In Latin the similarity between the nominative plural and genitive singular is most striking:

  • First: porta (Nom/Sing) and portae (Nom/Pl), portae (Gen/Sing) and portarum (Gen/Pl)
  • Second: servus (Nom/Sing) and servī (Nom/Pl), servī (Gen/Sing) and servōrum (Gen/Pl)
  • Third: māter (Nom/Sing) and mātrēs (Nom/Pl), mātris (Gen/Sing) and mātrum (Gen/Pl)
  • Fourth: manus (Nom/Sing) and manūs (Nom/Pl), manūs (Gen/Sing) and manuum (Gen/Pl)
  • Fifth: fidēs (Nom/Sing) and fidēs (Nom/Pl), fideī (Gen/Sing) and fidērum (Gen/Pl)

In the First, Second and Fourth declensions, Nominitive Plural is the same as Genitive Singular. The other two don't follow the same pattern, but don't deviate very far from it.

Even in English, we say:

  • dog (Nom/Sing) and dogs (Nom/Pl), dog's (Gen/Sing) and dogs' (Gen/Pl)

There are some other examples I've found especially while studying Russian:

  • ко́мната (Nom/Sing) and ко́мнаты (Nom/Pl), ко́мнаты (Gen/Sing) and ко́мнат (Gen/Pl)
  • There may be an explanation in spoken French. In French, the genitive is indicated by the preposition de (of in English), which has the same pronunciation as the numeral deux (two in English). Hence, the French genitive suggest the idea of plural. And since the preposition is understood to indicate plural, it no longer indicates genitive. Hence you get the nominative plural. I am thinking of preparing an article on this for the Journal of Irreproducible Results.
    – babou
    Jul 1, 2014 at 22:25
  • 1
    I've always wondered this. It is also true for Czech/Slovak and I think other Slavonic languages. But there it's mostly limited to certain feminine and neuter declensions. Can't figure out any good reason for it, though. Of course, it could also be because the declension paradigms are an illusion of grammar writing and the relationship happened on a very different level. Jul 1, 2014 at 23:36
  • 3
    Do you realise, babou, that you are talking complete rubbish? Are you really suggesting that Latin can be explained by means of "spoken French"? And likewise Russian??
    – fdb
    Jul 2, 2014 at 12:17
  • 1
    Also true in Latvian, at least for the feminine declensions in –a and –e: meita ‘girl’ (Nom/Sing), meitas ‘of the girl’ (Gen/Sing), meitas ‘girls’ (Nom/Pl). The same in Lithuanian, but word stress may differ.
    – neubau
    Jul 2, 2014 at 13:42
  • 2
    @fdb, OP is asking for an "intrinsic relationship", so I don't see why something happening in French couldn't be an example of that. I agree that "deux" having anything to do with it is far-fetched, though.
    – dainichi
    Jul 3, 2014 at 3:19

3 Answers 3


I am not sure about “intrinsic”. It is, however, main-stream Indo-Europeanist theory that the suffix * -es marks both the genitive singular and the nominative plural m/f in proto-Indo-European. Though I would concede that this does not really answer your question but merely projects it back to a hypothetical proto-language.

  • 1
    But it is interesting that the Latin 1st and 2nd declensions (-a stem and thematics), which I believe have taken their npl endings (-i and -ae) from pronouns, have taken the same forms for their gsg. I don't know the source of the latter endings (and I lent my Baldi to somebody and haven't got it back) but I suspect that there was at some point a feeling that the npl. and gsg. were linked.
    – Colin Fine
    Jul 2, 2014 at 18:22
  • I agree. There must be some sort of analogy happening here.
    – fdb
    Jul 2, 2014 at 18:35
  • 1
    An obvious guess would be a semantic widening: genitive -> partitive -> plural, but confirming that for IE would probably be hard if it goes back to before PIE. I guess one would have to look for occurrences elsewhere. Maybe the way French uses "de" for both the genitive and the partitive (which is kind of an indefinite plural) can be seen as a parallel.
    – dainichi
    Jul 3, 2014 at 3:02
  • Unfortunately, this falls apart because of the reasons given by @TKR. The genitive singular ending in the 2nd declension was always -i in the oldest Latin, but the plural was -os, -oi, -ei and finally -i. The analogy explanation is probably best, because if the two endings are similar (with -s) in other declensions, keeping them similar in the 1st and 2nd declension would be sensible.
    – siride
    Aug 27, 2020 at 3:02

There is a paper by Pavel Caha on Lingbuzz that addresses this exact question. Put simply, Caha argues, within a nanosyntactic framework, that the underlying structure of gen.sg=nom.pl nouns is something like this:

   [[[gen.sg] N] nom.pl]

That means that any gen.sg=nom.pl noun is viewed as a noun with a genitive singular complement that projects into the nominative plural.

But please, I'm just the messenger. Don't shoot me!


In the Latin first and second declensions the identity of the Nom. pl. and Gen. sg. is probably accidental. The former is imported from the pronominal declension, where the PIE ending for o-stems was *-oy, which regularly became -ei in early Latin and -ī in Classical Latin. The latter is of unknown origin, but is shared with Celtic. That these two are different suffixes is shown by the fact that (a) in early Latin inscriptions, only the Nom. pl. is spelled -ei, while the Gen. pl. is always spelled -i; and (b) even in Classical Latin, there is a difference between the forms in the case of words in -ius/-ium: the Gen. sg. may contract -iī to -ī (e.g. consilī for consiliī), but the Nom. pl. never does. That's for the second declension; the first declension simply took over these -ī endings from the second declension, giving e.g. animā-ī > animae in both Nom. pl. and Gen. sg. (although the older forms in -ās are still attested for both functions). (The fifth-declension Gen. sg. in -ī has the same analogical origin.) In the fourth declension, the identity of the Gen. sg. and Nom. pl. goes back to PIE (see fdb's answer); I don't know of any evidence bearing on the question of whether that identity is meaningful or accidental.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.