When talking about the locative case, Latin grammars generally say that its usage was mostly taken over by the ablative case in Latin. For example: Allen and Greenough say:

Relations of Place are expressed as follows:—

  1. The place from which, by the Ablative with ab , dē , ex .
  2. The place to which (or end of motion), by the Accusative with ad or in.
  3. The place where, by the Ablative with in (Locative Ablative). (§ 426.)

The grammars explain the extraordinary use of what looks like the genitive case for first and second declension nouns used without a preposition (e.g., Rōmae for "in/at Rome" and Brundisiī for "in/at Brundisium") as resulting from mergers with ancient locative forms) but then comfortingly tell you that some form of what we call the ablative case shines through in all other locative expressions (e.g., Carthāgine or Carthāginī for "in/at Carthage," Athēnīs for "in/at Athens," and in Itāliā for "in Italy). This seems perfectly logical and reasonable, especially as applies to words used in isolation or with the preposition in.

The grammars extend this use of the ablative by explaining that it is used for expressions of time that might have originally taken the locative, e.g. (nocte for "at night" or Īdibus Martiīs for "on the Ides of March"). I am further comforted by a vague memory that Sanskrit had locative absolutes that are the equivalent of ablative absolutes in Latin. Since poor Greek no longer had an ablative case, some functions went to the genitive, explaining its use of genitive absolutes for more or less the same purposes of the Latin ablative absolute.

We even have expressions like ā parte dexterā ("on the right (side)") that reinforce the idea that one use of the ablative was to express locative relationships.

Then we get to almost all other locative expressions, such as those introduced by the prepositions "ad," "prope," "apud," "post," "ante," "super," "sub," "suprā," and "infrā," and there is hardly an ablative usage in sight. I know that "super' and "sub" have some use of the ablative, but the overwhelming use of the accusative with the entire list of prepositions seems to demand some justification. I find it frankly embarrassing for poor Latin that many expressions in Romance languages express place with some form of ad plus a noun that is a descendant of a form with the accusative case ("à Paris," for "in Paris" or "a Roma" for in Rome). I think some forms of Spanish recognize the "lack of logic" :-) in this usage (according to my bias) and try to reserve en (a descendant of Latin in) for locative use and a (a descendant of Latin ad) for expressions of motion or where the emphasis is on where two things seem to come into contact by metaphorical motion.

Latin even has expressions using the accusative that seem to duplicate other expressions rightly used with the ablative, such as quam ob rem ~ quā rē and sub noctem ~ nocte. I know that language is flexible and there are subtle influences that could explain these particular similarities, but the overall picture seems quite surprising in a language that is supposed to keep the meanings of the case forms separate.

In ancient Greek, there seems to have been extensive use of the dative with prepositions as a descendant of locative forms. I don't recall any use of the accusative off hand, except to express explicit motion or to express extension in space with a preposition like ἀνὰ ("throughout," "along," etc.). The main competition in Greek usage of locative expressions seems to have been from what appears to be a partitive use of the genitive with many prepositions (e.g., ἐπὶ τῆς γῆς).

In standard modern German, there is a very consistent contrast between prepositions used with the dative to express location and the same prepositions used with the accusative to express motion. I think this was true of Old Irish as well, but my memory could be faulty. This contrast between place where and place to which also existed in Latin with the preposition in, but why did it not extend this usage to all locative expression generally. (I know that why questions are often futile when talking about linguistic usage, but I am trying to understand if there is something I'm missing in the underlying semantics,)

In Russian, you have a still viable accusative case used with nouns, but a completely different prepositional case used with most locative expressions.

The only explanation I can come up with for the Latin usage seems to be that it represents an instance of the accusative of extent used to apply to any locative expression thought of as describing an extended region of space. If ancient Latin speakers felt an opposition between localized space and extended space, that might also explain the use of the ablative of location with verbs of placement, in contrast with the practice in German. For example, in Allen and Greenough:

  1. Verbs of placing, though implying motion, take the construction of the place where:— Such are pōnō , locō , collocō , statuō , cōnstituō , etc.:—

“quī in sēde ac domō collocāvit ” (Par. 25) , who put [one] into his place and home. “statuitur eques Rōmānus in Aprōnī convīviō ” (Verr. 3.62) , a Roman knight is brought into a banquet of Apronius. “īnsula Dēlos in Aegaeō marī posita ” (Manil. 55) , the island of Delos, situated in the Ægean Sea. sī in ūnō Pompêiō omnia pōnerētis (id. 59), if you made everything depend on Pompey alone.

Is there another explanation? Do any other languages take this approach?

  • 2
    Perhaps the semantic value of ponere is more “to cause x to be located in place y” rather than “move x into place y.” Look at insula … in mari posita. In that usage, there’s no notion of movement at all, only location. Feb 4 at 19:03
  • 1
    As for your comment on an alleged "lack of logic" ("I find it frankly embarrassing (...) that many expressions in Romance languages express place with some form of ad plus a noun that is a descendant of a form with the accusative case (...)"), some scholars like Antonio Fábregas (septentrio.uit.no/index.php/nordlyd/article/view/110/103 ) have argued that, despite appearances, the Sp. preposition "a" in directional contexts like Sp. correr a la casa (e.g. see ex. (52) on page 189) in the link above) encodes location (place) but not directionality (path). So Sp. dir. "a" is not "to".
    – Mitomino
    Feb 4 at 19:19
  • 1
    Cf. also the ambiguity shown in John ran in the room (locative AND directional), compared to that in John danced in the room (only locative). Run-verbs are directional manner verbs, compared to dance-verbs, which are non-directional manner ones. As for related constructions like the ones discussed by you & Kingshorsey (e.g. quī in sēde ac domō collocāvit), I’d recommend you the following work by Gehrke (2007) academia.edu/1224158/Putting_path_in_place.
    – Mitomino
    Feb 4 at 19:20
  • "Some scholars like Antonio Fábregas have argued that, despite appearances, the Sp. preposition "a" in directional contexts like Sp. correr a la casa...encodes location (place) but not directionality (path). So Sp. dir. 'a' is not 'to'." That is an interesting idea. There is also the issue that in some dialects (as in the one I try to mimic), I must say "entrar a la casa" for "enter the house" and "quedar en la casa" for "staying in the house, using "a" or "en" according to whether the accompanying verb is a verb of movement or not. I learned not to say "entrar en la casa." Feb 4 at 20:03
  • Nota bene: sub noctem does not mean nocte (during the night), it means "around nightfall." Feb 4 at 20:43


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.