In many Germanic and Romance languages, the perfect tense is formed with the verb 'to have' or 'to be' plus a past participle. It's easy to find explanations ["I have an arrow (which is) made (by me)" --> "I have made an arrow"] but how did the same construction become so widespread over two branches of I.E.?

  • 1
    Doesn't the commonality suggest that the construction existed in the root language, before they split off into Germanic and Romance?
    – Barmar
    Jan 14, 2015 at 20:55
  • 2
    Stephie, isn't that what I said in the question? Barmar, no I don't believe Latin, for example, had that construction. Can't speak for Proto-Germanic. Jan 14, 2015 at 21:31
  • @DavidGarner: Sorry, must have missed the "be". No offense intended.
    – Stephie
    Jan 15, 2015 at 8:30
  • 1
    And none taken. No-one has jumped in and said, "That's a duplicate", which has surprised me. Jan 15, 2015 at 14:37
  • 2
    @Barmar No, it doesn't.
    – Atamiri
    Jan 16, 2015 at 14:57

7 Answers 7


The issue of the analytic perfect has been studied in the context of grammaticalization but I'm not sure that any one explanation for the spread of the construction has been agreed (perhaps somebody who knows the literature can elaborate). Most likely, different processes led to the emergence of the construction in different contexts.

The analytic perfect is also the feature of the Balkan Sprachbund meaning you can find it in Albanian and Greek as well as Romanian (Romance) and Macedonian (Slavic). This lends credence to the sprachbund explanation. The construction spread through areal contact and borrowing rather than through historical descent (although historical descent will have played a role in some cases).

But it would still be interesting to speculate as to the mechanism of the spread. Generally people talk about direct borrowing or imperfect learning as the main contact change mechanisms. Direct borrowing is a much harder case to be made with syntactic constructions but when one looks at the languages involved, it would be hard to construct a scenario where imperfect learning could account for the change.

Another possibility is simply parallel independent development. One of the features proposed as common to SAE (Standard Average European) was an analytic passive with a copula and a participle. It seems plausible that many languages could develop this type of perfect construction drawing on this features. Perhaps reinforced rather than directly introduced through contact.

It is also important to note that SAE tense and aspect systems are quite complex both diachronically and synchronically so looking at one form in isolation may not be the right approach.

  • Thank-you to Kaleissin and Dominic Lukes for explaining that language development doesn't always have a branching structure. Jan 18, 2015 at 16:00

This construction comes from Latin in the language families you mention but it's not unique to Germanic and Romance languages. This kind of perfect is fully grammaticalized in some Macedonian dialects (Macedonian is a Slavic language), e.g. ima bideno "has been", ima imano "has had", etc. There's also a similar construction in Northwest Russian which expresses possession with a PP and has an analogous perfect tense, cf. u menja dom "I have a house" vs. u menja vse sdelano "I have done everything". In the Lithuanian (i.e., Baltic) dialects of Belarus there's an identical construction: Possession is expressed with the adessive case and so is the agent in a structurally identical perfect. Since neither Russian nor Lithuanian were influenced by Latin there seems to be a general tendency to this kind of grammaticalization as far as the perfect tense is concerned, the other being the be-perfect (e.g., ich bin gegangen "I went" in German). In fact it occurs in many languages but in most of them it's not fully grammaticalized.

  • 1
    Thanks, Atamiri. When you say, "This construction comes from Latin", do you mean that the components - verbs To be and To Have, and past participles - come from Latin? As best as I can remember, Latin didn't have, for example "Habeo visum" for "I have seen". Jan 16, 2015 at 15:33
  • @DavidGarner I mean the grammatical structure. It may come from Vulgar Latin, I don't know at which moment the construction emerged, but it's called "habeo-perfect" by some. I'll try to find more information about its development.
    – Atamiri
    Jan 16, 2015 at 18:28
  • @DavidGarner - Protoslavic had the perfect with "to be" + "past participle". And also it had the plusquamperfect of the same kind, only "to be" was in the imperfect, severalmodern Slavic languages still keep those constructions. So I agree with Atamiri, +1.
    – Yellow Sky
    Jan 17, 2015 at 10:12
  • Thank-you Atamiri and Yellow Sky. There must be something about 'to have' in IE, since it also became part of a future-tense construction in Romance [and, will someone tell me, other] languages. Sprachbund is a fascinating concept [ en.wikipedia.org/wiki/Sprachbund ] which probably answers OTHER questions that I had in mind. Jan 17, 2015 at 11:10
  • 1
    @DavidGarner It seems that Latin had this construction as one could say, e.g., "librum scriptum habeo", but it wasn't fully grammaticalized since it only expressed a state.
    – Atamiri
    Jan 17, 2015 at 14:45

Using "to have" to show perfect is a key feature of the European sprachbund. It's happened repeatedly in European languages. For instance in Latin, than later with a new word for "to have" in descendant languages when the last round of suffixes have become worn down to invisibility. It's a "language hailing from Europe" quirk, and I think "sprachbund" is explanation enough.

I wouldn't call it widespread though: http://wals.info/feature/68A


It might be relevant that the Germanic and Romance languages don't have active past participles, as many other languages do (Indo-European and otherwise). E.g., in English one says "The food was eaten" but not *"The man was having-eaten the food", where "having-eaten" would be the active equivalent of the passive participle "eaten".

Thus, the have-perfect construction (have eaten, etc.) may have spread because it helped to create "symmetry" -- i.e., it created an active counterpart to past passive constructions like was eaten. (However, it's also possible that Germanic and Latin/early-Romance already had a counterpart construction like this, and the have-construction replaced it -- I don't know all the relevant history here.)

Since the Slavic languages do have past active participles -- e.g. Slovene jedel literally means "having eaten" (masculine singular), spal means "having slept", and so on -- it is not surprising that the use of "have" for the perfective has not widely spread to Slavic in the way that it has to Germanic/Romance, since Slavic does not have the same kind of "gap" to be filled. (It would be interesting, though, to know why the "have"-construction has taken root in some Russian and Macedonian dialects, as mentioned above.)

  • Depends on what you mean exactly, there aren't active past participles for transitive verbs, but you could consider the intransitive verbs that take "sein" in German or "être" in French as auxiliaries to have active past participles. Jan 22, 2015 at 23:11
  • In Macedonian, it's through language contact (Aromanian). In NW Russian it might have been via language contact, too (as Trubinsky speculates). But there's a general tendency to use possessive constructions to express state of completion. Czech and Russian have similar constructions that didn't emerge through language contact.
    – Atamiri
    Jan 24, 2015 at 8:42
  • @sumelic The participles of intransitive verbs acquired the active meaning through reanalysis that occurred probably later (when the habeo-perfect was already established). It can be found in many Slavic dialects also (Silesian: jeżech póńdzóny "I have left", literally "I am left", Macedonian: tamu e biden monastir "there was a monastery", lit. "there is been...", etc.) but I don't think it's relevant here.
    – Atamiri
    Jan 24, 2015 at 8:50
  • 1
    @Atamiri - There may be a general tendency to use verbs meaning "have" in certain perfective contexts -- e.g. English I have my trip planned, or Spanish Tengo planificada mi estrategia -- but that seems different from making this construction the default marker of the perfect. How widespread are the constructions you mention in Russian and Czech?
    – user8017
    Jan 25, 2015 at 1:25
  • @user8017 In Russian it's dialectal. In Czech it's used in a restricted context (with a clear reference to the present). It's not a tense (yet?) but it's getting grammaticalized.
    – Atamiri
    Jan 25, 2015 at 1:29

The Russian constructions literally mean "At me house" and 'At me all done' respectively. Since there is no 'have' verb used here, we cannot talk about parallels between these sentences and the have-perfect of Mediterranean-Euro-Atlantic languages. The discussed Russian constructions have their counterparts in Polish. The first one would be 'Ja mam dom'. Here 'Ja mam' means 'I have' and it refers to the fact of possession. The second Russian sentence in Polish must be rendered as 'U mnie wszysko (jest) zrobione (*At me everything (is) done). The verb have (mieć) is not used at all. The message is more or less 'In my section everything is done,' with no indication that the speaker has done it himself. The sentence 'Ja mam zrobione wszystko' (literally 'I have done everything') in Polish means that the job has been done by or for the speaker, with no indication of who has done it. It is not the equivalent of the Mediterranean-Atlantic have-perfect. Furthermore that construction works only with the verb meaning 'do/make'. A sentence '*Ja mam zobaczone wszystko' (literally 'I have seen everything') is meaningless in Polish. As far as I know Slavic Languages do not have the have-perfect construction, with the possible exception of Macedonian quoted above. I would like to know if the have-perfect constructions can be used in Macedonian with participles other than the equivalent of the English 'been,' and ’had.


The Latin perfect feci was later substitued by habeo factum (I have made, perfect with habere to have). This new system abolished with one stroke those numerous perfect forms that have become irregular.

I didn't have time to search for full information about the upcoming of the habeo perfect, here only some short information


Another problem:

We don't know what is behind the Latin verb endings of the perfect tenses, but they didn't fall from heaven. If you compare the endings of past perfect

veram veras verat verámus verátis verant

it looks like the forms of esse were the source: eram eras erat erámus erátis erant or the forms fueram fueras etc.

If you look at the endings

ueram ueras uerat etc

It looks like the source might have been forms of habere ( to have)

habueram habueras habuerat etc

Maybe the forms of esse and habere melted together and gave one basic set of morphems for the perfect tenses. (f)ueram and (hab)ueram can easily become one morphem.

Assuming that this theory has some probability then the invention of the habeo perfect was nothing else but the shifting of habere at the right side of the verb to the left side of the verb. The advantage was enormous, no melting process, no irregularities.


Here is another link to a website about the habeo perfect


Read the passage On the use of habeo + perfect participle in early and late Latin.

  • Your speculation about the pluperfect (what you call the "past perfect") does not work. The -v- in those forms is from the perfect stem (cf amavi, amavisti). Verbs which use the -s preterite in their perfect (eg scribo, scripsi) use that in their pluperfect. Your suggestion that the forms of esse are involved makes sense, but there is no reason at all to bring in habeo.
    – Colin Fine
    Jan 18, 2015 at 20:56
  • habuit can be shortened to buit and uit and vit and it. The Perfect with s is mostly a Perfect with an inserted s, mostly to distinguish Present as scribit and Perfect *scribit. By inserting s you get an unambiguous Perfect form and, of course, bs changes to ps and you get Perfect scripsit.
    – rogermue
    Jan 18, 2015 at 21:07
  • Yes, but what has that to do with my comment? Your answer, if I understand it correctly, suggests that the -averam etc. represents a form that contains habueram, on the basis that the -v- represents the hab-. That would not be unreasonable, except that my point is that the -v- is present in only some pluperfects: those where the perfect has a -v- infix; so there is no reason to suppose that hab- plays any part in these.
    – Colin Fine
    Jan 19, 2015 at 17:52
  • Actually I said we don't know where the endings of the perfect tenses in Latin come from but it may be that they are tiny remainders if forms of habere or even a mix of habere and esse or fui. But it would be a book to show what might be in those endings and how they evolved. Anyway, they did not fall from heaven, but were logically developed by paraphrases, simple compositions of two or even three parts that somehow expressed the specific time. There is a system in the endings of the perfect tenses that is no randon result.
    – rogermue
    Jan 19, 2015 at 18:43
  • You did say that. You also said "It looks like the source might have been forms of habere ( to have)", and that is the part I was addressing.
    – Colin Fine
    Jan 19, 2015 at 21:32

habueram looks like the original imperfect tense of esse has been bolted onto habui, as a suffix, to tweak the tense. This must have happened very early, as it is fully normal in the literature of the 2nd c. B.C.

Habeo factum et sim. can be observed, occasionally even in Classical Latin prose (i.e. of the 1st c. B .C.).

The 'new' future tense of Romance languages was similarly formed by adding a part of habeo onto the infinitive.

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.