Laudavēre is an (apparently older) alternative to laudaverunt. What is the origin of this ending? Is it connected with any other known endings or affixes?

Clackson & Horrocks say it is from an alternative 3rd-person plural ending -er plus a "primary marker" -i. I don't know this marker: I only know the primary paradigma as something resembling -m/s/t/mus/tis/nt. What is this marker about?

Palmer seems to be saying it contains the "impersonal" -r- also found in the infinitive (laudare) and in passive forms (laudatur). He, too, speaks of a primary ending -ri.

Both sources agree that -ĕrunt is from the perfective suffix -is-, found in laudav-is-se and laudav-is-ti(s) (and related to the -s- found in the Greek sigmatic aorist). The regular 3rd-person ending -ont was stuck on as expected. The suffix -is- (was) then rhotacized into -ir-, which in turn became -er-.

This would make me suspect that -ere also contains this rhotacized suffix -is-, with some mysterious suffix -e at the end. Neither of the sources mentioned seems to allow for this interpretation, but they don't fully explain -ere either: is this at all possible? Or are there any other theories?

  • Interesting question! I can't wait to see an answer. (Side note: Pings in comments don't reach anyone who has not edited or commented this very post.)
    – Joonas Ilmavirta
    Apr 24, 2016 at 20:16
  • 1
    The IE primary suffixes are –mi, -si, -ti etc. The secondary suffixes are –m, -s, -t etc. That is why i is called a primary marker.
    – fdb
    Apr 24, 2016 at 20:43
  • 2
    The "primary" -i is misnamed, as the endings that contain it are historically secondary: the original personal endings are -m, -s, -t..., but in the present tense the so-called "hic et nunc" deictic marker -i was added to these, giving -mi, -si, -ti.... (At least, this is the case with the singulars and the 3pl., though apparently not with the 1pl. and 2pl.) This pattern is still clearly visible in Greek, but in the Italic languages the -i was either lost or perhaps was never added in the first place.
    – TKR
    Apr 24, 2016 at 20:52
  • 1
    Btw your "Bauer" link actually goes to Palmer's The Latin Language.
    – TKR
    Apr 24, 2016 at 20:56
  • 1
    @TKR. In this context "primary" means "used in primary tenses" (e.g. Greek present indicative) and "secondary" means "used in secondary tenses" (e.g. imperfect). This is standard terminology in IE studies.
    – fdb
    Apr 24, 2016 at 22:44

1 Answer 1


(This answer is based on Weiss's Historical and Comparative Grammar of Latin, which is usually the place to go for this kind of thing.)

The most common Indo-European 3pl active ending is -nt(i), which is part of the familiar set -m(i), -s(i), -t(i), ..., -nt(i). (The 1pl and 2pl are a bit harder to reconstruct because they vary more in the individual languages). In these endings, the -i is originally a deictic marker meaning something like "here and now" (often referred to as the "hic-et-nunc i"); this is why the endings that have -i tend to be present-tense endings, while those that lack it tend to be past-tense endings. This situation is seen fairly clearly in Greek, where athematic verbs have e.g. pres. 1sg. -mi, while in the impf. and second aor. the 1sg. ending is -n, from earlier -m.

But there was also another, completely different set of active personal endings in Proto-Indo-European. These show up mostly in the perfect, where e.g. Greek has 1sg. -a, completely unrelated to -m. The 3pl ending of this set was -rs or -ers, and the latter became -ēr by a regular sound change (Szemerényi's Law) already in PIE.

This -ēr, with the addition of the hic-et-nunc -i, is what's thought to underlie the Latin pf. 3sg -ēre (which would originally have been -ēri, with a regular sound change of final -i to -e, also seen in e.g. the 3rd declension abl. sg., as pede from earlier pedi). It's a bit unclear to me why -i would be added to a non-present form, but then the Latin perfect does sometimes have present meaning ("they have gone" as opposed to "they went").

The ending -ērunt would then be, as you say, the result of a mix of -ēre with a different existing ending, -ĕrunt.

The reason that -ēre can't be descended from -is-e (or -is-i) is the long -ē-, which there's no regular way of deriving from the short -i- of -is-.

  • This is exactly what I was looking for, and no doubt worded better than its basis (although I have never seen Weiss). It's very interesting that the deictic i should be used in so many present endings. I may have known this, but I sure didn't remember. I remember it only from Latin haec and Greek touti. I presume imperatives on -i in Greek also display the deictic i, like gnôthi and...I think there were even imperatives on -ti?
    – Cerberus
    Apr 25, 2016 at 1:59
  • Like you, I am surprised that a present ending should end up in a perfect. But classical Greek perfect endings are often the same as present endings, like -mai. Did the Greek and/or Latin perfect originally have primary endings, perhaps? I should ask a separate question about these alternative active personal endings including the -r- forms.
    – Cerberus
    Apr 25, 2016 at 1:59
  • 1
    Do you have any idea why -m remained -m while -mi vocalised into -o, even though -mi doesn't seem any more likely than -m to result in an 'unpronounceable' consonant cluster, in neither Greek nor Latin?
    – Cerberus
    Apr 25, 2016 at 2:00
  • 3
    @Cerberus.The standard view is that IE had two different and unrelated endings for 1st person sing. primary: -ō for thematic stems and –mi for athematic stems. In Sanskrit and Avestan the thematic verbs combine both endings as *-ō-mi > āmi.
    – fdb
    Apr 25, 2016 at 8:24
  • 1
    @Cerberus: It's actually not just the 3pl Latin perfect ending that has this -i, but all the sg. endings as well. Weiss comments (p. 392), "This extension is not surprising given the essential non-past meaning of the old perfect. A similar process happened in Old Church Slavonic..." (where also the original perfect ending was augmented with -i). On -ō, there have been attempts to derive it from an earlier *-om, but these are not generally accepted.
    – TKR
    Apr 25, 2016 at 16:15

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.