5

The preposition expressions like "on top of the table", "under the tree", "above the building" are very well understood. Comparing these with the Germany noun cases "auf dem Tisch", "unter dem Baum", "die Farbe der Kleidung", "Ich sehe den Wagen". Also with Finnish noun case "Poydaallaa".

I really feel that essentially there are no difference between preposition/postposition with the noun case. I mean (roughly) if the preposition is separated from its noun then we will call it as preposition, but if the preposition is 'attached' to the noun then we will call it 'noun case'. So, what is the biggest difference between preposition / postposition and noun case?

  • 2
    A thought experiment on why case markers and adpositions must be different: many languages require that certain adpositions take specific cases to denote specific meanings. E.g. Polish Jestem z Polski 'I'm from Poland' vs. Jestem z Polską 'I'm with Poland*' - same adposition z, different cases (genitive vs. instrumental). – Mark Beadles Jul 1 '19 at 15:42
  • @MarkBeadles Is that really different from a preposition selecting prepositional phrases as their object (as in French au dessus or en dessous whose objects have to be marked with de)? – Eau qui dort Jul 2 '19 at 23:23
  • @MarkBeadles I don't think that argument works. After all, I could argue that these are four distinct case markings: genitive -i, instrumental -ą, ablative z- -i, and comitative z- -ą – Tristan Jun 24 at 13:22
  • the fact that adjectives can come between the z- and the stem but not between the stem and -i or -ą is still enough to make this analysis untenable (unless we're willing to accept all adjectives as affixes), but the fact z can pair with multiple cases doesn't help us with that – Tristan Jun 24 at 13:25
10

Semantically (in terms of meaning)? There's no real difference. Some languages might use an adposition for a certain meaning, while other languages use noun case. The underlying meaning can be exactly the same.

Syntactically, though—in terms of putting words together into sentences—there's one major difference. The case is part of a noun, and can't be separated from it. Adpositions can.

Consider in English:

I gave the book to Bob.
I gave the book to the blond man.
Where's the person I gave the book to?

The preposition and the associated noun can have all sorts of things inserted in between them, and one can move without the other following.

Compare to Latin, which uses a noun case for this:

Dedī librum Robertō.
Dedī librum flaviō virō.
Ubi est homō cuī dedī librum?

(The word order in these is somewhat stilted and unnatural, in order to line up with the English, but the point stands.)

The dative case marking can never be separated from the noun, by any means: it's a part of the word, and no syntactic movement will ever pull it off. The case is also applied to all words in the phrase: "blond" is flaviō instead of flavius, because it's taken on the case marking too.

| improve this answer | |
  • 3
    Case will also in some languages be marked on all words in a phrase, whereas adpositions only occur once. (AFAIK??) – curiousdannii Jul 1 '19 at 11:14
  • @curiousdannii Also true; I'll add that – Draconis Jul 1 '19 at 16:11
  • I guess that was the main thing that prompted me to write an answer too (although I see now that from the way I phrased it, it sounds like "adpositions never changing" was the thing that I wasn't seeing mentioned in other answers). – LjL Jul 1 '19 at 16:29
  • @curiousdannii: adpositions in Hungarian often occur twice. For example, az a ház (that the house) = "that house"; az előtt a ház előtt (that in-front-of the house in-front-of) = "in front of that house". – TonyK Jul 12 '19 at 0:26
5

Maybe this isn't a universal distinction, since other answers do not mention it, and I apologize if it's too Indo-European centric, but I understand a major difference between cases and adpositions (even though postpositions can look a lot like cases) is that the forms of adpositions can be predicted from the phrase they are attached to; they behave as words, or maybe especially with postpositions in cases such as Japanese, clitics.

On the other hand, different cases change words in ways that are not uniform nor necessarily predictable: for example, the Latin noun rosa in the nominative case becomes rosae in the genitive case, while the noun lupus in the nominative becomes lupī in the genitive. You cannot single out a "genitive postposition", because it would be ae in one case and ī in the other. At this point, one might be tempted to say, okay, they are different, but I can predict that a word in -us will get in the genitive; but this is not true, because currus becomes currūs in the genitive, not *curri.

Another major difference is that adpositions only need to occur once in the phrase they are attached to: before it in the case of preposition, and after it in the case of postpositions; furthermore, they occur before or after the whole phrase, including function words like articles. Cases modify individual words, and generally, a noun phrase that needs to be marked by a given case will exhibit agreement, whereby if the noun in the noun phrase is in the genitive case, any adjectives or other words that are subject to declension (articles in German or Greek are also declinable, for instance) will need to replicate the same case as the noun. This means that whereas "the good wolf" will be

  • bonus lupus in Latin
  • der guter Wolf in German
  • ὁ ἀγαθός λύκος in Greek
  • hyvä susi in Finnish

"of the good wolf" will become

  • bonī lupī in Latin
  • des guten Wolfes in German
  • τοῦ ἀγαθοῦ λύκου in Greek
  • hyvän suden in Finnish

where every declinable word (every word, in these examples) is modified by case.

| improve this answer | |
  • 4
    "adpositions never changing" isn't really true, though: Insular Celtic languages have conjugated prepositions; Slavic languages like Polish and Russian have phonological alternate forms for prepositions including w (we), z (ze), в (во); and Spanish has the forms conmigo, contigo, consigo. – Mark Beadles Jul 2 '19 at 16:34
  • @MarkBeadles you are right. I feel there is still a bit of "unchangeability" involved in adpositions compared to cases (alternate forms you mention are "just" for euphony, and those Spanish words are rather fossilized), but I can't pin it down, so I just changed my claim in the answer above to just one of predictability. It's a weaker answer that way, but after all, my "on the other hand" was already focused on the predictability part. Does it look better to you now? – LjL Jul 5 '19 at 23:23
  • 1
    Correction: "des gut-en Wolfes", and some inflections are oblique anyhow, i.e. female "der Frau". – vectory Jul 14 '19 at 18:22
  • @vectory thanks, corrected. I'm not entirely sure what you mean by oblique here (I know the term oblique cases, but if that's what you mean, I don't understand how it applies)... but if you just mean that some/many nouns stay the same as in their nom./acc., I didn't mean to imply otherwise: I just happened to think of an example where all the words change in all the languages I gave, which seemed neat, but I wasn't trying to make a point that they have to change... – LjL Jul 14 '19 at 23:46
3

This was addressed in Kuryłowicz's paper "Le problème du classement des cas" written in 1949 (!) and what he wrote is still true. As mentioned in the other answers, prepositions differ from case in form but not in function. There's a continuum case affix-clitic-preposition to signal grammatical relations. Note that even within one language there may be a formal distinction, e.g. in Lithuanian į mišką and miškan both mean "into the forest" — the PP is of type į+ACC while the NP is in the illative case. This is also the main reason why in the Universal Dependencies annotation scheme prepositions are marked with the label "case" on the corresponding edge and there's also a conflated version where the preposition is built into the edge label in order to make the syntactic representation more uniform (and therefore comparable) across languages.

| improve this answer | |
  • In many languages, the distinction between adposition and case marker is robust, e.g. most European languages. The Universal Dependencies annotation scheme gets this all wrong by conflating the analyses of the two. In essence, the UD scheme treats adpositions like case markers, ignoring the syntactic differences that distinguish between the two discussed and illustrated in the answers above. – Tim Osborne Jun 24 at 12:16
  • @TimOsborne Well, read the paper first, it's instructive. The UD is inspired by LFG, as they explicitly admit, and works very well for most languages. You need to stop slandering the UD project, you're the guy driving the wrong way up a motorway and complaining that everybody's on the wrong side. – Atamiri Jun 24 at 12:30
  • You might do well to address the linguistic issues I raise. – Tim Osborne Jun 24 at 12:31
  • @TimOsborne What exactly? The answers in this thread (including mine) point out the basic difference at the level of morphosyntax. The UD scheme (UD 2, which is actually the third version as it was preceded by two alternative schemes) has prepositions as standalone nodes in the trees so they do capture the difference, there's no controversy. It's actually more interesting to look at languages like Hindi/Urdu were the distinction is less clear cut. – Atamiri Jun 24 at 12:38
  • There's a host of linguistic issues indicating that prepositions are heads over their nouns. These issues are ignored by the UD scheme. I have argued this matter extensively in published works, and I have invited the UD people to discuss the issues in prominent linguistic forums. They consistently duck, dodge, and ignore, refusing to step up to the debate table. Let's take a concrete example, e.g. preposition stranding: Who did you give it to? If the preposition is a dependent of who, there's a nasty projectivity violation reaching from left to right across the entire tree. – Tim Osborne Jun 24 at 12:54
2

The biggest difference is that case is a grammatical category whereas adpositions are a syntactic category, i.e. cases are properties across (certain kinds of) words, and adpositions are one type of words. The case of a noun/pronoun/noun phrase/determiner phrase (hereafter just "noun") is a property of that noun; different languages mark cases differently, some languages don't even mark case. Adpositions are a category of words that typically express temporal/spatial relations.

For example, in the English sentence I won the argument with her, the pronoun I is in the nominative/subjective case, and the pronoun her is in the oblique case. In English, the rule is that subjects have nominative case and complements of prepositions have oblique case. The noun phrase/determiner phrase the argument could have accusative/objective case or not, depending on your view of English syntax. You may say that all English nouns have case, but case is only realized in pronouns, or you may say that in English only pronouns have case, not nouns.

While in English, case marking is part of the noun but adpositions are separate from their complements, in some languages it is difficult to tell case marking and adpositions apart. For example, here are some Japanese sentences.

1.  Taro-ga  pan-o     tabeta.
    Taro-NOM bread-ACC ate
    'Taro ate the bread'.

2.  Taro-kara Hanako-e  hanataba-ga  okurareta.
    Taro-FROM Hanako-TO bouquets-NOM were.sent
    'Bouquets were sent by Taro to Hanako.'

In Japanese, case is marked by the so-called particles (ga and o in (1-2)), which have the same distribution as the adpositions (kara and e in (2)). Some people consider both of them as case markers, and distinguish the two as syntactic vs. semantic case markers (Kishimoto 2018, 448). However, adpositional phrases may be case-marked in Japanese:

3.  Taro-kara-no  Hanako-e-no   tegami
    Taro-FROM-GEN Hanako-TO-GEN letters
    'letters from Taro to Hanako'

where the particle no marks the genitive case on the adpositional phrases taro-kara 'from Taro' and hanako-e 'to Hanako', making it difficult to distinguish between case-marking particles and adpositional particles on distributional grounds.


Kishimoto, Hideki. "Case Marking." Handbook of Japanese Syntax. De Gruyter Mouton, 2018.

| improve this answer | |
  • Both belong to the grammatical category. – amegnunsen Jul 1 '19 at 11:08
  • @amegnunsen No. See this and this. – Press Coffee Jul 12 '19 at 6:22
  • Syntactic category = Part of Speech. Every morpheme that cannot stand alone is considered a grammatical morpheme that is opposed to lexical morpheme. – amegnunsen Jul 12 '19 at 11:54
  • @amegnunsen I see where our opinions differ. I consider the notion of "case" to be different from specific mechanisms of its marking across languages. I would agree that case markers can constitute one or more syntactic categories, although I don't see any immediate value in that claim; we probably only care about the distribution of case markers in any given language. Still, I don't think cases themselves constitute syntactic categories - they are a kind of property, or feature, if you prefer. – Press Coffee Jul 13 '19 at 16:17
-2

It is a matter of meaning, and especially a question of syntactic functions. In some languages what marks an argument derives from adpositions like in English. So there are related diachronically, but not synchronically, even though they are pronounced identically.

Case markers indicate which is the object, the indirect object and/or the subject of a predicate. But adpositions indicate where a referent is situated in relation to another referent.

| improve this answer | |
  • 1
    In Persian, the postposition is used to mark the direct object. Other languages also have adpositions to mark the direct object. Even English uses the preposition to to mark the indirect object. – Yellow Sky Jul 1 '19 at 8:33
  • @Yellow Sky Yes, it is what I said. – amegnunsen Jul 1 '19 at 10:17
  • 1
    I'm downvoting this because I don't really believe there is such a semantic difference between adpositions and cases. It may arise in a few languages: for instance, in Latin, prepositions are used a fair amount, while base cases mark things like "the object, the indirect object or the subject", as you say. But that's not a general thing, it's specific to certain languages: in Japanesee, for instance, adpositions mark everything: topic, subject, object, as well as what I understand you mean as "where a referent is situated..."; in contrast, in Finnish, cases mark almost everything. – LjL Jul 1 '19 at 16:34
  • @LjL You are confusing two things: diachrony and synchrony & form/signifier and meaning/signified. To.IO is a grammaticalisation from to.PREP. Grammaticalisation can occur in different ways as in Japan. They have the same surface form but they have different underlying surfaces. – amegnunsen Jul 1 '19 at 18:02
-2

The best answer is: Adposition yields a more specific meaning/semantic, while 'case' yields a broader/deeper meanings/semantics.

| improve this answer | |
  • What languages is this true for? Many languages have many very specific cases. – curiousdannii Jul 6 '19 at 5:43

Your Answer

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