One of the things which it is important for languages to be able to do is distinguish who performed an action, who was affected by that action, etc. There are a number of ways in which they do this. English largely just uses word order – Luke loves Lucy does not mean the same as Lucy loves Luke. A great number of other languages are very like English in this respect as well. Another common way of achieving the same goal is through case: different forms of a noun (or pronoun) which realise this sort of function. Latin is a well-known example of a language with case. In Latin, endings called nominative are used with “subjects” (prototypically, nouns which perform or are responsible for actions) and endings called accusative are used with “objects” (prototypically nouns in some way acted upon). For example, compare this sentence –

Sentence: Lucia canem amat
Word-by-word gloss: Lucy(nominative) dog(accusative) loves
Natural translation: “Lucy loves the dog”

– with this one:

Sentence: Luciam canis amat
Word-by-word gloss: Lucy(accusative) dog(nominative) loves
Natural translation: “the dog loves Lucy”

The different endings on the nouns convey their different roles in each instance. English does something similar with some pronouns (compare “I” in I love Lucy with “me in Lucy loves me).

Other languages also use case for a similar purpose, but do things a bit differently. To illustrate this it will be helpful to introduce the distinction between “intransitive” and “transitive” verbs. With intransitive verbs only a single noun (or pronoun) is associated with the action, giving sentences like I fallLuke diedshe went away etc. With transitive verbs there are two associated nouns: I like linguisticsLucy loves Luke etc.

In Latin (and English) the same case – the nominative – is used for the subject of intransitive verbs as the subject of transitive verbs. Thus “Lucy arrived” in Latin is Lucia advenit (not Luciam advenit with the accusative, or any other case), and we say I fall and not me fall. But in many languages the form used with the “subject” of intransitive verbs is the same as that of the object of transitives, with a separate form for the subject of transitives. This makes the traditional terms “subject” and “object” a bit confusing when talking about different languages and many linguists prefer labels something like the following instead:

  • S: the “subject” of intransitives (I fall);
  • A: the “agent”, or “subject” of transitives (Lucy loves dogs);
  • P: the “patient”, or “object” of transitives (Lucy loves dogs).

Nominative, then, is a case used for S and A but not P; accusative is used for P only. A case used for S and P but not A is called absolutive; a case used for A only is called ergative.


An example of a language with an ergative-absolutive system is Yup’ik, spoken in Alaska. The following is an intransitive sentence in Yup’ik:

Sentence: Dorisaq ayallruuq
Word-by-word gloss: Doris(absolutive) travelled
Natural translation: “Doris travelled”

And the following is a transitive one:

Sentence: Tomam Dorisaq cingallrua
Word-by-word gloss: Tom(ergative) Doris(absolutive) greeted
Natural translation: “Tom greeted Doris”

(Sentences from Payne 1997, p. 135.)

Note that the same form is used for “Doris” in both sentences, whereas “Tom” takes a different ending.

In another type of system – the one in which I’m currently most interested – there are two (or sometimes more) cases which can occur with S (the intransitive “subject”). Typically one of these is the same case as used with the transitive subject A and the other is that used with the transitive object P: these can be referred to as agentive and patientive cases respectively.


In one variety of Tibetan, the agentive is marked with a suffix –s, whereas the patientive doesn’t take any suffix. This is seen with A and P in the following transitive sentence:

Sentence: ŋas stag bsadpayin
Word-by-word gloss: I(agentive) tiger(patientive) killed
Natural translation: “I killed the tiger”

Compare this with intransitive sentences where S takes the agentive –

Sentence: ŋas ŋuspayin
Word-by-word gloss: I(agentive) cried
Natural translation: “I cried”

– and the patientive (note the absence of the –s suffix):

Sentence: ŋa śibyuŋ
Word-by-word gloss: I(patientive) died
Natural translation: “I died”

(Sentences from DeLancey 1984, pp. 132-3.)

This is as if we in English were to say I cried but me died.

The exact criteria which decide whether the agentive or patientive is used vary between languages: roughly speaking the agentive is generally used when S is more “in control” of the action and the patientive when it performs the action involuntarily. Part of my research is aimed at trying to understand and explain these patterns across and within languages in more detail. I also want to explore the ways in which agentive-patientive languages relate to languages with other types of case system at the more abstract, underlying level within the mind which theoretical linguistics aims to understand.


DeLancey, Scott. 1984. Transitivity and ergative case in Lhasa Tibetan. Proceedings of the Tenth Annual Meeting of the Berkeley Linguistics Society, pp. 131-40.

Payne, Thomas E. 1997. Describing Morphosyntax. Cambridge: Cambridge University Press.