User:Elector Dark/Morphosyntactic Alignment

This text on basis of morphosyntactic alignment requires you have some knowledge of certain terms used in linguistics. If confused, consult a glossary such as SIL Glossary of Linguistic Terms. The author of this text is ~)  The Elector, Darkness Immaculate . If you find an error in tis text, report it on the talk page and provide explicit citations.

In linguistics, the general meaning of morphosyntactic alignment is that of the relations between the core arguments of intransitive and monotransitive sentences.

The three categories distinguished are the  S ubject, the  P atient and the  A gent. Traditionally divided into five basic combination of cases, there are also several less clear-cut alignments of these arguments.

The distinctions can be made morphologically, syntactically, or, where case-marking is relatively young or at the end of its life, both ways. There are no instances of a language aligning to only one of the alignments, and oftentimes other alignments are mixed in; we can draw conclusions and tendencies nonetheless.

Basic Division
The basic divisions of morphosyntactic alignments are the following:

Direct
The direct morphosyntactic alignment is the most common MSA (acronym) in the world, taking around half of the world's languages. In essence, the direct alignment makes no distinction between any of the three beforementioned categores, neither syntactically nor morphologically.

While many of the world's languages are categorised as direct-marking based on a lack of an overt morphological system on the nouns that shows relations, many are, in fact, not actually direct - the marking is done by the verb, or by syntactical features, such as word order. The most common secondary alignment in these languages is nominative-accusative by far, then very marginally ergative.

There are a few rare languages in the world that are truly direct-marking. A good example is Dakota, with extensive documentation on it. In Dakota, which doesn't have fixed word order, verbs take up most of the marking. Due to its, in essence, nonconfigurationality, Dakota's syntax permits various word order permutations, leading to ambiguous situations where there is no marking, neither on the verb nor nouns, to show actual relations. Lakota, its close relative, also has such an MSA.

The direct morphosyntactical alignment behaves as such:

I am not aware of any valency-decreasing voices that occur in direct-aligned languages.

Examples of a pure and impure direct language are Lakota and English, respectively.


 * Lakota: Wičhaša kiŋ wiŋyaŋ kiŋ waŋyaŋke (the man the woman see)
 * English: The man sees the woman / The woman sees the man

Nominative-Accusative
The nominative-accusative morphosyntactic alignment is the second most common MSA, after direct. Queerly, its speakers cound for a disproprtionate amount of the world's populace. In essence, the nominative-accusative alignment distinguishes between the Agent and Subject on one side, and the Patient on the other.

Nominative-accusative languages are very well known and well-studied. They are amongst the "most purely aligned" languages, so to speak, but are nonetheless oft corrupted by other alignments syntactically. A good example is English, with ergative verbs sprinkled around. These verbs cannot be traditionally passivised. For example:
 * I broke the window.
 * *The window broke by me.

No matter this exception, English is overwhelmingly nominative over ergative, in word order and pronominal marking. While Enlgish is moving towards a direct alignment in quite a haste, there are some languages that, in addition to distinguishing the nominative and accusative, also make distinctions between plenty other cases. Such an example is Modern Sanskrit, the most complex Indo-European language still spoken (but not the most complex attested, which is Vedic Sanskrit), having eight cases, three numbers and as many genders.

Continental nominative-accusative Indo-European languages appear to be losing their nominativity in modern times: west of German, the languages are becoming direct-aligning, as well as northwards. German is dialectically split between the northern, less marking languages, and southern, more marking. The Baltic, Slavic and Balkan languages are in between the two extremes, preserving nominativity, while as one goes more and more south-east, more alignments pop up, primarily tripartite, but even an occasional transitive.

The nominative-accusative alignment behaves as such:

The Subject and Agent are marked in the nominative, and the Patient is marked in the accusative.

A very common feature of nominative languages is the passive voice. The passive voice is a special valency-adjusting voice, demoting the verb from transitive to intransitive, and promoting the patient to a subject. The former agent may thus be added as an indirect argument of the verb:


 * John kills the bear > The bear is killed (by John)

Not all nominative languages have the passive - the Slavic languages have in its place a mediopassive voice, and the passive usually requires a complex and queer-sounding construction to work.

Marked Nominative
An unusual subset of the nominative-accusative languages are the nominative-absolutive languages, or, as more commonly called, marked-nominative languages.

They are primarily found in Africa, in some Berber dialects and in Nilo-Saharan languages of Sudan. An example from Middle Atlas Berber:


 * Ičča uryaz aksum
 * The man has eaten the meat

In the example sentence, "uryaz" is recognised as the subject, and "aksum" the object. The word "uryaz" is actually inflected by the prefix "u-", to mark it for the nominative.

There are various explanations as to why this marking has occured, but the one with more credibility than plenty of the rest is that it is an areal feature, an influence of a language that was previously ergative-absolutive.

Ergative-Absolutive
The ergative-absolutive morphosyntactic alignment is the second most common two-case alignment in the world. The distribution of ergative-absolutive MSA is pretty regional (although lone examples can easily be found), and the highest concentration of ergative-absolutive aligning languges is in the Caucasus (where the members of the three Caucasian families all have a very high tendency to be ergative) and Tibet. They are, of course, found in other places, such as, for example, on the western border of Spain and France (Basque).

Although all languages counted as ergative-absolutive have an ergative morphology, many have a nominative-accusative syntax; there are no documented or attested languages in our history to have had ergative syntax but a non-ergative morphology.

Good examples of very pure ergative languages are Basque and Lezgian. Example of Basque's ergativity:
 * Martinek egunkariak erosten dizkit
 * Martin buys the newspapers for me

Ergative-absolutive alignment in essence distinguishes between the Agent on one side, and the Subject and Patient on the other. As shown in the example sentence, the ergative agent, here Martin, is marked. So far, no absolutive-marking languages have yet been found.

The ergative-absolutive alignment behaves as such:

A very common feature of ergative-absolutive aligning language is the antipassive voice which works by deleting the patient and promoting the ergative agent to an absolutive subject. Basque has a queer-working antipassive in that it promotes the agent but doesn't delete the patient, leading to a situation where both arguments are marked with the absolutive. An example of this antipassive variation:


 * Gauza miragarriak ikusi ditut nik.
 * Gauza miragarriak ikusia naiz.

The first sentence means "I have seen beautiful things", while the other would have to be translated into "I am seen beautiful things".

Split Ergativity
A strange phenomenon in ergative-absolutive languages that happens in certain languages, dividing the alignment according to person, is called split-ergativity. A good example of this subtype is Sumerian. Sumerian behaves as a nominative–accusative aligning language in the 1st and 2nd person of present-future tense and incompletive aspect, but as an ergative-absolutive aligning language in other forms of the indicative mood. Due to severe influence from Akkadian, Sumerian had aquired an almost purely nominative syntax by the late 3rd millenium BC, leaning to a topic-prominent language.

A good example of Sumerian split ergativity:


 * Lugal-e e2 mu-un-du3
 * E2 ib2-du3-un

The first sentence means "The king built the house", and its alignment is ergative, while the other means "I built the house", aligned nominatively.

Transitive-Intransitive
The transitive-intransitive morphosyntactic alignment is among the rarest in the world, and is the rarest amongst two-case systems. It works by distinguishing only between arguments of an intransitive and a transitive verb, that is, the Agent and Patient on one side and a Subject on the other.

A good example of a transitive-intransitive aligning language is the split-transitive language of Rushani, considered by some to be a dialect. It has a transitive-intransitive alignment in the past tense, but is variably nominative-accusative or double-accusative.

An example sentence, comparing literal English and Rushani:


 * Mu tā wunt
 * Me you saw

These languages are not well documented due to their relative obscurity.

The transitive-intransitive alignment behaves as such:

Tripartite
The tripartite morphosyntactic alignment is a MSA with three cases to mark the core arguments of intransitive and monotransitive verbs. In essence, it seperates all three arguments. It is among the rarest agreements in the world, found only in a handfull of languages.

Good examples of tripartite languages are Ainu, of Asia, and Nez Perce, of Northern America. Ainu has a tripartite system in its pronouns, and Nez Perce in both verb and noun agreement.

An example from Nez Perce:


 * Hitw̓alapáyna
 * ʔaw̓líwaaʔinpqawtaca

The first sentence means "He arrived in the rain", and the prefix "hi-" marks for the third person subject, while the second means "I go scoop him/it up in the fire", with "ʔew-" marking for the first or second person agent, and "-ʔilíw-" marking for the third person object.

The tripartite alignment behaves as such:

A common feature of tripartite language is their usage of both the passive and antipassive voices.

Due to their very clean cut agreement, conlangs often feature tripartite alignments; one notable such example is the conlang Na'vi, from the movie "Avatar".

Example sentences from Na'vi demonstrate:


 * Nantangìl frìp tutet
 * Nantang hahaw

The first sentence means "The viperwolf bites the person", while the second means "The viperwolf sleeps". In Na'vi, the "-t" suffix is the accusative suffix, the "-ìl" suffix is the ergative suffix, while the intransitive takes no suffix.

Extensions by...
Besides the rudimentary division into five base alignments with rigid marking roles, there are other morphosyntactic alignments based on other criteria. Examples of such alignments are "active-stative", "direct-inverse", "hierarchical" etc.

...Hierarchy
Every language in the world features some sort of hierarchy which either obligates or tends to put certain things in a certain order. This is unrelated to word order, as when those things could be in various positions according to word order rules, they tend to follow a certain order set by hierarchy. Even though this hierarchy goes beyond just MSA, it is often a vital component. Hierarchies are ordered by their main feature. The most common orderings are by: '''
 * Animacy
 * Control
 * Contact with speaker
 * Respect'''

Each of these has a different ordering and works differently upon certain categories of words. In Indo-European languages today, the most common ordering is by respect, thusly followed by contact. Semantically, control is also found ("The man raged" and "The storm raged" vs. "The pebble raged")

Eastern Native American languages have a system of person hierarchy similar to both respect and contact, and Algoquin hierarchy is most well known.

Basically, the hierarchies tend to behave as such:


 * Animacy
 * 1) Superanimate (divine)
 * 2) Human (often male followed by female)
 * 3) Animal
 * 4) Ignic (fiery)
 * 5) Marine
 * 6) Inanimate
 * -The last three categories can be conflated into one inanimate category, and the superanimate doesn't need to occur.


 * Control
 * 1) Superanimate
 * 2) Human
 * 3) Animal
 * 4) Elementary force
 * 5) Inanimate
 * -The superanimate doesn't need to be present, but elementary forces often are


 * Contact
 * 1) First-Person
 * 2) Second-Person
 * 3) Third-Person


 * Respect
 * 1) Second-Person
 * 2) Third-Person
 * 3) First-Person


 * (Generalised) Algoquin Personal
 * 1) Second-Person
 * 2) First-Person
 * 3) Third-Person

Hierarchies meddle in morphosyntactic alignment in many ways, most highly in languages with verbs sensitive to animacy.

Languages which have special constructions due to a person hierarchy often belong to a category of "direct-inverse" languages, where verbs have a special conjugation to indicate which argument preforms which action. Inverse-marking can work only when the two personal participants are at a different level in the hierarchy.

An example of a direct-inverse language is Ojibwe, an Algoquin language in addition. Example of direct/inverse-marking, and a demonstration that it is seperate from voice:
 * Ojibwe
 * Obizindawaan
 * o-bizindaw-aa-n
 * -listen.to-
 * Obizindawigoon
 * o-bizindaw-igoo-n
 * -listen.to-
 * Bizindawaa
 * bizindaw-aa
 * listen.to-


 * English
 * He listens to the other one
 * he listen-s to the other one
 *  listen-  other 
 * ''The other one listens to him
 * the other one listen-s to him
 *  other listen-  
 * He is being listened to
 * he is be-ing listen-ed to
 *   listen- 

Here, it is demonstrated that, in Ojibwe, there is no direct relation of transitivity and "directivity". The English gloss is just for flavour (and for the more involved readers to piece apart, as English has a very complex system of personal cases so intervowen you'd confuse them for an Indian tapestry).

Ojibwe has an addititon to its Algoquin hierarchy, which is why the above works. In the glosses, you shall see "3P" and "3O", which stand for "third person proximate" and likewise "obivate", respectively. In Ojibwe, this split of the third person causes a minor restructuring of the hierarchy:

Ojibwe Personal Hierarchy (Algoquin Modified):
 * 1) Second-Person
 * 2) First-Person
 * 3) Third-Person Proximate
 * 4) Third-Person Obivate

...Control
This division is based on the control of the subject or agent over the verb.

The most common occurence of this division is part of the split-intransitive alignment group. In split-intransitive languages, or shortly, Split-S languages, the intransitive argument, denoted Subject, is marked differently based on whether the subject is in control of his actions.

The glosses often used for these are SA, for the agent-aligning subject, and SO. Although I haven't done enough research to have any conclusive statistics, I have observed a tendency.

In languages with the intransitive split based on control, there are a few tendencies for what case is taken by the intransitive argument. Primarily, volountary actions are like to take the more active marking. This can extend to verbs such as "stand", which can begin to have the intransitive argument that takes the more active marking as the default. Conversely, instinctual actions can grow to expect the opposite. Such verbs include "urinate" and other bodily functions.

It is important to note that the intransitive argument can get marked either the same as the object or agent. This extends to the glosses. Intransitive arguments aligning with the agent are glossed SA, while the ones aligning with the object are glossed SO.

The most frequent division is that where SA is the more active intransitive argument, while SO is the more passive one, thus mapping activity to control.

''I cannot currently find examples of natural languages exhibiting a split in control, with minimal pairs. If you have, provide on the talk page.''

It is also possible for languages to divide the transitive arguments based on their control. This naturally includes the agent, but can only in rare situations also extend to the object. To have a Split-A system, there should be a prerequisite Split-S, but for a Split-O to arise, there needs to be a volition distinction already present (and, I imagine, a very ergative basis to begin with and add the module on). Agents with a more passive marking would take the marking of an intransitive subject, while the more active ones would take on their own. This is just one of the many plausible scenarios, and this delves into the realm of tripartitivity.

Note that Split-O means that the object specifically is split for control, not that it is marked for the agent's control. If the latter happens, it's a long-distance marking for the agent. Split-O could possibly follow in the footsteps of Split-A, with tripartitivity.

...Verb Preference
An extension on verb preference is a (d)evolution of control-based systems. The most frequent is the split on the intransitive argument. Although we could gloss the different arguments as SA and SO, they behave differently from control-based SA and SO.

While the control-based split assigns the different markings based on the intransitive argument's control over the action, verbs in this extension have their own inherent preference of marking. Some verbs may take SA, but they usually take only SA whilst completely disregarding SO.

This system comes out of a control-based system, and there exists a continuum between the two forms, often extending into a unified marking system.

Certain linguists, such as some members of the Indo-Uralist group of which I am a part, consider Pre-Proto-Indo-European, named Proto-Indo-Uralic (the common ancestor of Proto-Indo-European and Proto-Uralic), to be such a language which has already begun to level the distinction.

...Feature-Conditioned Factors
Feature-conditioned factors are a specific case of a split alignment, relying on different morphosyntactic alignments based on certain marked features, most commonly ones of verbs. The most thoroughly studied one would have to be split ergativity conditioned by the past or perfect, as found in several Indic languages. There can be those which are simple, depending on one single feature's presence, or more complex ones, with a continuum of features, or involving multiple features. A good example of this would be Georgian's MSA, the one which resembles a tapestry of interconnected and intermarked things, ranging from "screeves" governing the MSA, to the actual uses of the screeves based on the needs of the speaker.