(L&N) Q&A Thread - Quick questions go here

[Salmoneus]

I've recently stumbled across the Gutenberg Project release of Chronicle and Romance: Froissart, Malory, Holinshed, by Lord John Bourchier Berners et al.

The original documents date back to the 15th century, so written roughly at the change from Late Middle English to Early Modern English, but this particular edition is more recent, and has been 'translated' into a more recent orthography. However, there's a particular word that I'm not sure about, in the first paragraph of the section "OF THE GREAT ASSEMBLY THAT THE FRENCH KING MADE TO RESIST THE KING OF ENGLAND":

The French king heard well what he did, and sware and said how they should siever return again unfought withal, and that such hurts and damages as they had done should be dearly revenged;[...]

Is that supposed to be "sever", with the intended meaning of "abandon"/"give up on"? I.e. "they should give up on the hope of going unfought if they were to return again"?

I'm no expert in late middle or early modern english, but my strong intuition on reading that sentence is that it's a typo for "soever". And then on top of that, I think it's using a positive indefinite in a negative sense - i.e. there's an implied negative - which I have a feeling is something that people did (and of course certainly did in French), but I can't actually provide any examples or expert support for that.

So I think it's saying "He heard about what had happened, and promised that they should [not] under any circumstances at all return again without being fought against".

[today, "soever" is usually found in compounds with an interrogative - in this case, howsoever. But originally it followed the wh-clause it modified - so it's licenced in some way by the "how", even though it's not adjacent to it. ]

[Nel Fie]

Thank you all for taking the time to answer, I hope you will not think ill of me for not responding at length to each of your thoughts, for I found the answer to the riddle, and it is in fact another thing altogether.

Salmoneus was on the right track: "siever" is a typo indeed, but not for "soever". The word that should be written is, in fact, "never".

I had a growing suspicion after reading your previous messages, Dormouse559 and Khemehekis, along with having already encountered another minor fault in a previous passage of the book. Gutenberg produces much of their texts through OCR, and misreads steal themselves into the resulting document from time to time, and past the vigilant eye of whichever proofreaders there might be. I sought out a scan of the book as it was printed and found one such on archive.org. As you may see for yourselves on the page linked, the text doth read:

The French king heard well what he did, and sware and said how they should never return again unfought withal, and that such hurts and damages as they had done should be dearly revenged;[...]

Sīc semper māchinīs.

[Davush]

Does anyone have any thoughts/ideas on why, in languages that use auxiliary verbs, the order V-O-Aux never seems to appear as a 'default' order, and when alternative orders are allowed, it is exceedingly rare? (Apparently Classical Latin allowed this order, but it was not common, and eventually disappeared*).

Perhaps it is just an accident of the available data, and this order did occur in the past, BUT languages with relatively free word order (e.g., Finnish, Basque) seem to also forbid/strongly disfavour this order, which would seem to indicate there is something fundamentally 'wrong' with it.

Now, to go all syntax-y on it, apparently this would fall foul of the 'Final-over-Final Constraint' – I understand this to basically be a rule whereby a head-final phrase doesn't permit a head-initial phrase within it, i.e., [V-O]-Aux is head-final if we take Aux as the head (being the 'syntactic' verb), but [V-O] is head-initial. However, this doesn't explain why the reverse is found! Aux-[O-V] is head-initial, but contains the head-final [O-V] and is very common all over Germanic (I have the book bought)!

Head directionality can't be the only reason, since many (if not most?) languages are not extremely rigid one way or another.

One explanation I have read is that OV languages moving to VO word order would go like this:

book buy FUT COMP
COMP book buy FUT
COMP FUT book buy
COMP FUT buy book

The change VO>OV is apparently the same, just starting from the bottom instead.

If we take FUT as being marked by Aux, then V-O-Aux wouldn't occur in either direction (but then this is quite abstract and I'm not sure if it actually holds up)...hmm!

Any thoughts welcome!

---

*Maybe Portuguese its forms such as cantar-te-ei (sing-to.you-I.will) could be analysed as V-O-Aux, but I imagine syntacticians would argue the O here is a pronominal clitic/infix, which is 'special' so doesn't count, or that the auxiliary isn't really an auxiliary (even though historically, at least, it ostensibly is...?)

[Arayaz]

Does anyone have any thoughts/ideas on why, in languages that use auxiliary verbs, the order V-O-Aux never seems to appear as a 'default' order, and when alternative orders are allowed, it is exceedingly rare? (Apparently Classical Latin allowed this order, but it was not common, and eventually disappeared*).

A large majority of auxiliary verbs evolve from regular verbs wherein the modified VP is the object of the auxiliary. Thus in an OV language, the structure is usually O-V-Aux, and in a VO language, it is usually Aux-V-O. (In a VSO language, Aux-S-V-O would be the ordering as a result of this ─ if the auxiliary is subsequently eroded to nothing, inverting the subject and verb signals a tense/aspect/mood shift here!)

[Creyeditor]

AFAIK, the FOFC is supposed to be asymetric, so [V-O]-Aux is banned but Aux-[OV] is okay, contrary to what the name suggests. I hasten to add that this in itself does not explain anything and could only be a first step on the way to a potential explanation.

[Davush]

AFAIK, the FOFC is supposed to be asymetric, so [V-O]-Aux is banned but Aux-[OV] is okay, contrary to what the name suggests. I hasten to add that this in itself does not explain anything and could only be a first step on the way to a potential explanation.

Edit: I had misunderstood (apologies!). BUT, this actually raises further questions! Why is a head-initial phrase contained within a head-final phrase banned (or strongly disfavoured), but not the reverse? What is it about this particular structure that is so 'disconcerting''? The syntactic change pathway explanation is quite nice, as the 'banned' word order supposedly does not diachronically ever appear, but does it hold up..?

A large majority of auxiliary verbs evolve from regular verbs wherein the modified VP is the object of the auxiliary. Thus in an OV language, the structure is usually O-V-Aux, and in a VO language, it is usually Aux-V-O. (In a VSO language, Aux-S-V-O would be the ordering as a result of this ─ if the auxiliary is subsequently eroded to nothing, inverting the subject and verb signals a tense/aspect/mood shift here!)

I understand that, but it does not explain the discrepancy why Aux-O-V is common, yet its reverse V-O-Aux is not.

[Arayaz]

A large majority of auxiliary verbs evolve from regular verbs wherein the modified VP is the object of the auxiliary. Thus in an OV language, the structure is usually O-V-Aux, and in a VO language, it is usually Aux-V-O. (In a VSO language, Aux-S-V-O would be the ordering as a result of this ─ if the auxiliary is subsequently eroded to nothing, inverting the subject and verb signals a tense/aspect/mood shift here!)

I understand that, but it does not explain the discrepancy why Aux-O-V is common, yet its reverse V-O-Aux is not.

I know; I don't have an answer for the overall question, but I thought I could offer some preliminary insights.

[Davush]

I know; I don't have an answer for the overall question, but I thought I could offer some preliminary insights.

You did, thank you. Syntax is not often an area I get into, because my questions tend towards the 'why' (a lot of which are probably unanswerable, but interesting to speculate on nonetheless)...

[Arayaz]

I know; I don't have an answer for the overall question, but I thought I could offer some preliminary insights.

You did, thank you. Syntax is not often an area I get into, because my questions tend towards the 'why' (a lot of which are probably unanswerable, but interesting to speculate on nonetheless)...

One intriguing thing is that in VO languages, V-O-Adverb is usually the favored word order when adverbs are involved ─ so if an adverb were to lexicalize into an auxiliary, it would result in V-O-Auxiliary ─ yet this occurs rarely in natlangs, so there must be something else at work! Perhaps it is akin to the thing about prepositions hardly ever becoming noun case, as opposed to postpositions doing it often ─ that is to say, annoying for conlangers!

In fact, depending on the analysis, Ngama might sometimes allow V-O-Auxiliary, for that exact reason ─ but rather than auxiliaries, they're probably best described as clitics.

[Davush]

One intriguing thing is that in VO languages, V-O-Adverb is usually the favored word order when adverbs are involved ─ so if an adverb were to lexicalize into an auxiliary, it would result in V-O-Auxiliary ─ yet this occurs rarely in natlangs, so there must be something else at work!

[...]

Since auxiliaries are usually the syntactic verb (i.e., they carry the grammatical information that a non-auxiliary would usually be inflected for), having them come from adverbs does not seem that likely, unless the adverb was first somehow verbalised, or gained verbal-like inflection (not impossible, but it just doesn't seem like a likely pathway when languages already have plenty of 'actual' verbs).

Re Ngama: There are probably some syntax tests to determine if it is more clitic-like or more aux (verb)-like...

The pathway change explanation is the only thing I've seen so far that addresses the asymmetry, in that the order V-O-Aux can't arise via the proposed pathway of change. However, if this holds empirically isn't certain...

[Arayaz]

Since auxiliaries are usually the syntactic verb (i.e., they carry the grammatical information that a non-auxiliary would usually be inflected for), having them come from adverbs does not seem that likely, unless the adverb was first somehow verbalised, or gained verbal-like inflection (not impossible, but it just doesn't seem like a likely pathway when languages already have plenty of 'actual' verbs).

So embarrassing, I forgot about that bit...

[LinguoFranco]

So, I read somewhere that Persian has the vowels /æ ɒː e i oː uː/. Why are some of them long and some short? Most languages that have phonemic vowel length tend to have them come in pairs like /a aː i iː u uː/, while in Persian, it seems that back vowels are long, and front ones are short.

Are there any other examples of natlangs with this feature?

[Arayaz]

So, I read somewhere that Persian has the vowels /æ ɒː e i oː uː/. Why are some of them long and some short? Most languages that have phonemic vowel length tend to have them come in pairs like /a aː i iː u uː/, while in Persian, it seems that back vowels are long, and front ones are short.

Are there any other examples of natlangs with this feature?

Tehrani Persian has /æ ɒː e(ː) iː o(ː) uː/ according to Wikipedia ─ long /ɒ/ and high vowels, and optionally mid vowels. Other dialects have more extensive systems and mostly have length contrasts for all vowels. But I'm not an expert on Persian, and I'm not sure of any other languages with a similar system.

[Flavia]

So, I read somewhere that Persian has the vowels /æ ɒː e i oː uː/. Why are some of them long and some short? Most languages that have phonemic vowel length tend to have them come in pairs like /a aː i iː u uː/, while in Persian, it seems that back vowels are long, and front ones are short.

Are there any other examples of natlangs with this feature?

I learned a little bit of Persian a few years ago; my teacher used a transcription system that had a i u for /æ e o/ and ā ī ū for /ɒː iː uː/. I can't quite remember what dialect it was though. I think you misremembered — its either /i: u:/ paired with /e o/, or as Arayaz wrote, with length contrast in mid vowels too. It doesn't seem that unusual to me, it's just a three-vowel system with length contrast, with the short vowels being reduced somewhat.

[Salmoneus]

Yep - well, I don't know if "reduced" is the right word because I don't think the short vowel are centralised, but yes, it's a 3-vowel system with length, historically, but where a quality distinction has been introduced by the long and short vowels shifting slightly. The long mid vowels in some dialects, iirc, arise from diphthongs, and mean that some dialects have (or had) 3 short vowels but 5 long vowels, which is rare and fun.

IIRC, in at least some dialects the 'long' vowels aren't actually consistently any longer than the short vowels now - it's become a qualitative system, with quantity being secondary, as in English and Romance (at different stages of the process, obviously).

[Davush]

Related to my recent question about word order, I have another one.

Supposing it is true that the shift from SOV starts "top down", what would this look like in practice?

Let's use Japanese as an example since it is rigidly SOV/head final. According to the proposed pathway of change, the order [OV]I]C] would start by "C" (which would be a complementizer) moving to the front. However, in Japanese, complementizers are usually particles that can't be separated from the verb/phrase, so I assume this means that a "new" pre-verbal complementizer would have to be generated? Where might this come from? Does this pathway actually seem likely? OR does it means that since the complementizer (in Japanese) is attached to the verb it would simply pull the entire Verb with it to the pre-O position?

Ignoring contact-related word order shifts for now, I'd be interested to hear any ideas on what (internal) shifts could cause strongly OV to move in the direction of VO...?

[Salmoneus]

I'm afraid I don't know any syntax (in a linguistics sense, imaginary trees and brackets and anagrams and whatnot), so I probably can't give the sort of answer you want.

[from my position of ignorance, your business with the complementizers seems counterintuitive, since at least in Germanic and I believe in various other languages as well the fronting of verbs occurs primarily and firstly in main clauses, with no complementizers-in-a-non-linguistics-sense, and subordinate clauses that do have complementizers-in-a-non-lingusitics-sense retain V-final position. But I don't know, I suppose that in linguistics "complementizers" are invisible and occur in all clauses or something like that]

But from a language (rather than linguistics) point of view, the main shift that would lead from OV to VO would be fronting of the verb to a pre-O position, although backing of the object to a post-V position (perhaps via a comment-topic-like structure: "I the cat hit" > "I (it) hit, the cat" > "I hit the cat") would also be possible.

In the case of Germanic, I believe the suspicion is that this is probably underlyingly at least partially influenced by the significance of Wackernagel's Position in Proto-Indo-European.

In PIE and (to varying extents) in its daughter languages, there is a strong tendency for weak particles to be attracted to second position, where they were probably (at least in stress terms) dependent upon the word in first position as clitics. This can result in strings of particles/clitics. In some cases it can be so strong a tendency that the particles can be inserted in post-initial position even when that means breaking up a phrase, or even breaking up what seems to us to be a compound word.

From this, we can suggest at least four motivations for verb-fronting (and specifically V2) (I'm no expert but I think these are in decreasing order of scholarly favour?):

- auxiliary verbs came to be destressed, which made them look like particles, so they were moved to Wackernagel's Position (WP). Subsequently, other finite verbs were moved to WP by analogy.

- particles in WP often had modal/aspectual/temporal significance (although not always - they also originally included pronouns, for instance). So the rule that WP contained TAM information was generalised, which meant moving finite (but not non-finite!) verbs to WP, since these verbs contained TAM information.

- the particles in WP could include "adverbs" and "prepositions" that were strongly semantically bonded to the verb, so may have attracted the verb to join them in WP. In the case of "prepositions" this involved compounding. Normally I think we assume that prepositional verbs were formed by verb-movement taking the original postposition with them ["I the field-over flew" > "I over-flew the field"], but postpositions could also be weak particles in WP originally, so maybe they moved first and brought the verb with them rather than vice versa?

- the tendency for objects to be indicated by weak pronouns in WP may have lead to WP being a weak position for objects. That is, in theory there would have been a distinction between weak objects in a S(O)V order, where (O) indicates an object referred to by a weak pronoun in WP, and strong objects in a SOV order, where O indicates an object referred to by a noun or strong pronoun... but obviously in practice, in simple sentences with no other elements or syntactic confusions, these two word orders are the same. So maybe in order to maintain a distinction between strong and weak objects, strong objects were moved out of that position, perhaps via a comment-topic structure [S(O)VO, with (O) then being dropped].


What I've seen of the linguistics about this is unfathomable to me and allegedly extremely complicated, with T and T-with-superscripts and C and SpecT and EPPs and who knows what else, and nobody can agree even on what's "happening" in any one old germanic language, let alone how it led to language change across the family, and since this whole debate goes on without ever needing to refer to any actual linguistic evidence (which we don't really have much of) I've never felt it worth diving into.

[Davush]

I'm afraid I don't know any syntax (in a linguistics sense, imaginary trees and brackets and anagrams and whatnot), so I probably can't give the sort of answer you want.

Thank you very much for this! To be honest, my knowledge of syntax/"linguistics" is very rudimentary, and I favour common-sense explanations over things like "Platonic complementisers" where possible, it's just that most of the literature seems to be in this vain and it's hard to find discussions in more concrete (i.e. language, rather than "linguistics") terms, so this is immensely more helpful!

I suppose a related question, then, would be what else could cause fronting of the V, since the PIE-Wagernackel thing seems very language-specific? Are there any examples where clefting has become the 'default' and thereby changed the default word order? To use Japanese as an example:

Default SOV: (watashi-ha) tabemono-wo tabeta (food-ACC eat)

With clefting: (watashi-ga) tabeta-no-ha tabemono desu (eat-NMLZ-TOP food COP) = 'What I ate was food / It was food that I ate'

(I suppose the clefting example is just a specific example of how a V may be fronted, but would it then decaying into a 'default order' be something that could (or has) happened...?)

[Salmoneus]

I'm afraid I don't know any syntax (in a linguistics sense, imaginary trees and brackets and anagrams and whatnot), so I probably can't give the sort of answer you want.

Thank you very much for this! To be honest, my knowledge of syntax/"linguistics" is very rudimentary, and I favour common-sense explanations over things like "Platonic complementisers" where possible, it's just that most of the literature seems to be in this vain and it's hard to find discussions in more concrete (i.e. language, rather than "linguistics") terms, so this is immensely more helpful!

I suppose a related question, then, would be what else could cause fronting of the V, since the PIE-Wagernackel thing seems very language-specific?

Well, I wasn't being sarcastic before: maybe the V just moves forward, before the O. Why not?

Many languages have multiple possible word orders, or even free word order. In such a language, one word order can become more common while another becomes less common. Reasons of topicalisation, focusing and emphasis are probably involved, but any initial purpose becomes bleached by analogy.

To give a concrete example: Old Latin was SOV. Romance languages are mostly SVO. At some point, Latin's SOV word order became a suggestion rather than a rule - SOV was a tendency, but word order was almost free in theory. At some late point it became a suggestion more honoured in the breach than in the observance - vulgar latin seems to have had SVO as its default. And at some later point in the various languages that breach became the new rule, and the old observance became ungrammatical - SVO became the rule, at least in main clauses with non-pronoun objects.

Having said that, there may also be one typological nudge in Latin's case: its prepositions and tendency for post-nominal adjectives conflicted with its tendency for final verbs, and moving to SVO resolved that. But that of course is just shifting the question (why did its mixed order originally arise?).

Are there any examples where clefting has become the 'default' and thereby changed the default word order?

I'd have thought so, surely. Why not? Reanalysis is something that happens all the time. For instance, Tibetan relative clauses apparently (or at least most of them?) used to be (really big and complicated) compound nouns...

[Davush]

Having said that, there may also be one typological nudge in Latin's case: its prepositions and tendency for post-nominal adjectives conflicted with its tendency for final verbs, and moving to SVO resolved that. But that of course is just shifting the question (why did its mixed order originally arise?).

Thanks! Yes, the Latin example does beg the question of how/why word order became 'free' in the first place, but I know these are questions that are probably ultimately unanswerable...

Reanalysis is also a good thing to keep in mind more generally wrt language change (and for conlanging purposes).