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

[Salmoneus]

Not an expert, but since we've raised the question, I'd guess there are three reasons why Latin developed mostly-free word order:

a) the word order was probably ancestrally never totally fixed anyway. Strong SOV tendencies in the earliest records may be in part a literary decision, a style of "correct Latin" that intentionally avoided 'sloppy', less-common orders. [there clearly was SOME fixation in this era because all Romans agreed that prepositions preceded nouns, which previously they hadn't, but the rest may have been variable in actual speech?]. So maybe later, freer word orders are a literary decision to abandon artificial rules and reflect speech more naturally

b) freer word orders may be a literary decision based on emulation of Greek in later centuries, as Greek has very mobile words too

c) freer word order may simply reflect a loss of redundancy: fixed word order was unnecessary in such an obsessively inflected language, so was not retained. A new fixed word order was imposed centuries later, when the loss of inflections in colloquial speech had made syntactic ambiguity too problematic.

[Davush]

Not an expert, but since we've raised the question, I'd guess there are three reasons why Latin developed mostly-free word order:

a) the word order was probably ancestrally never totally fixed anyway. Strong SOV tendencies in the earliest records may be in part a literary decision, a style of "correct Latin" that intentionally avoided 'sloppy', less-common orders. [there clearly was SOME fixation in this era because all Romans agreed that prepositions preceded nouns, which previously they hadn't, but the rest may have been variable in actual speech?]. So maybe later, freer word orders are a literary decision to abandon artificial rules and reflect speech more naturally

b) freer word orders may be a literary decision based on emulation of Greek in later centuries, as Greek has very mobile words too

c) freer word order may simply reflect a loss of redundancy: fixed word order was unnecessary in such an obsessively inflected language, so was not retained. A new fixed word order was imposed centuries later, when the loss of inflections in colloquial speech had made syntactic ambiguity too problematic.

Interesting! I suppose the variable placement of prepositions goes back to adpositions being more adverby in PIE?

---

An unrelated question:

Are there any languages where a two-way noun class system is not tied to gender at all? Extending this further, is it plausible to have a noun-class system that is not actually tied to semantic content at all and is purely formal? By this I mean that demonstratives/adjs/verbs/etc. show agreement with the noun class, perhaps based on the phonological shape of the word rather than groups of meaning/'natural' categories (male/female, animate/inanimate, mass/count, etc.) It seems that the vast majority of languages that have a two-way system, have gender (or animacy) involved. Maybe this is just a natural outcome since people like to categorise and having such a system with no connection to 'real-world' categories would make it decay quite quickly...?

[Creyeditor]

There are also binary gender systems based on shape (usually sphere-like vs. stick-like) and based on animacy. The thing is, of you have a purely formal system that assigns different genders to very frequently used nouns (which makes sense in terms of frequency) than it is likely that either man and woman will be in different classes - and therefore interpreted as sex-based (by linguists and speakers) - or frequent tools or animals will be in different classes - and therefore it will be interpreted as animacy-based. The former happened to many shape-based systems actually, where woman are classed as sphere-like and men are classed as stick-like.

[Visions1]

Is this a good place to mention Dyrbal? I haven't looked into it (it really seems like people are crazy over it. Are its features really so strange to everywhere and Australia?)

[Knox Adjacent]

Other than the rarity of ergativity extending to syntax, no. The gender system isn't as exciting as some people make it out to be, c.f. K Plaster 2007 ("Women aren't dangerous things: gender and categorization"). Otherwise it's a well-behaved eastern Pama-Nyungan language.

[Nel Fie]

A current curiosity of mine: for how many grammatical functions do natural languages employ word order? And what are those functions?

Though I'd wager the answer will probably be 'everything and anything', so perhaps a more specific question - among the languages anyone here might know about, which natural language concurrently employs word order for more grammatical functions than any other? And would the quantity of uses it makes of word order be unusual compared to most languages?

[Creyeditor]

I don't think there is an easy asnwer but here are some functions that I know of.

1) The difference between the two arguments of transitive verbs. Happens in English (and probably many other languages).
2) Information structure, so things like topic and focus. Happens in many languages including English, cf. Beans, I like.
3) Something that happens in noun adjective order but I forgot what it is. Happens in some Romance language, IIRC.

I don't think anything and everything is the correct answer. I know of no language that uses only word order to distinguish person, number, tense, aspect, modality, or evidentiality. But I might be wrong of course. I think, I had the idea of using word order in unusual ways for Pansi here (https://cbbforum.com/viewtopic.php?p=322245#p322243).

[Davush]

Another of my 'speculative' questions...

How might a language develop an animate/inanimate (or any other binary gender-like) distinction, on demonstratives specifically? I ask about demonstratives, since they are likely one of the first points in the development of a gender system (at least according to Corbett and others).

The obvious answer is that a simple demonstrative might gain a fused element, so "this" might split into "this" (animate) and "this.thing" (inanimate).

If we consider PIE *so (animate) and *to (inanimate), it looks like both are very old, being a single CV syllable each.

The demonstrative + lexical word idea is fine, but I'd also like to hear some other interesting pathways - maybe analogy or reanalysis could play a role.

Interestingly, I have read that demonstratives *may* be one of the few things in a language that have no lexical source (although they are often recycled and fortified via the addition of locatives or pronouns etc.), with proximal demonstratives often having a high vowel and distal a more sonorous vowel (so a kind of iconicity essentially).**

If the PIE example does come from an 'original' demonstrative + lexical word(s), this must have happened a very long time ago, with them having become eroded to such an extent. This made me wonder whether it is reasonable to assume that the animate/inanimate (or masculine/feminine) distinction could also arises 'spontaneously' without any specific lexical source - the driving factor is simply that they are (phonologically) different.

I'd be interested to hear any thoughts!

** Apparently there is evidence for this in post-eighth-century Chinese, where the distal demonstrative nà has no clear etymology, and various dialects all show a large variety of forms also without a clear etymology. This is even more unusual given that the preceding stages had a wealth of demonstratives, and the new one doesn't look like it is a recycling of old elements. The common factor is that the distal tends to be more sonorous overall (where the hypothesis is that an increase in sonority = an increase in distance, iconically, so the new one simply arose of its own accord).

[Nel Fie]

I don't think there is an easy asnwer but here are some functions that I know of.

1) The difference between the two arguments of transitive verbs. Happens in English (and probably many other languages).
2) Information structure, so things like topic and focus. Happens in many languages including English, cf. Beans, I like.
3) Something that happens in noun adjective order but I forgot what it is. Happens in some Romance language, IIRC.

I don't think anything and everything is the correct answer. I know of no language that uses only word order to distinguish person, number, tense, aspect, modality, or evidentiality. But I might be wrong of course. I think, I had the idea of using word order in unusual ways for Pansi here (https://cbbforum.com/viewtopic.php?p=322245#p322243).

Thank you for your insights. Admittedly, I'm surprised there seems to be so little - I would have expected much more, especially from analytic and isolating languages. But as your own work illustrates, it's maybe a nice avenue to explore new and unique ideas in conlanging.

As for point 3), the Wikipedia article on postpositive adjectives has some examples that seem to distinguish meanings by word order in French. Is this what you're thinking of?

When an adjective can appear in both positions, the precise meaning may depend on the position. E.g. in French:

un grand homme - "a great man"
un homme grand - "a tall man"
une fille petite - "a small girl"
une petite fille - "a little girl"
un petit chien - "a little dog (of a small breed)"
un chien petit - "a small dog (for its breed)"

[Davush]

A current curiosity of mine: for how many grammatical functions do natural languages employ word order? And what are those functions?

Though I'd wager the answer will probably be 'everything and anything', so perhaps a more specific question - among the languages anyone here might know about, which natural language concurrently employs word order for more grammatical functions than any other? And would the quantity of uses it makes of word order be unusual compared to most languages?

This is an interesting question, and it's also interesting to consider why word order isn't used for "anything and everything".

I think that cognition is probably quite involved: people generally seem to prefer to speak languages where the ordering of elements within a phrase is (largely) consistent. For example, overall head-final languages tend towards most other phrases also being head-final (and likewise with head-initial preferring head-initial). This seems like a cognitive motivation - competing word orders make things harder to understand.

When word order is involved in things other than basic argument structure, it seems to be very restricted such as English subject inversion for questions, or the restrictive/non-restrictive adjective placement in Spanish.

It also seems there is an overwhelming preference for the subject/agent to precede the object/patient - probably also because language is linear, and the subject/agent is more often than not what is being 'commented on' (topic-comment) and/or a causal agent (cause-effect), so it prefers to come first. However, people also like to be able to switch things around for focus, emphasis, etc. Maybe the object is actually being 'commented on', in which case it is often fronted. If word order then also had to take on the additional load of marking number, or TAM, or other things, it would (presumably) make it less available for more 'fundamental' functions.

Also pragmatically: if the order of S-O can switch around, their grammatical role will need to be marked, i.e., they will have affixes. If the nouns have affixes, the verb probably also have affixes (maybe there's a language somewhere with extensive nominal inflection but nothing on the verb, though?). This makes word order somewhat redundant, since roles are overtly marked.

I think you could get creative as long as it's fairly restricted though (assuming you want naturalism, that is). I could imagine something like N-Postposition signalling a singular noun, while Preposition-N signals a plural, where perhaps Preposition-N was originally something like 3pl-Postposition Noun. (Even so, I imagine such a system would be quite unstable, and would soon generalise to Preposition-N or N-Postposition).

So basically my intuition is that people are 'hardwired' to use word order primarily for argument & information structure, with some slight (and more specialised) curiosities here and there.

[Salmoneus]

Another of my 'speculative' questions...

How might a language develop an animate/inanimate (or any other binary gender-like) distinction, on demonstratives specifically? I ask about demonstratives, since they are likely one of the first points in the development of a gender system (at least according to Corbett and others).

The obvious answer is that a simple demonstrative might gain a fused element, so "this" might split into "this" (animate) and "this.thing" (inanimate).

Another answer would be just nouns - "person" vs "thing".

Or honorifics ("this.polite" > "this.animate")

Another might be distal vs proximal deictics ("this" > "this.animate" and "that" > "this.inanimate"; or "this with me" > "this.inanimate" vs "that with you" > "this.animate")

Another might be possessives ("your thing" > "that.inanimate thing", "his own thing" > "that.animate thing")

Another might be replacement of an existing demonstrative but only for animates. For instance, in Germanic languages, the function of relative "that" has been supplanted by interrogative "who" and "which" to some extent (although perhaps being reversed again in English). It's not hard to imagine that only happening with animates, so "the man who..." but "the table that...". And it's not hard to imagine something similar happening with demonstratives instead, or starting with relativisers and spreading to demonstratives.

An alternative approach might be analogical restructuring of the demonstrative to agree with the noun. If the demonstrative is "so" and inanimates end in -o, but animate nouns end in -a, perhaps a new animate demonstrative "sa" is formed.

[Nel Fie]

This is an interesting question, and it's also interesting to consider why word order isn't used for "anything and everything".
[...]

So basically my intuition is that people are 'hardwired' to use word order primarily for argument & information structure, with some slight (and more specialised) curiosities here and there.

Thank you very much for the extensive write-up! What you say makes sense indeed, although at the same time I'm wary of applying "common sense" to linguistics, considering there seem plenty of natural languages that defy common sense ideas. But these languages tend to come up in conversation like these, whereas nobody seems to know any natlang that does exceedingly strange things with word order, so I suppose it must indeed be one of those areas where natlangs are pretty comparable across the aboard, with no remarkable outliers.

[Salmoneus]

I don't think anything and everything is the correct answer. I know of no language that uses only word order to distinguish person, number, tense, aspect, modality, or evidentiality. But I might be wrong of course.

I can't name any, but I'd be surprised if there are no languages that at least secondarily use word order to distinguish tense, aspect and/or mood.

I say this because those things can often trigger current or historical structural shifts, including nominalisations, that require changes in word order, and that can have, or can be left with after sound changes and levellings, few or no other markers.

English for instance is SVO, but its nominalisation are often still SOV ("your mushroom-eating offends me"), so if English were to use a construction derived from a reanalysed nominalisation as its main way of indicating, say, a perfect, then that could involve, or even only involve, a change in word order from SVO to SOV.

I don't imagine such languages are plentiful, however.

[Arayaz]

I don't think anything and everything is the correct answer. I know of no language that uses only word order to distinguish person, number, tense, aspect, modality, or evidentiality. But I might be wrong of course.

I can't name any, but I'd be surprised if there are no languages that at least secondarily use word order to distinguish tense, aspect and/or mood.

I remember doing this:

Present: VSO
Future: go.to S {VO}

Then the go.to marker got elided into nonexistence (except in a few contexts) and then analogy made it so that present was VSO and inverting the subject and verb marked the future.

[Creyeditor]

I don't think anything and everything is the correct answer. I know of no language that uses only word order to distinguish person, number, tense, aspect, modality, or evidentiality. But I might be wrong of course.

I can't name any, but I'd be surprised if there are no languages that at least secondarily use word order to distinguish tense, aspect and/or mood.

I say this because those things can often trigger current or historical structural shifts, including nominalisations, that require changes in word order, and that can have, or can be left with after sound changes and levellings, few or no other markers.

English for instance is SVO, but its nominalisation are often still SOV ("your mushroom-eating offends me"), so if English were to use a construction derived from a reanalysed nominalisation as its main way of indicating, say, a perfect, then that could involve, or even only involve, a change in word order from SVO to SOV.

I don't imagine such languages are plentiful, however.

I don't think we disagree. As I said, I don't know any language that marks these categories by only changing word order. You said you can't name any.

I can see the diachronic pathway but I am convinced that there is more than diachrony. I would guess that in your scenario the marking on the verb is less likely to be dropped or more likely to be retained.

My guess at a functional explanation would be kind if the opposite of what Davush said. People like to switch around word order for information structural reasons or for dramatic reasons in (almost) all languages. If you also use it for other stuff, you have less to switch around and need to use more clunky constructions like English "speaking of X" or "as for X". Since almost every sentence has a non-redundant focus and topic but not every sentence has a non-redundant tense, word order is used for the former.
Of course, as is the nature of these kinds if guesses, this is all just speculation.

Also, minor clarification: Using word order for mood (in the sense of sentence type) is of course attested. In German, embedded clauses use SOV vs main clauses use verb-second. English uses Subject-Auxiliary Inversion in polar questions. Imperatives often differ in word order. But modality (in the sense of how much I think that a proposition is true (in a perfect world or our world) is among the categories that I have not seen marked by word order.

[Davush]

My guess at a functional explanation would be kind if the opposite of what Davush said.

(I pretty much said exactly the same thing)

[...] people also like to be able to switch things around for focus, emphasis, etc. [...] If word order then also had to take on the additional load of marking number, or TAM, or other things, it would (presumably) make it less available for more 'fundamental' functions.

Also, Semitic might be a "partial" example, where the imperfective is marked by verbal prefixes, and the perfective by suffixes (granted the two sets of affixes are different and there is stem-internal morphology too), but I could imagine this situation being extended to nouns overall. (Although as Salmons mentions, if this type of thing does occur, it's certainly not frequent.) Nonetheless, I agree that there is probably something more fundamental in our use of language that disfavours using word order for this type of stuff.

Another answer would be just nouns - "person" vs "thing".
[...]
An alternative approach might be analogical restructuring of the demonstrative to agree with the noun. If the demonstrative is "so" and inanimates end in -o, but animate nouns end in -a, perhaps a new animate demonstrative "sa" is formed.

Thanks! All nice pathways to consider (although the final one leads to me ask 'but why would inanimates all start ending in -o and animates in -a which I suppose is a chicken-and-egg type of a thing, since fused articles/demonstratives would be the easiest answer).

[Creyeditor]

@Davush: Sorry, I must've misread your post.

[Visions1]

I don't think there is an easy asnwer but here are some functions that I know of.

1) The difference between the two arguments of transitive verbs. Happens in English (and probably many other languages).
2) Information structure, so things like topic and focus. Happens in many languages including English, cf. Beans, I like.
3) Something that happens in noun adjective order but I forgot what it is. Happens in some Romance language, IIRC.

I don't think anything and everything is the correct answer. I know of no language that uses only word order to distinguish person, number, tense, aspect, modality, or evidentiality. But I might be wrong of course. I think, I had the idea of using word order in unusual ways for Pansi here (https://cbbforum.com/viewtopic.php?p=322245#p322243).

Doesn't Navajo have a whole thing with word order? It doesn't seem to mark anything, but it seems to organize things.

[VaptuantaDoi]

Definiteness can be kinda marked through word order in Hungarian


Péter olvas egy újságot
peter read.3SG INDEF newspaper.ACC
"Peter is reading a [specific] newspaper"

vs.

Péter újságot olvas
peter newspaper.ACC read.3SG
"Peter is newspaper-reading" (?)



The idea is that – crosslinguistically – indefinite objects tend to get more closely tied to the verb; in this case it happens by switching places (along with dropping the article, but I think that's fairly unimportant) to act as a verb modifier. It's like borderline NI.

[Creyeditor]

Which makes sense because definiteness could be construed of as a kind of information structure. It should definitely be added to the list.