I try to understand why mit taking the dative case, while ohne taking the accusative - because it seems to me that they are both taking the same role in the sentence, with an opposite meaning (with/without).

Can someone explain it?

  • 4
    There is no (consistent) logic to case frames. If there were, you wouldn't have to memorize them word by word. – Kilian Foth Jan 7 at 14:59
  • Better get used to ;) e.g. an etwas (akk.) glauben but an etwas (dat.) zweifeln, although it's the same preposition (with "negated" verb) – c.p. Jan 23 at 8:26

Even the Old High German variations »anu«, »ano« and »ana« (used in 8th century) were used with accusative case. Some etymologists say, that in even earlier times it also was used together with genitive and dative case, but I couldn't find any sources for it.

The word »mit« on the other hand seems to be in use with dative case since it exists, which also is for much more than 1000 years.

And other than the English pair »with - without« the German words »mit« and »ohne« are not related to each other. So, I don't see any reason why they should be used with the same grammatical case.

The reason why we today use certain prepositions with certain grammatical cases is because our ancestors did it this way. And they did it, because their ancestors did it. These are just inherited usages, and the origins of this usages have been forgotten over many centuries.

Living languages do not follow any logical rules. They develop like species of living creatures: The actual generation inherits everything from the previous generation and adds tiny changes. Many of these changes disappear quickly, but some of them will be inherited to the next generation.

All living species inherited absurd and impractical features1, and this is true for living languages too.

1● The photosensitive layer in our eyes is on the outer back side of our eyes. It would make more sense to have it on the inner side like it is the case in octopuses eyes. Because the sensor layer is on the wrong side, we all have a blind spot in our field of vision.
● The testicles of male humans only work properly in a sack outside the body. We don't carry any other organ outside the body for good reasons. It would be much safer to have all organs inside the body, including our testicles. (Elephants have their testicles inside the body where they work properly, although their bodies are exactly as warm as human bodies.)
● The recurrent laryngeal nerve is a nerve that connects your larynx (which is in your throat) with the brain. But from the brain it goes down to the heart and goes round the aorta before it goes to the larynx. A direct connection would be much better. Even in a giraffes neck this nerve goes from the brain down the whole neck to the heart and then back again the whole neck, up to the top of the neck where its larynx is. This was even the case in the necks of all dinosaurs. To connect the larynx, which was just a few centimeters away from the brain the nerve had to have a length of up to 28 meters in some species. This is not really the best way to connect the larynx with the brain.

There are much, much more examples for nonsense produced by evolution. This all did happen, because evolution doesn't make any use of logic. (And no intelligent designer would have designed such crap.) Evolution works only with inheriting and tiny variation. And sometimes you get results that are impossible to explain with logic. And because also the development of languages is driven by evolution, all natural languages contain features that can not be explained with logic.

  • Zu dem hübschen Beispiel mit der Larynx, was fast relevant für Sprache allgemein ist, sei bemerkt, dass es schon sinnvoll ist, die Pumpe direkt mit dem Abgas-Zuluft-Ventil zu verbinden. Eine indirekte Verbindung über das Gehirn wäre eben indirekt. Welche positiven oder negativen Konsequenzen das so oder so hat, sei dahingestellt, geht aus deiner Antwort jedenfalls nicht hervor. Am ehesten könnte man der Antwort entnehmen, dass Systeme immer irgendwelche Kompromisse eingehen, die kaum perfekt sind. Das betrifft ironischer Weise die Theorie, mit der du die Frage eher ausweichend beantwortest. – vectory Jan 22 at 23:21
  • @vectory: Die Pumpe ist nicht mit dem Ventil verbunden. Es läuft lediglich ein Signalkabel, das den Hauptrechner mit dem Ventil verbindet um ein weit entferntes Rohr herum. Es macht auch keinen Sinn die Flüssigkeitspumpe irgendwie mit einem Luft-Ventil zu verbinden. Das System, das das Blut durch den Körper pumpt hat nichts mit dem Schlucken von Nahrung und der Erzeugung von Tönen beim Sprechen und Singen zu tun. – Hubert Schölnast Jan 23 at 9:05

There is indication that ohne used to be linked also to the dative (and genitive) case. In the etymology section for the entry ohne at the DWDS, Wolfgang Pfeifer writes:

In präpositionaler Verwendung verlangt ohne, das seit dem Ahd. sowohl ‘nicht versehen mit’ als auch ‘außer, ausgenommen’ sein kann, von Anfang an den Akkusativ, doch ist in älterer Zeit Verknüpfung mit dem Genitiv (bei fließendem Übergang zum Adverb) und mit dem Dativ ebenfalls möglich.

English translation: In prepositional use, ohne, which since the Ahd. can be 'not provided with' as well as 'except, barring', has required the accusative from the beginning, but in older times linking with the genitive (with a smooth transition to the adverb) and with the dative had also been possible.

Unfortunately, Pfeifer doesn't provide any examples. Why nowadays the accusative is exclusive, remains unclear. I am afraid that it just evolved that way, without logic playing in the background.


To shed some more light on this question, we can look at other European languages which have kept more cases than German, particularly Slavic ones like Russian.


  • ohne - takes an Accusative because it needs a direct object. This is fairly natural and expected, there is nothing strange about it.

Now let's consider why mit (with) might be different and not fall into this category.

In Russian & Slavic languages,

  • with [mit --> с, s] in Slavic languages uses a case called Instrumental which no longer exists in German. Its purpose is to describe an "instrument" with which an action is performed. So with is therefore used with Instrumental.

Interestingly, the Instrumental case is morphologically quite similar to Dative. In Russian, they both almost identically use the letter m for Masculine and ei for Feminine. In Russian, the Dative masculine adjective ending -emu (e.g. нашему, nashemu, "to our") is similar to the Instrumental adjective masculine ending -im" (e.g. нашим, nashim, "with our"). In German the letter m also plays a role in masculine Dative, and analogously er in feminine Dative.

So what I suspect happened over time is that in the case of ancient German, the similar morphology of Dative and Instrumental got conflated into a single Dative. They sound close to each other, and maybe German erased the difference between them. Dative has absorbed Instrumental due to very similar morphology.

  • ohne doesn't have any issues, it takes an Accusative which is understandable and natural (because it needs a direct object). The question was about mit which had an "unnatural" Dative. – gene b. Jan 19 at 15:04

Dativ comes from Latin (casus) dativus (means "given") which was used to denote the receiver. Akkusativ comes from Latin casus accusativus (means "case relating to the indictment") which was used to mark the direct object.

A speculation by me is "with" had been understood as "with the given one", and "without" had been understood as "without that case (of absent one)". But in the end our ancestors did not write down the reason where the grammar case comes from.


The answer might lie in the origin of the prepositions.

(1) Proto-Germanic *midi is an isolated development, probably from Proto-Indo-European *me-, with some suffixation, (*meth₂) whence it would be akin to middle (*me-dʰi-o-). "among, amidst" and "with" also shows a compatible semantic similarity. Since *me- does not conform to basic assumptions about sylable structure that prohibits roots beginning or ending on a vowel (obviously not without exception), I would take this with a pinch of salt, more so when *'kom-, *'km- "with" is suggested whereby *'k regularly yields German H except in unstressed sylables under certain conditions (Verner's law). Thus *(h)m- is imaginable in principle, while the dental is in need of an explanation either way.

Although Icelandic með ("with") from the same root also commands dative, en.Wiktionary helpfully points out "+ instrumental" for *midia case that was leveled variously (i.e. lost) in later stages. For comparison, the supposed cognate Ancient Greek meta is used in three different cases for various senses.

This type of morphology is not fully understood or at least cutting edge research (meaning something within the last dozent years, give or take). This is especially difficult with particles that don't exactly conjugate in agreement with case, but appear rather fossilized (compare in, im, but in's?).

What's likewise difficult, and rather important to understand, is inflection in the supletive paradigm of pronouns. In Berlin you might actually hear "ohne dir", although it arguably sounds rather like dia. In Old English, acc. and dat. are indiscriminately rendered þē (normalized spelling; on that note, while "dia" is a common rendition of the Berliner Mundart in writing, it is certainly not standard, and the sound is certainly distinct from Dia-Schau). Also compare earlier quite different forms for "she", that rather belong to her

A contrary assumption to entertain would require conjugation, that is seen in *midi "mid" anyway, which has instr. *middju, which should remind a little of you, Frisian ju, for example. Even *meth₂ reminds of *tuh₂ "you" (not to be confused with *yu "you", cp. Ihr, euch). It seems therefore possible that the inflection was baked-in and persisted through changes to the case system (implying * ich komme mitia was primary, would have regularly developed to ich komme mit "I come with", and was reanalyzed as mit dir, substandard dialects notwithstanding), which is relatively unlikely because it can hardly account for other pronomials, but since mit mir is reletively rare, mit sich, uns indistinct, mit es non-existent, while Ihr, Ihm, Euch are akin to you (also note dem, der, des), this is barely a problem. It is very likel though that *-th₂ has another explanation that I'm not aware of.

Even if this might be entirely anachronistic, it is formaly equivalent to analyzing PREP + PRON + SUF. If dative *midjem "(to the) middle" thus reminds reasonably of "mit Ihm", it has to be noted that third person has a predisposition to appear in different cases then second person or first, potentially acommodated by different grammatical casing.

(2) ohne has no root beyond Proto-Germanic *enu, or *ena. Ancient Greek ἄνευ "without" has been suggested to reconstruct *h₁(n̥)new, with *(n̥) comparable to un-, Gr. an-, from *ne (the negative particle), but the rest of the formation unexplained. *h1 (the e-coloring laryngeal) is needed to explain the vowel shift, but has no meaning. *h1e-, *h1en- is a whole paradigm, but this cannot be guessed at. *new- ("new") is barely compatible, and a sigmatic *-w- difficult to come by (cp. with).

Following the same line of reasoning as above, *w corresponds to we (*wiz) but not us (*n-sme, showing the same vocalization of the nasal, by comparison to Latin nos, nostra).

(3) If I mentioned Berliner dir, it should also be noted that its dative use in place of dich needs to be compared phonetically, for /X/ may tend to a rhotic. This would be the same difference as for och "auch" and Norse ok "and" (cognate so far, including OE eke "and") or English or (cognate to oder, but where did the middle leave of?). You see, the difference between "as well" and inclusive "or" is no more difficult to bridge then between "and".

(4) It is not apparent how either particle would have been used originally, if they even stemmed from the same strata.

At any rate, ohne has a very different PoS in many ways, and it is not an anthonym to mit in all cases. Ich komme ohne dich is not the negation of ich komme mit.


There is no logic in grammar as you can clearly see by the dative case in German while almost every similar sentence in English would require accusative.

Language is not Mathematics. It's art not science.

  • 1
    So languages cannot be internally consistent because they differ among each other? Also, English does not distinguish between accusative and dative, so what are you saying? – Carsten S Jan 7 at 16:24
  • @CarstenS Read the question again. – Olafant Jan 7 at 23:27
  • 1
    I wouldn't say there is no logic, for example at least mit does consistently use the dative, it's just that the logic is more complex than it needs to be. As a mathie I can state that even math has its notational inconsistencies. For example in the composition of functions, g ∘ f means first perform f then perform g. This type of thing is usually down to "historical reasons", meaning there was a choice made at some point that seemed to make sense at the time, but in hindsight turned out to be wrong, and it's more trouble to fix it than to just live with the consequences. – RDBury Jan 8 at 1:09
  • @RDBury Consistency and logic are two different things. And notational convention has nothing to do with logic. But your example makes sense. Some things just evolve and there is no logic behind it. – Olafant Jan 9 at 9:39

Your Answer

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

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