Analytic proto-language for an agglutinative conlang

[gojin]

Hello,

as explained in the ACK on p. 105, languages can be seen as going through a cycle of agglutinative -> fusional -> isolating typology, at least in the long-term.

Here is another Q&A discussing the matter: https://www.quora.com/Historical-Lingui ... back-again

Now, what's kind of puzzling me is the role that syntax (word order) plays in that regard.

Assume that English is our proto-language, I find it perfectly reasonable that, over some time, a sentence like I'm going into the house turns into something like I'm going intothehouse -- where "into-" is a prefix meaning, well, "into", and "-the-" became a marker of definiteness, but is no longer a separate article.

OK, now if we stick to that rather silly example, it seems that the whole process is highly dependent on word order. If "into" In English is a preposition rather than a postposition, any derived language would have "into-" as a prefix, never as a suffix.

It seems that English into the house could never turn into something like a Finnish noun in illative case taloon, simply because the word order is entirely different.

So, if I have a conlang following the agglutinative paradigm, with a lot of case endings (and noun suffixes in general), am I basically stuck with that word order when creating a proto-language afterwards?

My impression is that if a language has a word like "taloon", the only way a proto-language could have worked is that it had "talo" as a separate word, as well as "on" as a separate word, where "on" must have been a postposition rather than a preposition.

Is this correct or were there any historical changes in natlangs that suggest otherwise?

(I left out possible sound changes to make it clearer that the question is about word order...)

Thanks,
gojin

[Travis B.]

There are other outcomes of the development of agglutinativity. Take your example, for instance: one could see into becoming an clitic on the verb turning it into, e.g. I'm going-into the house, and getting in turn transferred to the finite verb, becoming an applicative, e.g. I'm-into going the house.

(Note that English already can sometimes use prepositions as prefixes on verbs like applicatives like "The gazelle outran the lion", but this is not really fully productive.)

[gojin]

I was thinking the same thing -- I'm going into (the) house might become I'm going intohouse, but it might as well become I'm goinginto house. But then again, this depends on word order, right? into is not going to appear as a noun suffix all of a sudden.

You're suggesting that into, once it turned into a suffix, kind of starts wandering around in the sentence. Someone might choose to apply it to a verb that has a different syntactic role instead. Would it be reasonable (realistic) to assume that once into became a suffix, someone would apply it to a noun for some reason (thereby forming new case endings)? What is the linguistic terminology for such sort of change?

[Curlyjimsam]

Remember that word order isn't fixed: a language can shift from (say) postpositional to prepositional. So a postposition might fuse to a noun as a case suffix (not prefix), and then the language shifts to prepositional order. This has happened, or been theorised to anyway, in a number of real world languages.

[HoskhMatriarch]

Well, if you have more than one possible order of different elements in a sentence, the one that gives you syntax is not necessarily going to be the most common one. For example, Scandinavian languages and Icelandic have suffixed articles rather than prefixed articles even though earlier Old Norse had articles normally precede nouns, mostly because the stress was on the first syllable of the word always so for the definite article to become affixed onto the noun, it has to be in an unstressed position at the end of the word. However, to do that you have to have some flexibility of word order, however slight. Most (all?) languages don't have completely fixed word orders, not even generally analytic languages like English and Mandarin, so you can play around a lot.

[gojin]

@HoskhMatriarch
Thanks for posting these interesting examples.

[HoskhMatriarch]

I just edited my post a lot. Sorry.