Grammaticalization Quickie Thread

[bradrn]

The first thing that strikes me as probable is that this giant crufty qualifier thaŋmu tlaqufli would like to be smaller, and fall together into a single word; a bit of assimilation and loss of unstressed syllables and glottal stops, and we have "thamtlau", possibly simplifying further to mtlau, mblau, mlau, blau (I hope all that reduction won't make you blue, German or otherwise).

Well, firstly, that’s not even a single constituent: that would be *sasat thaŋmu ‘the wind’ (with a 2P past tense clitic). With sound changes and cliticisation, it ends up as sæssæŋ. (I think. I’m still working out the sound changes.) *Tlaquf is very much acting as a verb here — that’s why it gets an aspect suffix.

This change should suggest that verbs intimately connected with a nominal ought to follow it, inverting the order of waqli gilut to gilut waqli.

Why? I don’t follow.

This could, in turn, trigger the sequence of verbs to merge together — waq fawetlli > (disinflect the first verb) wappautli

Well, *fawetlli here is acting as more as a quotative than a full verb, so I’d say it would most probably end up as a quotative particle faw. What I’d really want is for the SVC *tlaqufli … waqli to fuse together into something like šɔʔ-wæ-l, where the verb has grammaticalised into a comitative. But I’d need to change word order to SOV before I could do that. (Though mind you, I could probably do it anyway — reanalysing *qalit thaŋ as an indirect object seems realistic.)

[Rounin Ryuuji]

Indications that I did that badly, prompting me to do much more reading on the subject.

Well, perhaps I ought to have just done what I do best — ruin the phonology to get a desired phonetic outcome (I think I understand some of the sound sequences you like, anyway, even if I think they're ugly not the most aesthetically-pleasing) and see if the pieces won't align themselves — from the beginning.

Phase 1: [ˈsasat ˈtʰaŋmu ˈt͡ɬaʔufli ˈʔalit tʰaŋ ˈwaʔli ˈɣilut ˈfawət͡ɬli | ˈnaʔəŋ ˈkʰajiʔənəbəj t͡sʰaːl ˈpawtiʔ]
(The language is as presented.)

Phase 2: ['saː.zəs.tʰam 't͡ʃou.lʲə 'aː.lʲəs.tʰəŋ 'waː.lʲə jiː.lət faut͡ʃəl | naːŋ kajiːnwe t͡sʰaːl pə'tiː]
(Absorbing the articles into the nominal stem and reducing unstressed syllables, with compensatory lengthening of some vowels; deleting a glottal stop lengthens a preceding vowel, and drags stress.]

Phase 3: ['saːstəm t͡ʃəu 'ais.təŋ t͡ʃəu 'wai jiut fau(t͡ʃ) | naːŋ 'kain.wə t͡sʰau psai]
(More reduction of inflectional endings; element [t͡ʃəu], now an unstressed conjunction, ends up redoubled, and is probably still conjugable, but I don't know what the morphology would look like.)

Phase 4: ['saːstəm 'ais.təŋ t͡ʃə'wai jiut | naːŋ 'kain.wə t͡sʰau psai | fəu]
(The first [t͡ʃəu] is dropped, and the second one reduced enough to be absorbed into the verb following. The route is a bit roundabout, but it gets you to the two being compounded. The element [fəu] is fully reanalysed as a quotative particle and put after the quotation, but I don't know what a verb for "said" to presumably come after it would look like.)

Phase 5: ['saːstəw 'ais.təj jiut t͡ʃə'wai | naːŋ 'kain.wə t͡sʰau psai | fau]7

The language completes its shift to being head-final. I have no idea what these ended up doing to the rest of the morphology, so pardon that probable weirdness.

Phase 6:['saːs.tʰoː 'eːs.t͡sʰe ɟ͡ʑyːc͡ɕʰ c͡ɕʰe'çɥeː | nu'woː 'c͡çʰeː.no t͡sʰoː psʰeː | xoː]
(I did say "ruin the phonology"; this one is a joke.)

[bradrn]

Indications that I did that badly, prompting me to do much more reading on the subject.

Well, perhaps I ought to have just done what I do best — ruin the phonology to get a desired phonetic outcome

Didn’t you already do that?

Phase 5: ['saːstəw 'ais.təj jiut t͡ʃə'wai | naːŋ 'kain.wə t͡sʰau psai | fau]

OK, this is an interesting approach: make various elements head-final so that the verb tends to become head-final as well. (I could probably do some interesting things with subordinate clauses here, not that they’re terribly prominent in the grammar anyway.)

Also, I note that in linking to the previous thread, I suggested an alternate approach: shift to SOV word order via SVCs with ‘take’ (cf. the Mandarin bǎ-construction), but then cliticise it to the previous word rather than the following, to give ergative case-marking rather than accusative. But this still has the problem that I can’t get an applicative marker.

---

I note, however, that this thread has drifted somewhat off my original question: I was originally more interested in getting a general overview of what my options are for going from SVO to SOV word order. Not that this more specific discussion hasn’t been helpful, but it has been somewhat limited.

[Rounin Ryuuji]

Indications that I did that badly, prompting me to do much more reading on the subject.

Well, perhaps I ought to have just done what I do best — ruin the phonology to get a desired phonetic outcome

Didn’t you already do that?

I think that was an improvement.

Phase 5: ['saːstəw 'ais.təj jiut t͡ʃə'wai | naːŋ 'kain.wə t͡sʰau psai | fau]

OK, this is an interesting approach: make various elements head-final so that the verb tends to become head-final as well. (I could probably do some interesting things with subordinate clauses here, not that they’re terribly prominent in the grammar anyway.)

I think it's what I was trying to express before (at least to some extent) — all the elements becoming head-final — but I didn't have the right vocabulary, or train of thoughts, to get it out quite right.

Also, I note that in linking to the previous thread, I suggested an alternate approach: shift to SOV word order via SVCs with ‘take’ (cf. the Mandarin bǎ-construction), but then cliticise it to the previous word rather than the following, to give ergative case-marking rather than accusative. But this still has the problem that I can’t get an applicative marker.

Do you want both of these things? If so, maybe throw in a new auxiliary or particle somewhere?

I note, however, that this thread has drifted somewhat off my original question: I was originally more interested in getting a general overview of what my options are for going from SVO to SOV word order. Not that this more specific discussion hasn’t been helpful, but it has been somewhat limited.

It might help if you play with the sound changes on a larger sample and see what develops.

[bradrn]

Well, perhaps I ought to have just done what I do best — ruin the phonology to get a desired phonetic outcome

Didn’t you already do that?

I think that was an improvement.

Ah, opinion is everything.

Also, I note that in linking to the previous thread, I suggested an alternate approach: shift to SOV word order via SVCs with ‘take’ (cf. the Mandarin bǎ-construction), but then cliticise it to the previous word rather than the following, to give ergative case-marking rather than accusative. But this still has the problem that I can’t get an applicative marker.

Do you want both of these things? If so, maybe throw in a new auxiliary or particle somewhere?

Possibly, but the question is still: where would it come from?

I note, however, that this thread has drifted somewhat off my original question: I was originally more interested in getting a general overview of what my options are for going from SVO to SOV word order. Not that this more specific discussion hasn’t been helpful, but it has been somewhat limited.

It might help if you play with the sound changes on a larger sample and see what develops.

I’ve been doing that all along. Applying sound changes to the first few sentences of my text gives:

Sæs sɔŋm šɔʔle æl sæŋ wæl yel fawle, ‘naʔŋ xæyʔənbəy sɔl pæws?’, aŋsæŋ sæwse læm yošəs. Ŋiyošə welse fawše sæ naʔ wæ aŋ sæŋ sef læm ul, ŋæy ni xæy.

But I’m not quite sure how this helps, given that phonological changes are largely orthogonal to grammatical changes.

(Also keep in mind that I’m still working on the sound changes; in particular I’m not quite sure what to do with some of the more horrible clusters, though these are at least largely absent from the sample. Furthermore it makes it look isolating like the protolanguage, whereas applying the grammar changes would actually make for rather long words.)

[Zju]

That will teach me to trust Wikipedia, then.

I’ve said it before and I’ll say it again: never, ever, ever rely on Wikipedia for linguistic information!

What's your go-to source for linguistic information then?

[vegfarandi]

That will teach me to trust Wikipedia, then.

I’ve said it before and I’ll say it again: never, ever, ever rely on Wikipedia for linguistic information!

What's your go-to source for linguistic information then?

Lately I've been reading a lot of Peter Arkadiev and Martin Haspelmath on Academia.edu and there's some great free stuff at Language Science Press. In some cases, it points to further reading outside of that sphere, often available from SIL. SIL being a Christian organization with intentions I find slightly dubious (linguistics combined with missionary work), it's not always my first stop, but they do have lots of good open source content.

[bradrn]

That will teach me to trust Wikipedia, then.

I’ve said it before and I’ll say it again: never, ever, ever rely on Wikipedia for linguistic information!

What's your go-to source for linguistic information then?

My standard resources are Dixon’s Basic Linguistic Theory, Payne’s Describing Morphosyntax, and ed. Shopen’s Language Typology and Syntactic Description. For specific languages or topics, I prefer finding a reference grammar or typological overview respectively. As vegfarandi says, LangSci Press also has some excellent resources of all types.

[Rounin Ryuuji]

Assuming none of the morphemes have changed in position, I suppose a gloss would look like this?

Sæs sɔŋm šɔʔle æl sæŋ wæl yel fawle, ‘naʔŋ xæyʔənbəy sɔl pæws?’, aŋsæŋ sæwse læm yošəs.
wind DEF.SG=PST follow-CONT sun DEF.SG do.IPFV-CONT argument say.IPFV-CONT, "Who strong-NMLZ=Q leave all?" person=MIR have.PFV-DIM cloak come.PFV-DIM

sæs = "wind"
sɔŋ-m = an umlauted form of the definite article (probably showing initial [tʰ] > [θ] > [s̪] > [s]), the vowel mutated by the lost historic /u/ a past-tense marker
šɔʔ-le = "and", but it functions verbally; presumably showing sound changes [t͡ɬ] > [ɬ] > [ʃ], deletion of an intertonic syllable (but not before) /u/ could umlaut the /a/ | suffixed with a continuative marker
æl = "sun"
sæŋ = the uninflected form of the definite article
wæl = appears to be a much-reduced form of /waqli/ "do" (continuative), showing loss of final /i/ (I suspect posttonic vowels and codas were regularly deleted), and the glottal stop (probably conditioned by another coda consonant) without compensatory lengthening
yel = "argument" (the language does not appear to have an indefinite article)
fawle = presumably from faw(etl)li; I suspect this terminal /e/ is the result of some sort of secondary stress falling on inflectional syllables in trisyllabic words) "say" (also continuative)
naʔŋ = "who"
xæyʔ-ən-bəy = "strong" | a nominaliser (in this case, probably changing the meaning to "stronger person, stronger thing") | a question word, potentially shifting the meaning of naʔŋ to "who?", though it may also be a pleonasm)
sɔl pæws = I expect this is a fixed expression, meaning "exceeding all (others)"?
aŋsæŋ = I expect aŋ means "person" and that sæŋ is a mirative suffix, expressing surprise at the person being there;
sæwse = In context, certainly meaning "having, who had", presumably sæw "have, possess" + -se, a perfective marker
læm = "cloak"
yošəs = presumably from a verb form "yošə" + -s, an allomorph of -se that probably appears after longer morphemes.

But I’m not quite sure how this helps, given that phonological changes are largely orthogonal to grammatical changes.

I don't think this is quite true. Sound change often forces structural change through erosion of morphological elements, as it did to the Latin case system. It wasn't a foregone conclusion that it would do this, but it was also not an unlikely outcome. What sound changes occur can consequently very much affect what your grammatical outcome is. You're more likely to replace one marker with another if that other can be more distinctly understood. You might also expect certain elements to be articulated in a single breath:

[Sæs sɔŋm] šɔʔle [æl sæŋ] [wæl yel] fawle, ‘naʔŋ xæyʔənbəy [sɔl pæws]?’, aŋsæŋ sæwse læm yošəs.

I would expect all of these to stick together, with the articles becoming fused case markers — Sæs sɔŋm > Sæsɔŋ (the sequence [ŋm] is rather difficult to articulate, so I think the [m] would drop), æl sæŋ > ælæŋ — sɔl pæws to become a fixed expression sɔpæws "be something above all" (or simply "above all", with a copula being inserted into sentences that contain it), and wæl yel to do something similar, though I'm not exactly sure how that last one would turn out, since I think wæl is the conjugable element.

With the erosion that's happened already, I expect the language has a tendency to put a rather strong emphasis on the main element of a cluster of morphemes, which tends to cause structural words to be deemphasised, and consequently more susceptible to erosion: as discussed before, šɔʔle > šɔl, and fawle > faw do now strike me as extremely likely developments, and you're only a tendency to parallelism of structure away from Sæs sɔŋm šɔl æl sæŋ šɔl wæl yel, which could then shift to Sæs sɔŋm šɔl æl sæŋ šɔwæl yel. If more words undergo semantic bleaching similarly, you end up with more markers. Looking for points of ambiguity and inserting more structural words to resolve them, then eroding away the word till it's just a little marker that can be reanalysed further, and so on and so on.

[Nachtswalbe]

How do you get poly synthesis out of an isolating language like English e.g poly personal agreement?

[bradrn]

Assuming none of the morphemes have changed in position, I suppose a gloss would look like this?

Yes, correct — as I said, no grammatical changes have been applied.

sɔŋ-m = an umlauted form of the definite article (probably showing initial [tʰ] > [θ] > [s̪] > [s]), the vowel mutated by the lost historic /u/ a past-tense marker

Correct. Though the past tense clitic really ends up as an affix on the verb, so this would actually become un-umlauted sæŋ.

(I suspect posttonic vowels and codas were regularly deleted)

Partially correct: IIRC I wasn’t deleting posttonic codas specifically, though this is true for vowels. (See François for details.)

yel = "argument" (the language does not appear to have an indefinite article)

The singular indefinite article is indeed null, but it would perhaps be better to analyse this as a coverb construction, in which a ‘light verb’ is complemented by a ‘coverb’. (Not converb! The two are entirely different!)

fawle = presumably from faw(etl)li; I suspect this terminal /e/ is the result of some sort of secondary stress falling on inflectional syllables in trisyllabic words) "say" (also continuative)

Well, it’s specifically posttonic unstressed vowel deletion. No matter whether the final vowel is stressed or not (I haven’t decided yet), it’s not posttonic.

xæyʔ-ən-bəy = "strong" | a nominaliser (in this case, probably changing the meaning to "stronger person, stronger thing") | a question word, potentially shifting the meaning of naʔŋ to "who?", though it may also be a pleonasm)

Not quite. *khayiq means ‘strong thing’ anyway; the nominaliser forms the abstract noun ‘strength’. The question clitic is indeed pleonasm. (And thanks for introducing me to this term.) This part will probably end up entirely different in the daughter language.

sɔl pæws = I expect this is a fixed expression, meaning "exceeding all (others)"?

I wouldn’t call it ‘fixed’ — it’s perfectly transparent. *naqeŋ khayiqene is inalienable expression: ‘the strength of who’. *tshaal is ‘leave, exceed’: the normal verb used in forming comparative constructions. So *naqeŋ khayiqenebey tshaal pawtiq is literally ‘the strength of who exceeds all?’.

aŋsæŋ = I expect aŋ means "person" and that sæŋ is a mirative suffix, expressing surprise at the person being there;

Again, correct.

sæwse = In context, certainly meaning "having, who had", presumably sæw "have, possess" + -se, a perfective marker

Basically, yes, though I wouldn’t translate it as subordinate: ‘The person has a cloak comes’ would be better, if ungrammatical in English. (This is another SVC.)

yošəs = presumably from a verb form "yošə" + -s, an allomorph of -se that probably appears after longer morphemes.

Just posttonic vowel deletion again: for now I’m stressing it *ˈyusaˌyetsi, so the final vowel gets deleted.

But I’m not quite sure how this helps, given that phonological changes are largely orthogonal to grammatical changes.

I don't think this is quite true. Sound change often forces structural change through erosion of morphological elements, as it did to the Latin case system. It wasn't a foregone conclusion that it would do this, but it was also not an unlikely outcome. What sound changes occur can consequently very much affect what your grammatical outcome is. You're more likely to replace one marker with another if that other can be more distinctly understood. You might also expect certain elements to be articulated in a single breath:

[Sæs sɔŋm] šɔʔle [æl sæŋ] [wæl yel] fawle, ‘naʔŋ xæyʔənbəy [sɔl pæws]?’, aŋsæŋ sæwse læm yošəs.

This is of course correct. Last time I thought about the previous iteration of these sound changes, I suggested the following outcome:

Sæssæŋ šɔ Ælsæŋ muŋiwæyelle faw, naŋ exæyənə pæws bəyesɔl, aŋseŋæ esæwyošətse læm.

The major development I haven’t mentioned yet is that the verbs all got subject cross-referencing. Also, as you predicted the coverb got absorbed into its head. But I seem to have muddled up the word order a bit: the second clause has ended up as SOV while the last is still SVO.

That being said, none of this is really relevant to word order, which is what I was trying to get at with my original statement.

[Rounin Ryuuji]

Sæssæŋ šɔ Ælsæŋ muŋiwæyelle faw, naŋ exæyənə pæws bəyesɔl, aŋseŋæ esæwyošətse læm.

The major development I haven’t mentioned yet is that the verbs all got subject cross-referencing. Also, as you predicted the coverb got absorbed into its head. But I seem to have muddled up the word order a bit: the second clause has ended up as SOV while the last is still SVO.

That being said, none of this is really relevant to word order, which is what I was trying to get at with my original statement.

I think I was getting at forcing structural change by breaking what grammar you already have through sound change, but you have some very resilient structure going on here. In the course of reading more on the subject, I've discovered that patterns of intonation are thought to sometimes affect word order — people might rearrange words based on where it's most comfortable to say them.

It also seems to be that Subject-Verb-Object is the most stable word order in human language, and that changing from it to Subject-Object-Verb is poorly-attested (even though the other way round is extremely common). It looks like the answer to the question of what your options are in getting to that, though you clearly can, by internal processes, would be "not very much". Languages with the desired structure tending to be postpositional and to put adjectives after the noun has been noted.

[bradrn]

In the course of reading more on the subject, I've discovered that patterns of intonation are thought to sometimes affect word order — people might rearrange words based on where it's most comfortable to say them.

This sounds interesting. Do you have any links?

It also seems to be that Subject-Verb-Object is the most stable word order in human language, and that changing from it to Subject-Object-Verb is poorly-attested (even though the other way round is extremely common).

Again, I’d be interested in a source for this. It does accord with my understanding, though: I’m aware of several examples of SOV→SVO, and at least one each of SOV→VSO and VSO→SVO, but none of SOV→SVO.

It looks like the answer to the question of what your options are in getting to that, though you clearly can, by internal processes, would be "not very much". Languages with the desired structure tending to be postpositional and to put adjectives after the noun has been noted.

Well, this language does at least put ‘adjectives’ (really, a subclass of nouns) after the head. It actually doesn’t have any adpositions, but their replacement (SVCs, mostly) have a distinctly prepositional flavour. The only really unstable part of this language is the fact that it has nouny adjectives but non-tensed verbs, which is why I predict the past clitic will end up becoming a true verbal tense affix.

[Rounin Ryuuji]

In the course of reading more on the subject, I've discovered that patterns of intonation are thought to sometimes affect word order — people might rearrange words based on where it's most comfortable to say them.

This sounds interesting. Do you have any links?

I actually tried finding it earlier today, but couldn't. It was in something about the loss of V2 order in the transition between Late Middle and Early Modern English.

It also seems to be that Subject-Verb-Object is the most stable word order in human language, and that changing from it to Subject-Object-Verb is poorly-attested (even though the other way round is extremely common).

Again, I’d be interested in a source for this. It does accord with my understanding, though: I’m aware of several examples of SOV→SVO, and at least one each of SOV→VSO and VSO→SVO, but none of SOV→SVO.

[url=file:///C:/Users/akiya/AppData/Local/Temp/scriptie_compleet_laatsteversie-1.pdf]I think it was this[/url]. I take some of the ideas it contains with a bit of salt, but the tendency to shift to Subject-Verb-Object unless there's some major impediment (like the Japanese adnominal use of verbs) seems very widespread. Granted, I think we also aren't presently observing the emergence of a complex case system at the moment. Classical Latin was Subject-Object-Verb and also had heavy nominal inflection, and the parts of the Romance languages that still are are the inflected pronominals. I think there might be more data on shifts to Subject-Verb-Object if we were able to observe the genesis of a similar case system.

It looks like the answer to the question of what your options are in getting to that, though you clearly can, by internal processes, would be "not very much". Languages with the desired structure tending to be postpositional and to put adjectives after the noun has been noted.

Well, this language does at least put ‘adjectives’ (really, a subclass of nouns) after the head. It actually doesn’t have any adpositions, but their replacement (SVCs, mostly) have a distinctly prepositional flavour. The only really unstable part of this language is the fact that it has nouny adjectives but non-tensed verbs, which is why I predict the past clitic will end up becoming a true verbal tense affix.

[/quote]
Verbs developing into adpositions is scarcely unheard-of, so I expect this language would probably do it to some extent, based on what I've seen.

[Zju]

I’ve said it before and I’ll say it again: never, ever, ever rely on Wikipedia for linguistic information!

What's your go-to source for linguistic information then?

My standard resources are Dixon’s Basic Linguistic Theory, Payne’s Describing Morphosyntax, and ed. Shopen’s Language Typology and Syntactic Description. For specific languages or topics, I prefer finding a reference grammar or typological overview respectively. As vegfarandi says, LangSci Press also has some excellent resources of all types.

Thank you for your input, but a paper book reply is what I secretly feared. Maybe I'll start using the online sources instead.

[bradrn]

I think it was this.

That’s a link to a file stored locally on your computer. (And in Dutch to boot, it looks.)

I take some of the ideas it contains with a bit of salt, but the tendency to shift to Subject-Verb-Object unless there's some major impediment (like the Japanese adnominal use of verbs) seems very widespread.

Are you sure? Plenty of languages around the world show no signs whatsoever of switching to SVO word order, including nearly all of Asia (except in the SE), most of North America, large swathes of South America, and much of New Guinea and Australia. The only families which really seem to have undergone a shift SOV→SVO en masse seem to be IE, Niger–Congo (to the extent it exists) and the various SE Asian families.

Verbs developing into adpositions is scarcely unheard-of, so I expect this language would probably do it to some extent, based on what I've seen.

Of course, and I expect some of these ones would absolutely end up as prepositions.

Thank you for your input, but a paper book reply is what I secretly feared. Maybe I'll start using the online sources instead.

Oh, I don’t use the paper copies! For the first two, I use the copies distributed with the Stack. The last one is harder to track down, but I don’t have a hardcopy of that either.

[Rounin Ryuuji]

That’s a link to a file stored locally on your computer. (And in Dutch to boot, it looks.)

Oops; I've ended up closing out the tab, but I'll dig it up again in a bit.

I take some of the ideas it contains with a bit of salt, but the tendency to shift to Subject-Verb-Object [from Subject-Object-Verb] unless there's some major impediment (like the Japanese adnominal use of verbs) seems very widespread.

Are you sure? Plenty of languages around the world show no signs whatsoever of switching to SVO word order, including nearly all of Asia (except in the SE), most of North America, large swathes of South America, and much of New Guinea and Australia. The only families which really seem to have undergone a shift SOV→SVO en masse seem to be IE, Niger–Congo (to the extent it exists) and the various SE Asian families.

Missed out some important words.

Edit: Also, I thought Mandarin was Subject-Object-Verb by default?

[bradrn]

I take some of the ideas it contains with a bit of salt, but the tendency to shift to Subject-Verb-Object [from Subject-Object-Verb] unless there's some major impediment (like the Japanese adnominal use of verbs) seems very widespread.

Are you sure? Plenty of languages around the world show no signs whatsoever of switching to SVO word order, including nearly all of Asia (except in the SE), most of North America, large swathes of South America, and much of New Guinea and Australia. The only families which really seem to have undergone a shift SOV→SVO en masse seem to be IE, Niger–Congo (to the extent it exists) and the various SE Asian families.

Missed out some important words.

I did understand that you were talking about a shift from SOV word order specifically. All the areas I mentioned have predominantly SOV word order.

Edit: Also, I thought Mandarin was Subject-Object-Verb by default?

What, Mandarin‽ It does show some tendencies towards SOV word order (namely the bǎ-construction I mentioned earlier), but for the most part it is quite straightforwardly SVO.

[Rounin Ryuuji]

Edit: Also, I thought Mandarin was Subject-Object-Verb by default?

What, Mandarin‽ It does show some tendencies towards SOV word order (namely the bǎ-construction I mentioned earlier), but for the most part it is quite straightforwardly SVO.

Oops, typed it backwards. I thought you were excluding most of Asia in the not shifting to SVO, but Sinitic seems to be there (Old Chinese was also apparently Subject-Verb-Object — I think my brain just likes typing the sequence "subject-object-verb" for some reason); I could, incidentally, see Sinitic developing into an inflected language; it would be interesting to see what sort of Sinitic we have in 2500.

[Darren]

How do you get poly synthesis out of an isolating language like English e.g poly personal agreement?

Check out this thread from the old board, it's got a whole section on how polysynthesis can arise.