Latin, a finite clause with ***quod, quia*** or ***quoniam*** replaces the AcI after verbs of saying and knowing: *scio **quod** filius meus es* rather than *scio tē esse filium meum*. This is the ancestor of the Romance *que, che, que* + finite verb. The subjunctive retreats in the same move: many clauses that took it now take the indicative.

On our data (UD Latin PROIEL training set, NC, counted only; 2026-10-22):

| text | AcI (accusative subject of an infinitive) | *quod/quia* complement clause | subjunctive share of finite verbs |
| --- | --- | --- | --- |
| Caesar, *Gallic War* | 295 | 11 | 32.4% |
| Cicero, *Letters to Atticus* and *De officiis* | 473 | 19 | 29.7% |
| the Vulgate (New Testament) | 241 | 340 | 13.8% |

In Caesar and Cicero the AcI is the norm and the *quod/quia* clause an exception (4%). In the Vulgate the *quod/quia* clause has overtaken it (59%). The open treebanks agree: the subjunctive is 12.5% of the finite verbs in CIRCSE (Seneca's verse, Tacitus) and 12.8% in the LLCT charters (UD 2.18, 2026-10-22).

Greek turned the same way in the same centuries: the accusative with the infinitive against ὅτι clauses is 583 : 69 in Herodotus and 398 : 875 in the New Testament ([grc chapter 14](https://md.mova.elamur.ai/free/en/docs/languages/grc/14-subordination/), §9).

> *Vidit Deus lucem quod esset bona.* (Vulgate, Gen. 1:4) — "God saw the light, that it was good": a *quod* clause after a verb of perception, following the Hebrew *kî* (ki).

## 12. How Mova handles it

**In the UD trees:**

- a subordinate clause is `ccomp` (complement), `advcl` (adverbial), `acl:relcl` (relative) or `csubj`; its conjunction is `SCONJ` attached as `mark`;
- the subjunctive is `Mood=Sub` in the verb's `FEATS`;
- an AcI is a `ccomp` (or `xcomp`) infinitive whose subject is `nsubj` with `Case=Acc`;
- the ablative absolute is `advcl:abs` or `advcl` of a participle (by treebank: [chapter 6](https://md.mova.elamur.ai/free/en/docs/languages/la/06-nonfinite/)).

**In Mova.** `en xpos-fill` maps the subjunctive to an ordinary finite tag, so the mood is lost from the tags (it stays in the `FEATS`). The parser learns the clause relations from the trees. A Latin frames reader needs, in order:

1. **The mood from `Mood`** and the conjunction from `mark`, to read purpose, result, condition (and its type, §8), fear and indirect command as not asserted (`irrealis`/`cond`, as the English reader marks *if, so that, want to*).
2. **The AcI as reported speech:** its subject is the accusative `nsubj`, and *sē* inside it refers to the speaker of the main verb (§3), the cross-clause link the English reader does with *he said that he*.
3. **The sequence of tenses (§2) for time:** a subjunctive's tense gives its time relative to the main verb, not to the speaker's present.
4. **The late-Latin *quod/quia* clause (§11)** read like the AcI, as reported content.

These are the same needs as the French reader's ([fr chapter 11](https://md.mova.elamur.ai/free/en/docs/languages/fr/24-coordination-subordination/)): the two languages are one family seen at two dates.

- [← previous 9. The simple clause: agreement and word order](https://md.mova.elamur.ai/free/en/docs/languages/la/09-simple-clause/)
- [next → 11. Numerals, dates and the lexicon](https://md.mova.elamur.ai/free/en/docs/languages/la/11-numerals-lexicon/)
- [↑ back to Latin (la) — a reference manual for Mova](https://md.mova.elamur.ai/free/en/docs/languages/la/)