In the Thesaurus of Iohannes Matthias Gesnerus under cogito: (clearer link)

inter cogitare et deliberare hoc interest, cogitare est dubitare, deliberare, confirmare. Haec Nonius 5,28. Interdum sane, qui cogitare se dicit nondum constitutum sibi certi quidquam fatetur. Sed non est perpetua haec differentia. Nam deliberare saepe dicitur qui consilium petit, et cogitat de profectione qui proficisci iam ante constituit.

I have it all except the sentence in bold, which I can't make grammatical. Not sure on so many things (ceteri: adj or gen. of Noun; indirect speech or not - where does it start if so?), but every combination I tried failed to make sense.

(Frustra de hac sentenia cogitabam, nunc igitur delibero et peto auxilium vestrum. Gratias vobis)


Here's how I would translate the bolded phrase:

Sometimes, of course, he who says that he is thinking [cogitare] is claiming that nothing certain has been decided by him yet.

Here's a breakdown:

qui cogitare se dicit...

This is the subject: "He who says that he is thinking"


This is the verb: "is claiming"

nondum constitutum sibi

This is a dative of agent (thanks to TKR for the correction), meaning "not yet decided by him."

certi quidquam

The certi is an example of a partitive genitive, e.g. "quid novi" = "something new." In this case, "quidquam certi" means "something certain." There is an understood "esse" that links this to the previous: "nothing certain has been decided by him yet."

  • 5
    I'd take sibi as dative of agent -- "nothing certain has yet been decided by him" (taking the implied esse with constitutum). – TKR Jul 23 '20 at 19:29
  • 1
    Thanks! this missing esse made it harder for me. How frequent an "implied esse" in indirect speech is used? – d_e Jul 23 '20 at 21:12
  • @TKR I think you're right! Edited accordingly – brianpck Jul 23 '20 at 21:20
  • 2
    @d_e I find that esse is rarely made explicit in indirect speech. – Figulus Jul 24 '20 at 16:16
  • @TKR Descriptively speaking, I agree that here sibi is a "dative of agent". However, to the extent that datives of agent can be theoretically reduced to datives of possession (cf. latin.stackexchange.com/questions/12844/… ), I'd eliminate the correction over "possession". Incidentally, note that the presence of a dative of agent here is a nice evidence for the existence of an elliptical esse. Importantly, as claimed in my link, datives of agent, unlike ablatives of agent, cannot directly depend on a participle. – Mitomino Jul 24 '20 at 16:51

What follows is not an answer but just contains some (grammatically relevant) remarks on the translation given above by brianpck. I've decided to include these remarks here as an answer since I'm afraid I've already posted too many comments in brianpck's post...

Let's start with the following fragment:

nondum constitutum sibi "This is a dative of agent (thanks to TKR for the correction), meaning 'not yet decided by him.'"

This translation could set some learners of Latin like d_e wondering: "fine, but then is there any semantic difference between a "dative of agent" (sibi) and an "ablative of agent" (a se)? Should both be translated as "by him"?

When dealing with the difference between "datives of agent" and "ablatives of agent", some (good) grammars of Latin often include relevant pairs like the following one:

Mihi consilium captum diu est (Cic. Fam. 5.19) 'lit. 'For me there is a decision taken for long time' (NB I: the participle captum is NOT part of the verb(al form) but is rather acting as an adjectival predicative of the nominative subject consilium; NB II: for reasons clarified below, the existential "For me there is..." is better rendered in many languages via a possesive verb: "I have + Direct Object + predicative of DO").

A me consilium captum est 'The decision {has been/was taken} by me' (NB: the participle captum is part of the complex verbal form).

Cf. the following relevant remark made by the French latinist Lavency (1985: 165), which, I hope, is quite transparent for English speakers (otherwise, please let me know): "on notera la différence entre Mihi (Dat.) consilium captum est (présent correspondant à: Mihi consilium captum fuit) et A me (ab + Abl.) consilium captum est (passé correspondant à: A me consilium capitur)". Importantly, note that the participle captum is adjectival in the former example with mihi, whereas it is verbal in the latter with a me (for further discussion on the important distinction between "adjectival passives" and "verbal passives", see this link).

As I've pointed out in some comments on brianpck's answer, the notion of possession is relevant when dealing with so-called "datives of agent". For example, authors of excellent descriptive grammars of Latin like Lavency (1985: 165-166) have noted parallelisms like the one between Caesari equitatus coactus erat and Caesar equitatum coactum habebat (Caes. BG 1.15). Indeed, there is a close relationship between so-called "datives of agent" and "datives of possession". So, for the purposes of the present post, note that a parallelism can also be put forward between the following constructions. The former involves a dative of agent sibi (which is in fact a dative of possession!), whereas the latter involves the possessive verb habere:

nondum constitutum (esse) sibi certi quidquam

nondum constitutum habere se certi quidquam

To recapitulate, consider the full translation given by brianpck:

Sometimes, of course, he who says that he is thinking [cogitare] is claiming that nothing certain has been decided by him yet.

Note that this translation has transformed:

(i) a Latin adjectival passive INTO an English verbal passive (please note that in Classical Latin the adjectival participle constitutum, when coappearing, for example, with a "dative of agent", does NOT form a perfect infinitive with esse);

(ii) a Latin dative of agent (a dative of possession; see above) INTO an English by-phrase (the latter being more similar to the Latin "ablative of agent").

In short, note that brianpck's translation is rather a translation of nondum constitutum a se (certi quidquam), where constitutum does form a perfect infinitive with an elliptical esse.

Please don't get me wrong! I'm NOT saying that there are some "mistakes" in brianpck's translation. The aim of the present "answer" is just to exemplify how some subtle grammatical facts of the original text can change when translating them to another language. As is well-known, translations often involve, i.a., a non-trivial process of (natural) adaptation of the original information to the target language, whereby some subtle nuances in the original text can be lost. To rephrase/reuse an admittedly polemical title (I know some of you strongly disagree with Newmeyer's statement!), "translation is translation and grammar is grammar".

Take-home message: As a Spanish colleague of mine (Ignacio Bosque, a famous grammarian of this language) likes to say, "en la lengua —y en particular en la Gramática— todo son matices" ('in language -and in particular in Grammar- everything is a matter of nuances').


  • For related discussion on the different grammatical behavior (e.g., the different syntactic distribution) of "datives of agent" and "ablatives of agent", see this link.
  • Take a look at this excellent descriptive article on datives of agent (NB: it is not technical at all but is written in French).
  • Romance languages like Catalan (my native language) can account for the minimal contrast Mihi hoc est constitutum and A me hoc est constitutum as follows: '(Jo) tinc això decidit' (lit. 'I have this decided') and 'Això {ha estat/va ser} decidit per mi" (lit. This {has been/was} decided by me'), respectively. – Mitomino Jul 25 '20 at 22:53
  • A couple of ways come to mind in which the classification of the OP's sentence as a dative of possession might be tested. 1 The sentence has an omitted copula; are copulas ever omitted in the possessive dative construction? I don't think I've ever seen that; if it's rare or unattested, that would be a problem for this analysis. 2 Taking the omitted est as a present-tense copula rather than part of a periphrastic perfect with the participle implies that this construction should never trigger secondary sequence; that too is in principle falsifiable, though it may be hard to find relevant data. – TKR Jul 25 '20 at 23:56
  • 1
    To clarify, my "Heri constitutum est a me / *mihi..." was not meant as a description of actual usage, but as a prediction made by the analysis you cite -- I have no idea whether or not it's borne out by the usage facts. – TKR Jul 26 '20 at 4:00
  • 1
    I'm not sure how to decide whether the esse of Nunc est (nobis) bibendum is existential -- if it's never elided, that would be good evidence, but I don't know if that's the case. The parallelism you note is indeed suggestive. But the copula ellipsis difference still argues against the dative of possession analysis in my mind. – TKR Jul 26 '20 at 4:31
  • 1
    That existentials can diachronically acquire obligational sense doesn't mean that they're the same thing synchronically. As I see it: the dative of possession, as that term is normally used, definitionally depends on an existential esse; existential esse definitionally cannot be elided; therefore, despite semantic commonalities, the dative of agent (for which this is not true) cannot be reduced to a dative of possession. (Of course, if it turns out esse can be elided in the possessive dative construction, my argument fails; I've just posted a question on that topic.) – TKR Jul 26 '20 at 5:43

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.