Grammar Changes in Languages

[Acid Badger]

Can German even be SOV? From what I learned, the Subject can't be separated from the verb, so we get:

SVO
VSO (If it's starts with a prepositional phrase)
OVS

You also get verb final in subordinated clauses:

Ich weiss, dass er den Buch schlug.

And that's assumed to be the underlying order. You get V2 in finite clauses, but it's through movement of the verb to that other position.

*das Buch. Also, he hit the book?

[roninbodhisattva]

Can German even be SOV? From what I learned, the Subject can't be separated from the verb, so we get:

SVO
VSO (If it's starts with a prepositional phrase)
OVS

You also get verb final in subordinated clauses:

Ich weiss, dass er den Buch schlug.

And that's assumed to be the underlying order. You get V2 in finite clauses, but it's through movement of the verb to that other position.

*das Buch. Also, he hit the book?

Yup, I had originally put das Buch schrieb and then meant to change it to den Mann schlug....but forgot to change Buch > Mann. :/ I'll fix that.

[merijn]

Since my freshman year as a linguistic student I have never come across a linguist who would describe Dutch, and by extension German, as SVO underlyingly and not SOV*. The reason for this is that there are three exceptions to the SOV rule:
1) a normal clause where the order is V2:
2) when the object is a sub-clause
3) in imperatives and questions which are verb-intials,
When you assume SVO you still need exception 1) for when the clause starts with a non-subject. You also need exception 2 because objects that are sub-clauses follow adverbs that other objects precede. So you need a rule that moves the sub-clause to the right. 3) is a rule that you still need for questions though admittedly not for imperatives. In addition to those exception you also need to expain the different order in subclauses and for verbs other than the inflected verbs.
In mainstream Generative Grammar this is reflected by the idea that at D-structure (in GB) the object precedes the verb in Dutch and exceptions 1-3 are explained by movement. Usually different frameworks have different ways to express the same instinct, but I am usually not that interested in word-order so I don't really know how other frameworks do this, but I can imagine that for instance in Construction Grammar type of framework instances of SVO that are the result of exception 1 are the result of a construction [Topic] [Verb] [The rest] rather than an instance of a construction [Subject] [Verb] [Object]. (I have to admit that I don't really know that much about Construction Grammar so I may be completely on the wrong track)

*Except people who follow Kayne and believe that all languages are underlyingly SVO, but even they would say that there is a stage in the derivation of seemingly SVO sentences where the object precedes the verb.

[Ċeaddawīc]

I'm not claiming to be an expert, but it seems to me that the main factor of grammatical/syntactical change is reanalysis, where one more common construction in reanalyzed and less common ones follow.

[roninbodhisattva]

Extension is a very important process. In extension, a construction/grammatical element is extended from one grammatical context to another. My example of the irrealis construction in Salish is a good example of this.

[Basilius]

Since my freshman year as a linguistic student I have never come across a linguist who would describe Dutch, and by extension German, as SVO underlyingly and not SOV*.

This is what I called theory-bound analysis above.

You are using a framework that implies an early generation of a string already including all three of S, V, and O, with all the three already ordered. This is a peculiarity of a particular class of theories, not of German language. Can you point to a fundamental reason to stick to this class of theories for all languages forever? Even with transformation-based analyses, I can point to some (logical, even if not trendy) alternatives.

And it's only for the class of theories in question (and not all of them; you rightly mentioned Kayne's as one exception) German is typically analyzed as "underlyingly SOV". Note also that without adding "underlyingly" (which immediately makes the statement a theory-bound one) you get simply a lie, for the default surface WO in German is unquestionably SVO or, better, V2 (or, still better, a WO with no more than one argument permitted in pre-verbal position).

And, as I said, it's all irrelevant for diachronics where you need to explain the default surface WO first. (Do I need to explicate this?)

[cybrxkhan]

I am simply wondering about was whether, in general, there are grammar change rules or commonalities like that of sound changes.

OK, this sounds like something I can try to help with.

One rule of that kind that may look indeed as a rule:

Today's morphology is yesterday's syntax

(Givón, IIRC; not an exact quote; don't ask me where he said this...)

Is this the type of stuff you're interested in?

More or less.

Yes, I may (and do) wonder why, for instance, there are VSO, SOV, and SVO Indo-European languages when PIE obviously couldn't have been all of them at once; or why Sanskrit is ergative in a few situations when obviously languages like English or French aren't (or at least Wikipedia mentioned that somewhere...). HOWEVER, I'm more interested in whether there are any common patterns that cause these kinds - and all kinds - of grammar changes in general that allow so many different and - to me, it seems - ad hoc changes to occur.

But the rule you stated does sound like what I'm getting at.

[Count Iblis]

Alright, thanks for telling me about the issue with German; I'll have to look into it more, then. My main point stands, regardless; I'm curious as to what kind of situations and catalysts would bring about these changes. I'm not asking specifically about the grammar changes of German, or Welsh, or English, or Latin from PIE or whatever. I'm just wondering if you guys know if there are any typical reasons why changes in grammar happens through all languages in general. Regardless, thank you all for your replies and input.

This kind of typological change can occur from influences of foreign languages, especially substrate languages. If, for whatever reason, language A begins to dominate language B the speakers of language B will learn to speak A. Adults don't typically learn foreign languages well so aspects of their native language, like word order and phonology, tend to creep into the dominant language.

[cybrxkhan]

This kind of typological change can occur from influences of foreign languages, especially substrate languages. If, for whatever reason, language A begins to dominate language B the speakers of language B will learn to speak A. Adults don't typically learn foreign languages well so aspects of their native language, like word order and phonology, tend to creep into the dominant language.

Ah, right. I actually had forgot to mention that, it was something that popped in my mind. Influences from foreign languages can be pretty important, and I probably don't really need to list any examples; but I'm wondering also, and mainly, about changes that are largely independent of foreign influences, or are minimally influenced by foreign influences.

[Bedelato]

I don't know about the goose->geese example you posted

I can't believe nobody brought this up yet. It's just Germanic umlaut at work. Germanic umlaut is actually a great example of how a sound change can work its way into grammar.

Once i-mutation became phonemicized in pre-Old English, words like "goose" and "man" were left with a vowel alternation in their singular and plural forms, and the vowel change was analyzed as the plural marker.

[Jipí]

For those who don't know how it worked: an [​i​] or [j] raised an [a] in the previous syllable to [e] if there wasn't a back-consonant cluster (I think? I also seem to recall that there was something up with nasals) in between. In German this later spread to [o] and [​u​] as well, raising them to [ø] and [y]. As those [​i​]'s were usually unstressed, they withered away to [ə] (Nebensilbenabschwächung); word-final [ə] has a tendency to get dropped altogether. Such fossilized umlaut actually accounts for a lot of irregularities in German. In my class we illustrated this with Attila, who appears as Etzel in the Nibelungenlied:

attila → aʦila → ɛʦɪla → ɛʦələ → ɛʦəl

--
You might want to know there's the Zero Width Space at U+200B. I don't know the alt-code of that. This way you can circumvent writing [ i ] and still use BBCode.

[Ser]

You might want to know there's the Zero Width Space at U+200B. I don't know the alt-code of that. This way you can circumvent writing [ i ] and still use BBCode.

It would be theoretically Alt + 8203, but most programs don't support Alt codes above Alt + 255. In fact, the only one I know that does is MSN Messenger.

[linguoboy]

Yes, I may (and do) wonder why, for instance, there are VSO, SOV, and SVO Indo-European languages when PIE obviously couldn't have been all of them at once.

Is it obvious? I mean, yeah, the common wisdom is that non-configurational languages don't exist in nature. But there are enough of them with a lot of flexibility in word order, and this seems particularly true of those with a lot of congruence. Some languages fit much more comfortably into these typological classifications than others. Fronting even English allows!

[roninbodhisattva]

I mean, yeah, the common wisdom is that non-configurational languages don't exist in nature.

Citation?

[tezcatlip0ca]

You might want to know there's the Zero Width Space at U+200B. I don't know the alt-code of that. This way you can circumvent writing [ i ] and still use BBCode.

It would be theoretically Alt + 8204, but most programs don't support Alt codes above Alt + 255. In fact, the only one I know that does is MSN Messenger.

It doesn't work for me on MSN Messenger, but it does on Office (and only if the last two digits of the hex code are less than 0x20)

[Ser]

It doesn't work for me on MSN Messenger,

It does work for me though. Screenshot:

but it does on Office

Hadn't noticed, it works on Word and PowerPoint too! Although not in Excel.

(and only if the last two digits of the hex code are less than 0x20)

Can you explain this bit to me?

[tezcatlip0ca]

For example, In the Cyrillic block, "ie with grave" to "pe" don't work, because they map to C0 control characters.

[Ser]

I still don't understand what you mean, they work for me, too (Alt + 1049...1055 uppercase and Alt + 1081...1087 lowercase). What do "0x20" and "C0 control character" mean?

[tezcatlip0ca]

Look, 0x20 means hexadecimal 20, that is, 32. The C0 control characters are the controls in the original ASCII (from NULL to INFORMATION SEPARATOR ONE).

[Ser]

Oh, I see. Where did you get that they don't work unless the last two digits are less than 0x20 then? Because ‹й›~‹п› work perfectly in my computer's Word/PowerPoint entering them with Alt codes (and they end in 0x39 ~ 0x3F).

[tezcatlip0ca]

Capital "ie with grave" (not breve) 0x00 to "pe" 0x20.