The forms ipse, ille, and iste look like vocatives instead of nominatives, and one would expect to see ipsus, illus, and istus instead. In fact, ipsus is attested but far rarer than ipse, but the other two forms seem to be entirely absent. Is there a reason for this unusual morphology? Is it related to the vocative (perhaps having a similar origin) or is the similarity coincidental? Is there something analogous in other languages or with other pronouns, or anything else that might shed light on these unexpected forms?

Some pronoun singular nominatives are quite wild (is, qui, illud, …), but the three I brought up are only mildly weird, and therefore I hope that an explanation actually exists. If I remember correctly, the singular masculine vocative of the second declension is obtained from the pure stem (without the nominative ending -s), with a vowel change o>e. Any evidence for tendency to drop the -s for pronouns in singular masculine nominative would be interesting if that is the case.

It appears that ipsus was more popular in pre-classical Latin, and kkm remarks in a comment that olle and ollus are attested for ille in old Latin. This gives the impression that the expected -us ending was there but was later lost. Any insight as to why or how this happened would be great.

  • 2
    Iirc ipsus is attested in older Latin. But I don't think *illus or *istus are.
    – Draconis
    Mar 7, 2018 at 4:26
  • @Draconis I checked PHI and indeed ipsus is attested, also classically, but illus and istus aren't. I updated the question.
    – Joonas Ilmavirta
    Mar 8, 2018 at 23:19
  • olle and ollus are attested in old Latin for ille, if I am not mistaken. In general, a connection to voc. looks rather doubtful. The se and te reflexes are very transparent; and IIRC there is a relation of "distance" of demonst. to the grammatical person in many PIE daughter languages (this = near me, that (far) = near him, that (near) = near you = is + te). Mar 12, 2018 at 5:12
  • @kkm I agree, they are unlikely to be directly connected to vocatives. But still the question remains: why not -us? The old attestations are interesting, as it makes it seem like the ending has changed.
    – Joonas Ilmavirta
    Mar 12, 2018 at 9:53
  • @kkm, Joonas, what if -e is derived from/has a common ancestor with the adverbial ending instead?
    – Rafael
    Mar 12, 2018 at 16:54

1 Answer 1


Because the second element is the PIE deictic pronoun *so-/*to-, which gave Latin -se;

so we may reconstruct PIE nom. sg. mas. as *só (Weiss, p. 337).

  • *is-so > *is-se > ipse (Weiss, pp. 346-347); cf. de Vaan *e(s)-pe-so > *ispse > ipse;
  • *is-to (or *es-to) > *is-te > iste (Weiss, p. 345);
  • *ol-so > olle > ille.

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.