Morphosyntactic alignment

Morphosyntactic alignment refers to the rules used to distinguish between arguments of a transitive verb and those of an intransitive verb in a language. This distinction can be made by case marking, verb agreement or word order.

Marking core arguments

Transitive verbs have two core arguments, the agent (A) and the patient (P). For example, in the following sentence:

(1) The child throws the ball.

the child is the A and the ball is the P. Intransitive verbs have one core argument, the subject (S). In the sentence

(2) The child laughs.

the child is the S.

Most languages have only two kinds of marking for the three types of core arguments S, A and P, using either one marking for S and A and another for P, or one marking for S and P and another for A. The different morphosyntactic alignments are:

1. Accusative alignment (also nominative-accusative): In a language with accusative alignment, one marking, the nominative, is used for S and A ('subject') and another, the accusative, for P ('object'): S=A. Most European languages have accusative alignment. Many accusative languages can detransitive transitive verbs by raising the P to S and demoting or omitting the A. This is called the passive voice.

2. Ergative alignment (also ergative-absolutive): In a language with ergative alignment, one marking, the absolutive, is used for S and P and another, the ergative, for A: S=P. Examples include Basque and most Caucasian languages. Many ergative languages can detransitivize transitive verbs by raising the A to S and demoting or omitting the P. This is called the antipassive voice.

Very few languages are completely ergative in their alignment; typically there is some subset of the grammar that obeys accusative alignment. (Linguistic universal #3)

3. Active alignment: In a language with active alignment, the category of S is split, i.e. some intransitive subjects are marked like an A and others like a P. The criterion for this split is a semantic one, the agenthood of a subject. Consider the following three sentences:

(1) The child throws the ball.

(2) The child laughs.

(3) The ball lies in the sand.

The sentences (2) and (3) are intransitive, but in an active languages, their subjects receive different markings. In (2), the subject is an agent (S_a) and receives the same marking as the A in (1) (S_a=A). In (3), the subject is a patient (S_p) and receives the same marking as the P in (1) (S_p=P). Active languages are less common than accusative or ergative languages; examples are Georgian and Lakhota.

There are two subtypes of active alignment. In split-S languages, there are two distinct classes of intranstitive verbs, one marking their S as A, and one marking their S as P. In fluid-S languages, the two classes overlap to some degree, such that some intransitive verbs can use either marking depending on the semantics. Consider the following sentences:

(4) My brother arrived yesterday.

(5) Your letter arrived yesterday.

In (4), a fluid-S language would use A marking for my brother, as the brother is (most likely) an actively travelling person and thus an agent. In (5), a fluid-S language would use P marking for your letter, as the letter is an inanimate object that is being carried by some other (unmentioned) agent.

4. Hierarchical alignment (also known as direct-inverse alignment) occurs, for instance, in the Algonquian languages of North America. In languages with hierarchical alignment, arguments are ordered in an animacy hierarchy with 1st or 2nd person pronouns at the top and inanimate nouns at the bottom. A marking on the verb indicates whether the more animate NP is the A (direct voice) or the P (inverse voice). For instance, in the sentences

(6) The girl hit the ball.

(7) The ball hit the girl.

sentence (6) would display direct marking and sentence (7) inverse marking. If both NPs have the same rank on the animacy hierarchy, one of them is marked as obviative, giving it a lower animacy rank.

5. Austronesian alignment (also known as a trigger system) is a particular system of alignment found in many languages of the Austronesian family. In these languages, there is one unmarked case which can function as either the nominative (S=A) or the absolutive (S=P), and two further cases, the ergative and the accusative. The function of the unmarked case is marked by verbal morphology (agent trigger: accusative alignment; patient trigger: ergative alignment).

6. Tripartite alignment is rare; it uses three different markings for S, A and P.

7. Neutral alignment (also occasionally nicknamed clairvoyant among conlangers) means that no difference is made between S, A and P. This is restricted to certain subsystems of the language (case marking, verb agreement or word order) as all languages need a way to tell A and P apart.

This is not to be confused with the lack of overt case-marking; it is very typical for core arguments to be marked only by their position in the sentence (see word order).

8. An unusual and rare alignment is transitive alignment (in conlanger circles also called "monster raving loony") uses one marking on A and P, and another on S. Like neutral alignment, it is limited to some subsystems as the language needs to tell A and P apart.

9. Logical languages do not fit easily into any of the alignment types listed above and are perhaps best considered their own class of morphosyntactic alignment.

Dative and dechticaetiative languages

Another distinction is between dative languages and dechticaetiative languages. These languages differ in the treatment of indirect object. Consider the sentence:

(8) The boy gives the flower to the girl.

In a dative language, the direct object (the flower) receives the same marking as the P in a monotransitive sentence (e.g., sentence (1) above), while the indirect object (the girl) receives a different marking, the dative case. In a dechticaetiative language, the indirect object receives the same marking as the P in a monotransitive sentence, and the direct object receives a different marking. Dative languages are more common than dechticaetiative languages.

Split and mixed alignments

A language may show more than one morphosyntactic alignment type. In some languages, a split alignment occurs, depending on the sentence type. In many Indo-Aryan languages, for instance, the alignment is accusative in the present tense and ergative in the past tense. Georgian displays accusative alignment in the imperfective aspect and active alignment in the perfective aspect. Hittite inflects animate nouns accusatively but inanimate nouns ergatively. Generally, accusative alignment is more likely to show up in present tense, imperfective aspect and animate nouns, and ergative alignment conversely in past tense, perfective aspect and inanimate nouns.

Also, the various subsystems of a language (word order, case marking, verb conjugation) may show different alignments.

External links

• Wikipedia:Morphosyntactic alignment
• Alignment of Case Marking of Full Noun Phrases in the World Atlas of Language Structures