Talk:Linguistic typology

From Wikipedia, the free encyclopedia

Contents 1 German 2 whining about redundancy 3 subject weak, declension, inflection 4 topic comment? 5 Linguistic typology and inflexional languages 6 Koine 7 difference between ergative and accusative 8 Proposal: Standard typology description in articles about languages 9 Word order and agent 10 Word order and agent 2 11 SVO, AVP, etc. 12 Re: Linguistic Topics link 13 A "tentative" field?

[edit] German

I would argue that German is more basically SOV/SVO, not VSO. This is a minor word-order, akin to the English Have you seen my dog?, in which the non-analytic verb is have, and the word-order there VSO (but in the main, it is SVO: I have seen your dog.) Without Im Wald, the given sentence perfectly demonstrates the two major word orders of German:

Ich habe einen Fuchs gesehen
I   have a   fox   seen

SOV occurs when the verb is non-finite, and SVO occurs when a finite verb appears in the main clause. It's not unusual for some languages to have fairly free and variable word-orders, and the word-order is very rarely completely rigid. In Wichita, for instance, the two major word orders are SOV and OVS, but SVO can occur in isolated sentences translated from English. In Basque, the profusion of cases means that word order is almost entirely free in basic sentences. thefamouseccles 23:58 25 Oct 2003 (UTC)

[edit] whining about redundancy

I don't quite see the point of a bunch of pages

   * Subject Verb Object
   * Subject Object Verb
   * Verb Subject Object
   * Verb Object Subject
   * Object Subject Verb
   * Object Verb Subject

that merely repeat information from the Linguistic typology page. Couldn't we reduce all of those pages to a single line "See Linguistic typology." ?

DavidCary 02:49, 20 Feb 2004 (UTC)

[edit] subject weak, declension, inflection

Where do concepts like subject weak, declension and inflection fit into this? I'm sure I'm not understanding the classification properly so maybe some extra information for us beginners would be useful (thanks) --KayEss 21:50, 29 Mar 2004 (UTC)

[edit] topic comment?

I'm really quite interested in Linguistic Typology, as it seems to offer a great deal of insight into languages. Difficulties with the SVO/VSO/... system of classification have been pointed out, primarily in dealing with polysynthetic languages whose rich morphology allows speakers to permute the word order as they see fit.

While this is a valid concern, as a speaker of Chinese I am somewhat more concerned by its aparent failure to deal with topic comment (or sometimes topic predicate) lanugages. ASL (American Sign Language) is another example of a topic-comment language, as is Japanese (or so I've read, I speak/sign neither of these languages). Essentially, in a topic comment language the structure of the sentence is typically 'topic' followed by a predicate, or comment, which comments on the topic in some way. Lingusitic typologists seem wont to classify Chinese as SVO, but it becomes very difficult to satisfactorily model the language in this way. Rather like the tendency of speakers of nominative-accusative languages to confuse agent/experiencer with subject and patient with object because the tripartite distinction makes very little sense in the context of their own languages, it seems to me as though typologists are drawing heavily on their intuitive understanding of European languages and attempting to extend that framework globally.

The categorization of Chinese as SVO seems to be motivated by the fact that the subject typically precedes the verb, and that the object (if there is one) typically follows the verb. However, this is overlooking the fact that in most sentences, the subject is the topic of the sentence. As a result, one has a situation where a satistically significant number of sentences seem to begin with a subject, but in fact, they begin with a topic.

And, as a consequence, virtually any word can be regarded as the subject, the verb or the object in a Mandarin sentence.

Typically, a chinese sentence is broken down into topic-comment, comment, comment, comment .... (with the comments all being predicates that share the same topic, introduced at the sentence head). In fact, the Chinese notion of a sentence is very different from an English sentence -- the above construction would invariably lead to a run-on sentence in a language like English, but is perfectly acceptable in both written and spoken Chinese.

As I've implied, the topic need not be the subject, but may in fact be any part of speech. The canonical example is the sentence 七岁时死了他父亲 (literally 7 years old died his father) which of course means "His father died when he (the child) was seven" but leverages Chinese topic-comment sentence structuring to make the age of the child the topic of the sentence. The intrasitivity of the verb 死 (to die) allows the "subject" to be placed anywhere without confusion (as the subject is actually the experiencer and not the agent).

This accounts for the frequent use of sentences that seem (to speakers of Indo-European languages, at any rate) to be in the passive voice, but are in fact not in the passive voice (Chinese has a well marked passive voice). This is because in those sentences, the object acts as the topic and is brought to the beginning of the sentence, with the subject taking second place (or being ommited completely). An example of this might be 车洗好了 (literally, car washed COMPLETE) which introduces the car as both the patient and the topic (and ommits the subject). This could be translated (inaccurately) into an English framework in two ways: I washed the car (actually 我已把车洗好了, which introduces "me" as the topic of the sentence, and uses the active voice), or The car has been washed (actually 车被洗好了, which keeps the car as the topic of the sentence, but emphasizes -- through the use of the passive voice -- the existance of an agent). The actual sentence in chinese is talking about the car, and is commenting on it: it's been washed. This is rather hard to do in English.

With this background, I'd like to hear from some of the linguists around what their thoughts are on topic-comment structures. It's essentially impossible to speak or understand Chinese if you adhere strictly to a SVO classification, and it's different from languages like German which are nominally SVO but SOV and VSO characteristics in certain instances, because in those languages, the grammar (rather than the semantics) completely predicts deviations from the norm (ie, German using SOV structure in subordinate clauses, and VSO when a prepositional phrase is brought to the head of the sentence). In Chinese, there are no "rules" that predict when a sentence will depart from SVO -- the only way to know how to structure a sentence is to drop the SVO fallacy completely and accept the topic-comment model.

Is Topic-Comment a part of linguistic typology? It occurs frequently in literature (esp with regards to Chinese, Japanese, and ASL).

If it is, maybe we ought to have an article on it. If not, maybe we ought to have one anyway, and point this out as a possible limitation of Typology. 202.96.246.206 06:56, 14 Dec 2004 (UTC)

Oh, if you're interested in a better summary than mine, with more examples, check out this PDF. http://www.latrobe.edu.au/linguistics/stlapolla_data/PublicationItems/Miscellaneous/clause_structure.pdf Just found it on-line. 202.96.246.206 07:30, 14 Dec 2004 (UTC)

Of course the Topic-Comment problem can be a part of typology. In fact, the S-V-O order is one of very numerous criteria used for classifying languages. You can divide languages in groups depending on the number of phonemes in them, on number of personal pronouns (modern English use "you" for both singular and plural while Old English had two distinct forms here instead - and it can be count in a different type of languages), on presence or absence of dual, or even on presence or absence of the voiceless laryngeal spirant [h]. So, why not divide languages in groups taking into consideration e.g. the position of the topic in the sentence?

At the same time I do not think that the Topic-Comment problem could ruin the S-V-O classification. Basing on my scarce knowledge (and on books I have read) I can state categorically that Mandarin Chinese belongs to SVO type. And as far as I know, such a word order is enough rigid. But Mandarin is an isolating language in which words are not strictly divided into parts of speech, such as we know from IE languages. For example, the word zài (4th tone) can be interpreted as a verb ("is", "there is") in some sentences while as a part of a coverb (e.g. zài X li = in X, zài X shang = on X etc.) in others: Tāmen zài nèi jiān kètáng li means They are in that classroom but Tāmen zài nèi jiān kètáng li shàngkè means They are studying in that classroom. However, notice that the second sentence can be also interpreted as They are in that classroom, studying.

There are at least two grammatic facts which make us accept that Mandarin is a language of SVO type.

1. Changing the positions of the object and the subject causes the passive meaning: tā wánchéngle jìhuà = he has performed the plan, but jìhuà wánchéngle = the plan has been performed. Both sentences can be interpreted as SVO or SV. So, when the object is at the subject slot, it becomes subject in fact.
2. Changing the positions of the object and the verb is impossible at all. Instead, Mandarin uses a special kind of an auxiliary verb to fill the slot of V. Just compare wǒ zhǎodàole gāngbǐ = I have found a pen and wǒ bǎ gāngbǐ zhǎodàole, which can be translated as I have done a pen found with the "particle" bǎ understood as an auxiliary verb similar to English do.

Of course, Mandarin grammar knows "exceptions" like Láile kèrén - Guests have come (literary come-COMPLETE guests) or like sǐle fùqīn (die-COMPLETE father) in the example mentioned in the original comment above. However, as far as I know, such sentences are impossible with transitive verbs at all, so they do not break the basic order SVO. Who said that the word order in an intransitive sentence must be SV if the order in a transitive sentence is SVO?

And one more comment: 七岁时死了他父亲 (in the original text qī suì shí sǐle fùqīn) is not a sentence but a clause, i.e. a part of a sentence. It lacks the subject (rén) which was mentioned in the main clause.

I have used Latin script (script, not transcription!) for Chinese sentences for two reasons: it is a very hard task to write hànzì on a non-Chinese operation system, and Latin writings are much more readable for those who (like me) cannot read the ideographic script.

--Grzegorj 15:01, 17 Mar 2005 (UTC)

[edit] Linguistic typology and inflexional languages

There exists a myth that languages with rich morphology (and especially with noun cases) have free word order in sentences. I do not think it could be true in any language.

I think that a kind of a neutral word order is always essentially fixed in a given sentence ("given" - because it can depends on lexical facts, i.e. whether such or another verb / object / subject is used). Anyway, it is so in Polish, which is my mother tongue. Polish has noun cases, and it is said to have very free word order. In fact, the neutral word order in Polish is SVO in most sentences. The problem is, however, that this word order can be changed - but not freely, with no reason! In other words, the word order is "free" (in a special sense - you can find many sentences with the word order other than SVO in any Polish texts) but it cannot be freely changed.

On the other hand, there can be many reasons to change the "neutral" word order. Such a change is frequent, among others, in poetry (e.g. for preserving the metrum, the rhythm or the rhyme), or when you want to "topicize" a given part of a sentence. For example: co widziałeś w lesie? - w lesie widziałem wilka (what-ACC saw-2sg in forest-LOC - in forest-LOC saw-1sg wolf-ACC), which means what did you see in the forest? - I saw a wolf in the forest. Another example shows breaking the SVO order itself: Piotra bije Paweł (Peter-ACC beat-3sg Paul-NOM) - Paul is beating Peter. In this sentence the speaker gives the additional information that it is Peter who is beating Paul (Peter and not another person). A similar effect can be achieved in English with the help of intonation: Paul is beating Peter! with "Paul" spoken at a higher pitch of voice.

But there is another fact which is often omitted by typologists: some words tend to change the basic word order. Some examples. An intransitive word causes the SV word order as a rule. There are two verbs that can be used with the meaning "exist", they are być istnieć. The first one is just to be. It needs the VS word order in this meaning: jest Bóg (is-3sg God-NOM; God exists or there is God), w lesie są grzyby (in forest-LOC are-3pl fungi-NOMpl, there are mushrooms in the forest). But the synonymous word istnieć needs SV in simple sentences, like Bóg istnieje (God exists) while VS in more complex sentences (e.g. w kosmosie istnieją różne rodzaje gwiazd - in space-LOC exist-3pl various-NOMpl kinds-NOMpl stars-GENpl, various types of stars are/exist in the space). In boli mnie głowa - ache-3sg me-ACC head-NOM (I have a headache - or more literally my head is aching me) the VOS order is frequent (along with głowa mnie boli - SOV). In the subordinate close in chcę cię spytać (want-1sg you-ACCsg ask-INF, I want to ask you) we have OV, while in chcę spytać Piotra (I want to ask Peter) we have VO (no subject as it is the same as in the main clause). As you can see, here the word order depends on what part of speech is used, a pronoun or a noun. Etc...

As you can see, neutral word order does exist in Polish (and I belive, in all the other inflexional languages) even if it can be easily broken (e.g. in poetry, for emphase, topicization etc.). Which is more, it depends on lexical and semantic facts - it is not so that all verbs demand the same word order. Nevertheless, a given word needs "its own" word order (even if the word order cannot be the same as that one which is prevalent in the language).

Does it all mean that such languages like Polish cannot be classified within the S-V-O typology? No! But you should take only natural, neutral word order into consideration, such as in sentences with "typical" words - or just more frequent order (statistically) which can be found in texts. Taking this important notice, Polish belongs to the SVO type, Latin - to SOV etc., even if virtually any of the six orders can be found in Polish or Latin texts. Statistics seems to be the only method in such instances. And hence I am sure that Basque (or any other language) can be classified within one of the six types.

--Grzegorj 13:19, 17 Mar 2005 (UTC)

[edit] Koine

The use κοινη without proper accentuation, modifying the English word "Greek" was a ridiculous pedanticism, so I removed it. Besides, Modern Greek or Mycenean Greek are satisfactory examples of this phenomenon as well, so you may as well say simply "Greek."--Jpbrenna 22:35, 13 Jun 2005 (UTC)

[edit] difference between ergative and accusative

I didn't understand the difference between these two. Maybe some examples would clarify the section? Someone who understands it could maybe classify English, Spanish, French, Russian, etc.?

ANSWER:

All these languages are accusative; any use of ergative about them is mistaken (although 'ergative' has often been used in this context regarding English). Setting aside voice such as the passive, in accusative syntax, the 'subject' (instigator or experiencer) of an intransitive construction such as the 'I' in 'I smile' shares a case with the 'subject' of a transitive construction such as the 'I' in 'I love you'. This case is the nominative. The 'direct object' is excluded from this case, and is given its own 'odd-man-out' case, the accusative. The nominative participant is basic (i.e. if only one argument is marked on the verb, it must be the nominative). English distinguishes nominative from accusative by word order (nominative+accusative+verb), although 'he'/'him' variations occur as well. In ergative syntax, the 'subject' of the intransitive (the 'I' of 'I smile') shares a case with the 'direct object' (the 'you' of 'I love you'). This case is called the absolutive. The 'subject' of the transitive verb (the 'I' of 'I love you') is then the odd-man-out, and is called the ergative. The ergative cannot be marked on the verb before the absolutive. So, the difference lies in the case that is used for the intransitive subject; in accusative languages, it is the same as the transitive subject, but in ergative languages it is the same as the direct object. This is NOT obligatory passivisation, as passives are not really transitive. Note that many languages have split systems, which use accusative and ergative syntax in different contexts.

[edit] Proposal: Standard typology description in articles about languages

I suggest to introduce a standard section, "Typology", in language articles. Since I am not an expert linguist, I cannot suggest a good format. Any ideas?

Can it be as simple as an infobox? (I don't think that {{Infobox Language }}, already big (seee eg English language), can house typology as well, since the typology has several facets). `'mikka (t) 01:10, 22 April 2006 (UTC)

Indeed, I wouldn't put it in Infobox Language. I'm not sure it is a good idea to have a separate section on this — I doubt it would be of interest to non-typologists. Currently, typological features are often simply discussed in the standard sections about grammar and morphology and that's fine with me.

[edit] Word order and agent

The section about word order in the beginning is misleading. The employed concepts are indeed Subject, Verb and Object (syntactic) and not agent and patient (semantic), as the section now suggest. Consider English, a rigid SVO language:

1) The girl hits the ball 2) The ball is hit by the girl

'Girl' is agent in both sentences, but subject only in the first. The subject of the second sentence is 'ball'. This illustrates that subjects always precede the verb in English, while agents do not.

145.18.233.228 10:30, 7 August 2006 (UTC)

Yes, I've changed it back to the original form. User:Angr 20:42, 23 August 2006 (UTC)

[edit] Word order and agent 2

Agent and patient are preferable as not all languages 'have' subjects and objects but all languages have means to express agent and patient. The examples from English show only that in passive the reversed Patient Verb Agent constituent order is found, but it doesn't mean that this is the most frequent order in 'neutral' contexts, thus English is an Agent Verb Patient word order, though it allows for variations. And one more comment, in English subjects often but not always preceeds the verb (what about interrogative sentences). --Newydd 01:29, 24 Aug 2004 (UTC)

What languages don't have subjects and objects? And don't say "ergative languages", because that simply isn't true. (And even if it were, the terms SVO/SOV etc. are too well entrenched in the literature for us to change it here. Doing so would be original research.) "Subject" and "object" are terms of syntax, while "agent" and "patient" are terms of semantics. Word order is part of syntax, not semantics, and as the poster above mentioned, in English, active sentences are Agent-Verb-Patient while passive sentences are Patient-Verb-Agent, but both are Subject-Verb(-Object). In English interrogative sentences, do-support ensures that even they remain SVO ("Did Mary_S kick_V the ball_O?" Wh-questions can become OSV (What did Mary kick?), but even there the subject precedes the verb. User:Angr 04:59, 24 August 2006 (UTC)

Answer: Dear Angr, I am not insisting on replacing "subject" and "object" with "agent" and "patient" as they are indeed too well entrenched in the literature, but may be it is worth mentioning that the alternative terminology is quite widespread in the typology as well (see, for instance, Payne T. 1997. Describing morphosyntax. and authors mentioned in this introductory work in the chapter on Grammatical relations). Speaking of languages without subjects. I meant that in some languages, for instance, those with syntactic (and not morphologic) ergativity, the situation is not that straightforward. In these languages the syntactic properties traditionally considered as the key identifiers of grammatical relations (say, agreement, case, but also relativization or behavior in raising and control constructions) do not converge on a single set of GRs (thus, for instance, according to verb agreement one argument seems to be the subject but according to case a different one). So, there is no single subject in a way, so it does not make sense to speak of subject in such languages. I can provide a number of references, but it goes far beyond the scope of this article.
One more comment, Agent and Patient as used by (at least some) typologists are not purely semantic but true semantico-syntactic roles.
As for English, I meant such questions as "What did you do yesterday?" but also sentenses as "Dogs I hate." --Newydd 13:02, 24 Aug 2004 (UTC)

[edit] SVO, AVP, etc.

The article is interesting. However, if you are not a linguist, it can be difficult to understand as it doesn't fully explain terms like AVP, VAP, etc. It is only by reading the discussion that I understood them. Adding a paragraph to define the different terms and highlight the difference between say SVO and AVP would help clarify it immensely. For instance, you could give a quick example like "The dogs bites the cat" compared to "The cat is bitten by the dog" where both are SVO but the former is AVP and the latter PVA.

[edit] Re: Linguistic Topics link

Link in question: http://www.uio.no/studier/emner/hf/ikos/EXFAC03-AAS/h05/larestoff/linguistics/Chapter%204.(H05).pdf

Looking through the source it looks to be pretty good resource as an external link to provide. It think it would be a very good addition to the article. However the link as provided provides no information about author, copyright, or anything. Neither did the link description prescribe any authority to the link. I'm really sorry, but it shouldn't be here if we don't know who wrote it, if it's legal, etc --Puellanivis (talk) 21:46, 15 December 2007 (UTC)

Well, it's freely available on the Web, so why not? The link does no more and no less than simply link to a freely available source. If we wrote a description on how to access it via Google, would that then be legitimate? This way the author might come upon the link and claim the credit that is due to him/her. Tkeu (talk) 22:36, 15 December 2007 (UTC)

"freely available on the web" is not really a good standard. While I'm sure someone could steal some proprietary source code and distribute it over the internet, it's still not "freely" available on the web. Likewise, if this is simply a scan of a book, or otherwise illegally reproduced, I think it's kind of the duty of a responsible community to not endorse copyright infringement. Honestly, it looks like there are a few chapters to this, someone should be able to walk up the tree, or whatever, and find out who wrote it, and if it's an authorized copy of the material. --Puellanivis (talk) 02:13, 16 December 2007 (UTC)

OK, you're right. I've 'walked up the tree' and this is what I've found: the file comes from the University of Oslo website. Witness the following link, which is a root link for the file: http://www.uio.no/studier/emner/hf/ikos/EXFAC03-AAS/h05/larestoff. Next, expand the link further, and you discover this: http://www.uio.no/studier/emner/hf/ikos/EXFAC03-AAS/h05/larestoff. So it is freely available. I'll re-add the link and add information about the authors. Tkeu (talk) 04:24, 16 December 2007 (UTC)

[edit] A "tentative" field?

The SOV, SVO, et cetera classifications seem to be more tentative than absolute. I think this article explains the difficulties fairly well. Perhaps one might add that linguistic typology seems to be fairly young as a distinguished scientific field; note that the Association for Linguistic Typology was founded as late as 1994. The trouble is, that we have a tendency to believe that "the name is the thing".

In other words, according to the article, Russian is classified as an SVO language, since frequency investigations of the occurring orders in a lot of sequences revealed that the SVO order was more common than others. Actually, since Russian has a fairly free world order, other investigations might come to other conclusions; this might even be a mode thing. If some speakers/writers of Russian should be using another order somewhat more frequently, and SVO somewhat less, then you might conclude that their Russian should be classified differently; although most would agree that their language is as much Russian as that sampled before.

However, it is easy enough to read the statement as Russian is an SVO language, and take this to mean that this is an absolute property of some languages, Russian among them. I do not believe that the linguistic typologists themselves would make that kind of mistake to a high extent, but for us others, it is all to easy to assume that a term like "SVO" covers an objectively existing "essence" shared by a large group of languages.

So, I like this article, but I think that we should be careful not to refer to specific languages as "being" e.g. SVO or SOV, especially not in the introductions of the separate articles about the languages. JoergenB (talk) 21:08, 8 May 2008 (UTC)