Generally, a relative pronoun agrees with its antecedent in gender and number, while its case is determined by its grammatical function in the relative clause, e.g.

Do pecuniam filio [dat.] quem [acc.] amo.

Amo filium [acc.] cui [dat.] pecuniam do.

Equus filii [gen.] qui [nom.] me amat fortis est.

See Allen & Greenough §305:

A relative agrees with its antecedent in gender and number; but, its case depends on its construction in the clause in which it stands.

Allen & Greenough notes an exception, though, in §306.a:

a. A relative occasionally agrees with its antecedent in case (by attraction).

sī aliquid agās eōrum quōrum cōnsuēstī (Fam. 5.14)
if you should do something of what you are used to do
[For eōrum quae]

This so-called attractio relativi is a common phenomenon in Greek. (See Smyth §2522ff) Curiously, though, Smyth notes that this often occurs from the accusative into the genitive or dative, while a relative pronoun is "very rarely" attracted from the nominative or dative. Smyth gives a lot of other interesting notes about the phenomenon of attraction in the linked section.

My question, however, has to do with relative attraction in Latin. Allen & Greenough says this happens "occasionally," but doesn't specify how often or in what instances:

  • Is this a relatively rare phenomenon?
  • Are there situations (like in Greek) where it is more common or even preferred?
  • Are there situations where it is impossible?

(My question is inspired by an answer to another question which suggests "nemo dominus cui mortuus est" rather than "nemo dominus ei qui mortuus est.")

  • 2
    I would like to know this too. It is also interesting how languages like English and Dutch suffer the same phaenomenon. Could there be something universal about it?
    – Cerberus
    Commented Apr 26 at 4:21


Your Answer

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