I'm making a conlang and right now it is SVO. I want to make it more in line with an SOV language. Currently Nominatives are unmarked while only 1 Accusatives is marked in a group. So it reads "Bob kicked Susan and Dan-ke" rather than "Bob kicked Susan-ke and Dan-ke". If I shift this to SOV it would be "Bob Susan and Dan-ke Kicked" which to me looks a little confusing... Another thing is that groups are not seperated by commas. They are all connected by "and" so if we add another name it would be "Bob Greg and Susan and Dan-ke kicked".

Looking at it here I can see that it's fairly obvious where the Object is, but it also seems like this would be confusing, especially as spoken. So do SOV languages tend to employ a Nominative marker or an Subject marker so that it'd read something like "Bob blah Greg and Susan and Dan-ke kicked"? Or is it just me overthinking this?

  • It depends on the language; some SOV languages mark the nominative (Japanese, Korean), some don't (early Latin), and some don't even have a nominative-accusative system (Basque, Nepalese)... I think your ke marker is fine; it's called an enclitic, though we usually use = instead of - in front of clitics, a convention you might want to adopt. Dec 30, 2016 at 17:55
  • In Turkish, which is an SOV language, the subject would be marked off with a comma: "Bob, Susan and Dan kicked"
    – BillJ
    Dec 30, 2016 at 17:58
  • @WavesWashSands forgive me if you knew this already, but the "-" is just there to make it clearer what I'm pointing out. I would normally write it as "Danke" but considering that is a name already, that would make it confusing. ^.^
    – Durakken
    Dec 30, 2016 at 18:01
  • So some have a marking, but not all... do you know if there has been a survey that says if they more often do or do not? Or in your general experience do you think/feel they do or do not?
    – Durakken
    Dec 30, 2016 at 18:04
  • @Durakken Yep, but in linguistic discussions where we need to segment morphemes, we generally limit - to affixes and use = for clitics. I'm not aware of such a survey and I can't find one on Google Scholar yet, though if there is one, I'd be somewhat sceptical of whether there's a causal link between overt nominative marking and SOV unless some functional explanation is given for the tendency. Zero-marking for nominatives is an effect of economy - Nominative arguments are the most common, and can thus be made covert - and I'm not sure if SOV alone is related to that. Dec 30, 2016 at 18:12

2 Answers 2


Zillions of possibilities here. If I understand correctly, you are mainly concerned about the possibility that there is no obvious boundary between the subject and object in case there is a multiple of one or the other. But I do not see that as a problem at all:

1) Minimal marking by word-order

If the language requires subject, then it is fairly clear the first word is a subject. Then you have a transitive verb that requires an object, so know that the stuff before the verb is object. Bob Susan, John and Dan kicked is perfectly interpretable because the verb requires object, thus Dan is part of the object, then its pair noun John too. Now you know Bob is subject and you consider only whether Susan is subject or object. It is clearly object because if rules of pairing with and are similar to English, the and is always before the last element of the chain.

Considering that human mind does not favour long chains of elements on the same level (John, Peter, Susan, George, Nathalia,.... went to school.), it is highly unlikely that even this minimal marking by the word order itself would pose significant problems.

2) Prosodic marking

Syntactic units typically share also their own intonation/pitch contour. Thus it is likely that even with very minimal marking, the phrasing would be Bob | Susan John and Dan | kicked thus clearly separating the subject from object.

3) Syntactic restrictions

The position between subject and object may work like the weak position in many Slavic languages and attract all the clitics, thus separating the subject from the object. Connected with some obligatory particles (say Japanese-like topicality markers), this could pretty much work in every single sentence.

4) Overt morphological marking

Case endings, clitics, whatever your heart may please. Subject can have even zero-morpheme and it is the object that bears the marking (this is actually fairly typical).

5) Covert morphological marking

Congruence between the verb and either the subject or the object. I.e. in Bob Susan and John kicked, the verb would be conjugated for 3SG in case of congruence with the subject, thus clearly marking it is just the very first noun, or 3PL in case of congruence with the object.

6) Combo

Languages tend to be fairly redundant in conveying information in order to make sure it goes through even in case of noise in the channel, thus more/all of the above possibilities are likely to go together.


Of course, in conlang you can do whatever you want. So you can have parsimonous inflection markers like in your question (maybe, they are just postpositions).

Natural languages with inflection work differently from your conlang: They tend to apply the inflection markers abundantly and use agreement and repetion to create cohesion and redundancy.

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.