Are there any examples of constructed languages that have verbs with different forms depending on the object of the verb?

  • Especially after reading Circeus's response, I wonder if your question's wording couldn't improved to specify more clearly what you mean by "change", etc.? It seems like a very broad & general question!
    – elemtilas
    Feb 14, 2018 at 0:31

4 Answers 4



I am not familiar with any “well-known” conlangs, but a quick search reveals that Klingon verbs inflect for both subject and object. This is a phenomenon called polypersonal agreement and it is very common in natural languages, and so it is only expected that it would also crop up in several conlangs. Among them, for example also my own, for which I just yesterday put up a puzzle that involves decoding the verbal system here.

I cannot currently think of any examples of a language which inflects for only the object, but it’s a rather low-hanging fruit and I cannot imagine that it has never been done before.


CALS lists 9 conlangs where the verb only agrees with P in transitives, 20 where it can agree with either A or P, and a further 127 where the verb agrees with both A and P, though of the ones listed, the only really well-known one is Klingon, which Adarian also mentioned. Including Mark Rosenfelder's langs (which aren't in CALS) in the at least relatively well-known bucket, Wede:i and Old Skourene both inflect the verb for both A and P to different extents (though note that Old Skourene works on an ergative basis).

Of the languages that mark P but not A, most of them are listed as having ergative agreement and so presumably agree with S as well, however two of them, 'Yemels and Snahhian are listed as accusative, and so presumably has the verb agree only with P.

Interestingly enough, despite conlangers often liking to solve problems by throwing affixes at them, a complete lack of personal agreement is actually overrepresented in CALS, when compared to WALS.

  • 1
    Note on how to read that last graphic: the lefthandside represents data from WALS (natural languages) the righthandside from CALS (conlangs). The darkened area represents the difference between the left and right. Thus a large dark green area represents overrepresentation of a feature in conlangs based on the data in those two databases (which are both not representative to any rigorous standard) Feb 11, 2018 at 20:47

Әřant has a change that, while it may not be what you're asking for, does fit the question in that depending on the animacy valency of the object, the verb conjugates differently:

men xshanni - - tassu han-ne!

it drop.2S.IMPER - - thou.2S INTERJ-hey.EMPH


san xshannos - - tassu han-ne!

her drop.2S.INDIC - - thou.2S INTERJ-hey.EMPH

The person has a higher animacy valency, so requires a (slightly) more polite idiom.


I believe your question conflates two related, but crucially distinct concepts, patients (a thematic relation and semantic concept) and objects (a verbal argument and syntactic concept).

By conventional syntactic definition, if only one argument is being agreed with, that argument is the subject. In that sense, it's by definition impossible for a verb to agree only with an object.

However, the patient of a verb does not have to be its object! In ergative languages, it's generally acceptable to analyze that the syntactic subject of a transitive verb is its patient.

In summary: it makes no sense to say that a verb agrees only with its object, because it would causes that object, by definition, to be the subject, merely reversing the agency alignment of the verb (cf. English I like X vs. Spanish me gusta X).

  • 2
    I would like to see some citations here, you’re making some strong claims that do not agree with the definitions I am aware of, which define the subject as being either the S or A role of a clause. Dixon’s Ergativity has a whole chapter about how that notion of subjects still applies to ergative languages. Feb 14, 2018 at 15:13
  • I have made no assertion whatsoever that "subjects [do not] apply to ergative languages". You are making the exact conflation I was pointing out in my answer. I have asserted that it's arguable that in ergative languages patients can be syntactic subjects of transitive verbs. What that means, simply put, is that in ergative languages, the case that marks subjects is often the absolutive, causing patients to be treated as subjects for syntactic purposes: who left in "he hit her and went away" is not going to be the same in Dyirbal as in English.
    – Circeus
    Feb 14, 2018 at 15:45
  • 2
    @Circeus Have you actually read Dixon? No such conflation is made, as Dixon (and a lot of other authors) use the term "subject" specifically to refer to an S/A grouping which have arguable universal linkages, and the term "pivot" to refer to categories controlling syntax in specific languages (which may then either be S/A, S/O, a mixture of those depending on context, or a relation not defineable in terms of syntactic roles (e.g. topicality)(though note that some authors, including Dixon, prefer not to call the latter a "pivot", and refer to such languages as "pivotless")).
    – Gufferdk
    Feb 14, 2018 at 16:36

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.