Talk:Morphosyntactic alignment

I rewrote this page in an attempt to clarify the concept. However, after re-reading it I suspect that the new page is more confusing than the old one, perhaps even flatly wrong. (Specifically, my use of the words "subject" and "object" may be incorrect.) Would please some expert have a look, and fix or revert the page if necessary? Thanks, and sorry for the inconvenience...
Jorge Stolfi 03:54, 4 May 2004 (UTC)

I've just reverted it. In fact there were a few errors, but the main thing is that most of the new content could be found in the linked pages. I'm taking some time soon to write a new version, since the original one is a bit too terse and starts out flatly technical. Of course, many non-technical readers could not be bothered to learn about "morphosyntactic alignment" in the first place. The page needs examples, so the concept can be grasped more easily. Morphosyntactic alignment is not exactly an issue of subjects and objects, but of syntactic pivots -- i. e. which argument of the verb is the most basic, which can be left out and assumed to remain the same in propositions such as "he saw the deer and ran". Since English has accusative alignment, the implicit subject of "ran" is "he" and not "the deer". --- Pablo-flores 01:37, 7 May 2004 (UTC)

Well, I have tried to restore my introduction while keeping the old contents of the page. I agree that the page still needs examples. Also, the pages ergative-absolutive language and nominative-accusative language are fairly short, so perhaps they could be merged into this one, and turned into redirects.

But now I have a problem, I do not understand the difference is any more. To distinguish EA languages from NA languages, one must be able to distinguish the two roles A and P in a way that does not depend on noun inflection (otherwise the definition is circular). Then NA languages are those that mark A same as S, and EA languages are those that do the opposite; right?

But, one cannot use the semantical agent/patient idea to distinguish A from P, because of problematic examples like "to receive", "to suffer", "to wait", "to tolerate", etc.. One cannot say that A is the argument with which the verb agrees, because in some languages the verb is invariant, and in others it must agree with both arguments. Finally one cannot say that A is the sentence pivot because then EA and NA become identical...

So what is the way out? Could it be that "morphosyntactic alignment" is just a complicated way of saying in some languages, the arguments of some verbs are exchanged with respect to their closest English equivalents -- e..g as in Spanish "me gusta el libro" versus Portuguese "eu gosto do livro"? Help...

Jorge Stolfi 19:39, 12 May 2004 (UTC)

Sorry about the delay, I've had my comp in repairs for ten days already... The difference btw Erg-Abs and Nom-Acc is how they group S, A and P. This grouping is done by some means, either morphology or syntax (or both). I agree that, once you abstract S, A and P without assigning them semantic properties, then EA and NA are quite interchangeable, in the sense that for each one there are two arguments grouped and one left alone, and the names S, A and P really don't mean much.

One cannot always say that A is the argument with which the verb agrees, as you say, but if the verb is invariant it's quite common to have a fixed word order, and in those cases 95% of languages place A (=the subject) first, and P (=the object) second (SOV, SVO, or VSO). A is also the most agent-like argument for most verbs. More importantly, Erg-Abs and Nom-Acc are not interchangeable because of the fact that many EA languages have mixed alignment (erg-abs case marking but nom-acc syntax), so they're not symmetric.

Verb forms like "a mí me gusta el libro" vs. "I like the book" are examples of a common pattern that doesn't interfere with the main tendency of the language in question. Topics in the dative case occupying the subject slot are known in Russian, Latin and many other langs; they don't have to do with MS alignment but with pragmatics.

The Milewski's typology

I have added information on extremally interesting (but little known) typology created by a Polish linguist in 1960s.

--Grzegorj 18:19, 17 Mar 2005 (UTC)

I cleaned up Grzegorj's English and attempted to clarify the examples and table he provided. I think it's o.k. the way it is now, but frankly i don't know the first thing about the subject, instead working off of my own editorial experience and guesswork. I suggest someone look it over and take another crack.

Suggest removing Milewski's typology

This is a highly obscure and out-of-date take on the subject which is out of place in this article: if Milewski is to be treated in this much depth then there are at least half a dozen other scholars (especially Dixon, Du Bois come to mind) who should be treated in equal depth. A particular problem is that Milewski's use of terminology does not comply with that used in the rest of the article or the field as a whole. —The preceding unsigned comment was added by 82.33.156.75 (talk) 16:31, 21 January 2007 (UTC).

Well, I've made the deletion: if someone wants to reinstate it I suggest it goes into a separate article.

Dear Mister Nobody, if you decided to move this part to another article, don't you think you should have placed a link to it in this article?

Is it a common practice in Wikipedia that a Mr. Somebody unassigned requests removing some parts of an article, and then another Mr. Somebody, also unassigned, just does it without any further discussion?

Grzegorj 20:06, 18 June 2007 (UTC)

Transitive case?

Is there such a thing (or is it just another way of saying the 'accusative case')? Link is needed regardless. - IstvanWolf 14:16, 26 April 2006 (UTC)

Not sure in warrants its own article, though it wouldn't hurt. Rushani has one case for intransitive verbs, and a second for transitive verbs for both agent and patient - rather like a combined ergative + accusative. kwami 02:41, 27 April 2006 (UTC)

Cleaned up

The article was getting grammatical role (S, A and P) horrifically mixed up with semantic role (agent and patient). I have tried to clean this up by removing as much semantic role terminology as I could. A =/= agent, P =/= patient! —The preceding unsigned comment was added by 82.33.156.75 (talk) 16:44, 21 January 2007 (UTC).

Cleaned up

I cleaned up the artcle a bit and deleted some interesting but irrelevant information under nominative-accusative (the same can be said about languages with ergative alignment, but has nothing to do with ergativity or accusativity itself):

In some languages, there is more than one case for a single category; for example, Baltic-Finnic languages use both the accusative case and the partitive case to mark the O, in different distributions. Languages without case marking can identify the arguments through word order; for example, in Subject Verb Object languages the A argument precedes the verb while the O argument follows. Caseless languages may alternatively identify verbal arguments by coreferential agreement markers on the verb, or by incorporated pronouns.

Moreover, I would suggest to avoid such obsolete terms as nominative-accusative languages or ergative-absolutive languages, as there are hardly any purely ergative languages and a lot of languages with dominantly nominative-accusative alignment have traces of ergativity. In the same mode I would suggest to rename articles on Nominative-accusative language, etc. into articles Nominative-accusative alignment.
Finally, I would replace the Japanese example with an example from some other language, because as far as I know the wa marker in Japanese has also been analyzed as topic marker and not as case. --User:Newydd 13:19, 22 Jan 2006 (UTC)

Also, if I'm not mistaken, the "sentences" of the Japanese list "The", while the grammatical comparisons more correctly omit "the", as there is no comparable article in that language (iirc). As a simple workaround, I'm going to make the minor addition of putting parenthesis around (The) to hopefully show that this is implied, although I don't know enough about Japanese to feel secure writing up a footnote. Any takers (or anyone wanna switch the language entirely and avoid these issues?). 209.153.128.248 19:33, 6 November 2007 (UTC)

(1) I feel the translation should be colloquial English, and therefore we should not have parentheses around (the). No translation can do justice to the original, which is why we have the interlinear gloss. Also, to be consistent we'd have to do the same with the Basque.

(2) Ga is nominative, more or less, not a topic marker. However, it is optional, which means that it isn't case in the strict sense. Perhaps it's more of a nominative focus marker, but there isn't much agreement as to exactly what it does. Maybe not the best example, but the difference is a bit esoteric for the level of description we're giving. kwami 00:34, 7 November 2007 (UTC)

Added some graphics

I have added some nice graphics, that can improve this article a bit, I hope.

--dnik (talk) 00:02, 12 December 2007 (UTC)

Ditransitive and genitive/construct state alignment

Ideally the article should also treat how the indirect objects of ditransitive verbs are marked in various languages, vis-a-vis A, P, and S; and how genitives or construct states are marked relative to other cases in various languages. --Jim Henry (talk) 22:10, 12 September 2008 (UTC)

Do you have some references discussing these topics that we can draw on? --JWB (talk) 01:49, 13 September 2008 (UTC)

There are references in the ditransitive verb article itself for some of this; maybe some of the text there, and its corresponding references, could be copied here and abridged. As for genitive alignment, see this CONLANG list message and the books and articles referenced in said post. --Jim Henry (talk) 23:45, 14 September 2008 (UTC)

Odd sentence

I'd like to call attention for the odd phrase "Note that while the grammatical role labels S, A, and O/P are originally short for "subject", "agent", and "object/patient", the concepts are distinct both from "subject" and "object" (the terms that S, A and O supersede (so the label of 'object' supersedes itself?) and from "Agent" and "patient" (which indicate thematic relations, not grammatical relations: an A need (absence of third person 's') not be an agent, an O need (absence of third person 's') not be a patient)." —Preceding unsigned comment added by 189.46.210.43 (talk) 22:23, 16 April 2009 (UTC)

No, O is not "object". But I removed the word 'supersede' when I reworded for the next complaint. — kwami (talk) 10:43, 15 January 2012 (UTC)

Too difficult and needs more examples

I am a foreign language teacher and my degree included linguistics (at a prestigious ancient university), yet I find this article impenetrable. Speaking as a teacher, I think it needs to work with examples much more. Right at the start, it needs clear examples to explain the A, O and S terminology, which are confusing. There is one little example (Julius Brutum vidit - helpful, but even that's in Latin), and that's all. Obviously this is never going to be an article for the reader with zero background in language study, but it needs to be accessible to people who have such a background. Wikipedia articles are not textbooks for academics. In particular, I think the first twenty lines or so of an article need to be comprehensible to a generally informed reader, even if later sections may become increasingly complex. APW (talk) 09:06, 15 January 2012 (UTC)

I took a crack at the first section, and also added the English situation to the lead (and fixed a few factual errors). It should read a little better.

The problem with giving clear examples is that, in English at least, there really aren't any. Or, if there are, you could probably write a textbook with them. — kwami (talk) 10:41, 15 January 2012 (UTC)

It's a big improvement. Many thanks! APW (talk) 07:50, 21 June 2012 (UTC)

A/O/S being confused under "Semantics and grammatical relations" section

At the beginning of the section, in the first paragraph, it's written:

″Transitive verbs have two core arguments, labelled A (the more active or in-control) and O, which in a language like English are subject (A) and object (O). Intransitive verbs have a single core argument, labelled S, which in English (but not in all languages) is also a subject.″

So according to this paragraph, transitive verb sentences take A and O, while intransitive verb sentences take only S.

But then further down, in the fourth paragraph, an example is given of the English verb "moved", and the usage of A/O/S is contrary to paragraph one!

″The best-known system is the English type, with a distinct O (object). These are called nominative–accusative languages, or just accusative languages, after the nominative and accusative cases, which are how A/S and O are distinguished in Latin. The best known of the other systems is the ergative system, named after the ergative case, which is how A is marked in many languages (such as Inuit and Basque) which distinguishes A from S/O. An example in English of a verb used in the ergative case is the verb "moved" in the two sentences “John moved the stone,” and “The stone moved.” In the first sentence, since the subject (S) (John) performs the verb upon the object (O) (the stone), the verb (moved) is transitive. In the second sentence, however, the stone becomes the subject (A), which performs the verb, and the object disappears, and thus, because the verb does not take a direct object, it is intransitive. ...″

In the example of "John moved the stone" and "the stone moved", the transitive sentence is said to take S and O, while the intransitive is said to take A!

Can someone please clarify which label system is correct, and change paragraphs 1 and 4 to match up with each other. Thank you. — Preceding unsigned comment added by Jclu (talk • contribs) 02:43, 24 June 2014 (UTC)

A Typology of Marked-S languages

The following book is on a more specialized subfield of alignment, but it might be instructive nevertheless. It is an Open Access Book, which is good, but I am not sure in how far it is on topic, so I leave it for other people to decide whether or not to add it to the external links

"A typology of marked-S languages" Corinna Handschuh

http://langsci-press.org/catalog/book/18

A typological study of the rare marked-S language type which overtly marks the single argument of intransitive verbs (S) while one of the arguments of transitive verbs (either A or P) is left zero-coded. The formal (overt versus zero-coding) as well as functional aspects (range of uses of individual case forms) of the phenomenon are treated. The book covers languages from the Afro-Asiatic and Nilo-Saharan languages of Africa and of the North America Pacific Northwest and Pacific regions. — Preceding unsigned comment added by 130.133.5.142 (talk) 12:34, 28 October 2014 (UTC)

"moved" example seems incorrect/out of place

From the article:

An example in English of a verb used in the ergative case is the verb "moved" in the two sentences “John moved the stone,” and “The stone moved.” In the first sentence, since the subject (A) (John) performs the verb upon the object (O) (the stone), the verb (moved) is transitive. In the second sentence, however, the stone becomes the subject (S), which performs the verb, and the object disappears, and thus, because the verb does not take a direct object, it is intransitive. Therefore, the verb "moved" is an ergative verb because it can be used both transitively and intransitively.

Firstly "a verb used in the ergative case" doesn't seem to make sense, nouns have case not verbs. Assuming it means "a verb used with the ergative case", it's still not correct - there's no ergative case here, there's a subject case (used for both A (John) and S (the stone in the intransitive sentence)) and an object case (used for O (the stone in the transitive sentence)). The final sentence says "moved" is called an ergative verb because it can be used both transitively and intransitively which may be true but is irrelevant to this article.

I am going to delete the entire example. --Yym (talk) 10:14, 23 May 2017 (UTC)