What on earth is ergative?

A couple of times now I have mentioned "ergative" or "ergativity." Some of you know what this is, while others may not. It has to do with how languages divide up the duty of representing three categories: subject of an intransive verb, subject of a transitive verb, object of a transitive verb.

Theoretically there are four possibities, but in actuality two are usually found in natural languages. The shorthand names for them are accusative and ergative. Here is how they work:

Accusative languages:

transitive subject
intransitive subject
transitive object

Ergative languages:

transitive subject
intransitive subject
transitive object

In each system the roles with the same color are marked the same way. The role marked with a different color is marked differently. If, like me, you are a native speaker of an accusative language, the ergative system looks quite bizarre. However, as you can see from the way I have listed them, it is just a matter of how they divide up grammatical space. Neither is inherently more logical than the other. Each gets the job done.

In each case, the system is named for the role that is marked differently. Since in English the direct object is the role distinguished, the system is give the name "accusative," after the traditional name of the case for direct objects in Latin, Greek, etc. The ergative role is the distinctly marked role in the ergative system. The term comes from a Greek root meaning "work" and was coined in the 1970's by linguists working on such languages.

Ergative languages are quite common. All Mayan languages are ergative, as are Basque, Eskimo, many Australian aboriginal languages, and even a few Indo-European languages such as Gujerati (spoken in India). In most ergative languages there is a "split," that is, in certain constructions ergative case marking is used, while in others accusative or non-distinguished case marking is used. Sometimes the trigger is a particular verb tense or aspect, other times it is subordination. The latter is the case in Tektiteko, the language about which I wrote my Masters thesis. In this language, ergative marking is the norm, but when a clause is subordinated (e.g., "when I went...", "so that I can..."), all three cases are marked the same way, with ergative prefixes. This may sound confusing, but in the context of a real life situation, there is rarely any ambiguity.