Linguistic Miscellany Thread

[zompist]

We agree on a lot, so I will try to focus on remaining points of disagreement (or maybe it's different backgrounds; I weigh pragmatics very highly).

Another puzzle about epistemic contradictions is that, as Beddor and Goldstein (2018) point out, it's perfectly reasonable to personally hold a belief in p, while also acknowledging that that belief might be mistaken and that it might be the case that not p. But asserting both of these things at once, in the same breath, ends up sounding odd (to most people, I guess).

It's odd only if you view eliminating contradictions as possible in practice.

I assume you know Makinson's paradox of the preface. Authors usually state that "remaining errors in the book are mine", and yet if asked, would maintain that they've fact-checked everything and are not aware of any errors. This is hard to address with classical logic, but on a human level it's neither strange nor uncommon.

Yeah, this kind of data is one of the big arguments for the "must is weak" camp—e.g. Lassiter (2016), who argues for a probablistic semantics for must. As you note, it's quite tricky to deal with this for the "must is strong" camp, and you're forced to say something that perhaps isn't so satisfactory.

I'm glad we agree on this, though I'd note that in form these sentences are identical to "It must be raining, but perhaps it's not." Classical logic at least doesn't let us peer into a proposition P to see if it's inherently difficult to test.

I'm not sure if I understand what the disagreement is then—neither I nor von Fintel and Gillies (nor Lassiter, nor Mandelkern, etc.) are challenging the idea that must indicates deductive grounds/indirect evidence for our belief---I think everyone (now) acknowledges that this is probably true. But the question that people are arguing about in the literature is whether must p entails p, or if it doesn't.

It seems to me that you want to divide up "must" into two things, one of which we agree on.

I don't see that there's a pragmatic or deductive part that you can take away, leaving some sort of "essence of must" where must(p) -> p. To me the parts you want to set aside-- the things we both acknowledge weaken its meaning-- are key to its meaning. But perhaps I just don't understand you... what is the "other part" of "must"?

The only candidate I can see is the central meaning of must-- absolute necessity. It's hard to come up with loophole-free examples... maybe something from basic physics, or Boolean logic. If it helps, I agree that if P is absolutely necessary, then P. But this is one of those words that are almost always used in a much lesser sense. Not even "it must be raining" is a claim of absolute necessity.

When you say "emphasizes where language does not work like logic", I wonder what kind of logic you mean? I think most semanticists well understand that human language doesn't work on the principles of classical logic—this is why we make use of the tools of modal logic, or sometimes multivalued logics, etc.

It goes beyond "if", which he discusses. The words "and, or, each, every", and more, don't exactly match their logical counterparts.

This doesn't mean that there's something wrong with logic, just that we can't just say that e.g. "and" means AND and that's it. And conjunctions are relatively simple.

[priscianic]

E.g. the Wikipedia article on "Syntax" ideally wouldn't begin with "syntax [...] is the set of rules, principles, and processes that govern the structure of sentences (sentence structure) in a given language, usually including word order", having a dedicated section to S-V-O order afterwards and no other specific topic in particular. Way too many people in conlangery at large at least seem to think syntax is word order plus esoteric frameworks and trees. The long list of terms at the end of the article includes things like "gapping" and "subordination", but they're hardly in focus...

Maybe I should edit that Wikipedia article and get rid of its focus on word order, but the thing is it also reflects a widespread attitude among language hobbyists (language learners, conlangers, people who like linguistics but are not academics) towards formal syntax.

Yeah, this is one of my personal bugbears about conlangers and syntax—they think it's all about word order, but we actually don't really think about word order all that much...

Yet, if I see a ‘syntax’ section in a reference grammar, it’s usually all about word order. So, if that’s not ‘syntax’, then what is syntax actually about, and where does word order fit into it?

(This is also one of my personal bugbears about (many) documentary linguists too...)

Syntax is, very abstractly, just about the structure of sentences. I think one useful way to think of syntax, especially for conlangers, is as the answer to the following question: what do you need to know in order to put together a well-formed sentence from its individual bits? Obviously word order is a part of that, but I hope it should be clear that there's a lot more to it than just word order. There's more information you need to know besides just what goes where.

For instance, what exactly are you ordering? When people say things like SVO, what is S, what is V, and what is O? Those aren't theoretically-innocent letters: they're syntactic notions that people take to be in some sense "fundamental" to how languages are structured. That's syntax too! (and there's plenty of disagreement and quibbling to be had about what S, V and O mean and whether they're actually coherent, useful concepts.)

Some more specific questions you could ask: how do I put together a polar question? how do I put together a content question? how do I put together a relative clause? etc.

To really explore that small space, you also have to get away from that small class of things that are immediately verifiable. I'd put it to you that statements like 16a sound much more plausible in sentences like this:

Gravity must be mediated by a graviton, though we haven't discovered one yet. But perhaps it isn't.
Life must exist in the Tau Ceti system... but perhaps it doesn't.
God must be about to punish this city full of sin, but perhaps he's not.

These amount to saying "What I know requires a conclusion p. But these are things that are hard to know and I recognize or fear that ¬p is after all not ruled out." They have an air of anxiety, because normally we don't like to confront the fact that our reasoning and beliefs are shaky. But they're certainly not meaningless or impossible.

Yeah, this kind of data is one of the big arguments for the "must is weak" camp—e.g. Lassiter (2016), who argues for a probablistic semantics for must. As you note, it's quite tricky to deal with this for the "must is strong" camp, and you're forced to say something that perhaps isn't so satisfactory. Von Fintel and Gillies (2020) try to address this kind of data, and argue that it involves a change in context—in particular, a change in which possibilities are deemed relevant at a given point in time (in their words, you're "expanding the modal horizon"). Understandably, if you allow the context to so rapidly change like this, you worry that your theory loses any bite at all—or as they put it, you worry about whether they're putting their theory of "pragmatic life support" to shield it from any counterexamples. So they try to support this idea by trying to control for shifts in the "modal horizon" (though some of their data is, as they admit, somewhat shaky).

As a non-syntactician, I would think that these sort of sentences would best be analysed by assuming ‘must’ is a radial category. In that case, the prototypical meaning of ‘must’ would be something like ‘X is true in all possible worlds where the premises hold’, but humans have then extended that to just act as a general intensifier. That explanation doesn’t seem at all unreasonable to me (if anything, it seems natural!) — is there any reason it hasn’t been adopted more widely?

I'm not sure if it's true that that kind of approach hasn't been adopted more widely. I'm personally more familiar with model-theoretic formal semantics, and in this area of the field we don't really make use of the idea of radial category. But there are plenty of people who work on meaning that aren't model-theoretic formal semanticists, and they might well be writing bucketloads about radial categories and modals—unfortunately, I'm really not the person to ask about that.

As for why model-theoretic formal semanticists don't employ radial categories, I think some abstract knowledge of what model-theoretic formal semantics looks at and how the field works might be instructive.

The kind of core things the field is trying to explain are when sentences are true and when they are false (or, potentially, when they're undefined), and when one sentence implies another sentence. In other words, the "truth conditions" of a sentence. An additional thing to explain is how we get to those truth conditions from the individual morphemes that make up those sentences—the problem of compositionality (as such, the field is fundamentally opposed to the idea that natural language meaning is noncompositional).

The process generally looks like this, in a very idealized form:

• Try to figure out the truth conditions of a sentence (e.g. by eliciting truth value judgments in different contexts, or by checking entailment patterns between different sentences, etc.);
• Come up with compositional rules and a syntax that are able to combine the various bits that make up the sentence (for instance, function application: we can say that one morpheme is a function that takes another in as an argument);
• Come up with formal denotations for each of the various bits that, when combined following the compositional rules and the syntax, ends up with the right truth conditions.

In principle, you could imagine formally defining some notion of "radial category" that fits into the broader picture. But I think, in practice, that that would end up looking like one of two things:

• A morpheme whose denotation is underspecified, such that it ends up looking like it means different things at different times
• A morpheme that has multiple denotations (i.e. is polysemous/homophonous)

And we actually currently do make use of both these kinds of things, though we don't think of them as "radial categories".

I also think these options lose some of the fundamental intuitions and insights of the concept of "radial category" (e.g. the notion that there's a "core meaning" is lost, for instance). Personally, I'm not sure how radial categories could fit into this broader picture without looking some of that fundamental character. I guess you could imagine taking the second, polysemy/homophony option, and just declaring by fiat that one of these denotations is "core". But I don't think that would actually have any strong impact on the theory itself, besides just painting one denotation a special color.

It's also worth noting that this kind of semantics sort of just pushes lexical semantics aside as an issue for someone else to solve, and instead focuses more on functional items like quantifiers, conditionals, tense, aspect, modality, etc. In contrast, if I'm not mistaken, the prototype theory/radial category people very much start from lexical semantics and the meaning of lexical items, building off of insights from psychology/cognitive science of concepts and meaning. (Similarly, those people don't do as much work on the nitty-gritty of things like quantifiers, conditionals, tense, aspect, modality, etc., as far as I'm aware.)

[bradrn]

Yeah, this is one of my personal bugbears about conlangers and syntax—they think it's all about word order, but we actually don't really think about word order all that much...

Yet, if I see a ‘syntax’ section in a reference grammar, it’s usually all about word order. So, if that’s not ‘syntax’, then what is syntax actually about, and where does word order fit into it?

(This is also one of my personal bugbears about (many) documentary linguists too...)

Syntax is, very abstractly, just about the structure of sentences. I think one useful way to think of syntax, especially for conlangers, is as the answer to the following question: what do you need to know in order to put together a well-formed sentence from its individual bits? Obviously word order is a part of that, but I hope it should be clear that there's a lot more to it than just word order. There's more information you need to know besides just what goes where. … Some more specific questions you could ask: how do I put together a polar question? how do I put together a content question? how do I put together a relative clause? etc.

A clarification: when I talked about ‘word order’, I was including stuff like relative clauses and questions — I consider those to be fundamentally a matter of word order as well.

For instance, what exactly are you ordering? When people say things like SVO, what is S, what is V, and what is O? Those aren't theoretically-innocent letters: they're syntactic notions that people take to be in some sense "fundamental" to how languages are structured. That's syntax too! (and there's plenty of disagreement and quibbling to be had about what S, V and O mean and whether they're actually coherent, useful concepts.)

That last statement is pretty surprising — to me, V/S/O (or V/S/A/O if you want to split up S) have always seemed like pretty useful and coherent concepts. Could you expand a bit on why this may not be?

In principle, you could imagine formally defining some notion of "radial category" that fits into the broader picture. But I think, in practice, that that would end up looking like one of two things:

• A morpheme whose denotation is underspecified, such that it ends up looking like it means different things at different times
• A morpheme that has multiple denotations (i.e. is polysemous/homophonous)

And we actually currently do make use of both these kinds of things, though we don't think of them as "radial categories".

I also think these options lose some of the fundamental intuitions and insights of the concept of "radial category" (e.g. the notion that there's a "core meaning" is lost, for instance). Personally, I'm not sure how radial categories could fit into this broader picture without looking some of that fundamental character. I guess you could imagine taking the second, polysemy/homophony option, and just declaring by fiat that one of these denotations is "core". But I don't think that would actually have any strong impact on the theory itself, besides just painting one denotation a special color.

It's also worth noting that this kind of semantics sort of just pushes lexical semantics aside as an issue for someone else to solve, and instead focuses more on functional items like quantifiers, conditionals, tense, aspect, modality, etc. In contrast, if I'm not mistaken, the prototype theory/radial category people very much start from lexical semantics and the meaning of lexical items, building off of insights from psychology/cognitive science of concepts and meaning. (Similarly, those people don't do as much work on the nitty-gritty of things like quantifiers, conditionals, tense, aspect, modality, etc., as far as I'm aware.)

Thanks for explaining! That makes a lot of sense — if you’re already allowing underspecified and multiple denotations, there isn’t really any need to introduce a whole new concept of ‘radial categories’.

[priscianic]

Another puzzle about epistemic contradictions is that, as Beddor and Goldstein (2018) point out, it's perfectly reasonable to personally hold a belief in p, while also acknowledging that that belief might be mistaken and that it might be the case that not p. But asserting both of these things at once, in the same breath, ends up sounding odd (to most people, I guess).

It's odd only if you view eliminating contradictions as possible in practice.

I assume you know Makinson's paradox of the preface. Authors usually state that "remaining errors in the book are mine", and yet if asked, would maintain that they've fact-checked everything and are not aware of any errors. This is hard to address with classical logic, but on a human level it's neither strange nor uncommon.

I'm not actually aware of the paradox of the preface—thanks for alerting me to it!

It's hard to address with classical logic, but that's why natural language semanticists don't use pure classical logic :p
It's very common to use tools from modal logic and multivalued logics, as I mentioned previously, and as well to embed the denotation of individual sentences within the an explicit theory of discourse context (i.e. to recognize that pragmatics exists, and should be explicitly theorized). (There's a whole subfamily of semantics called "dynamic semantics" which treats the meaning of sentences not as truth conditions but rather as functions from discourse contexts to discourse contexts, and as such takes very seriously the idea that utterances are made in a context in a very specific way).

Yeah, this kind of data is one of the big arguments for the "must is weak" camp—e.g. Lassiter (2016), who argues for a probablistic semantics for must. As you note, it's quite tricky to deal with this for the "must is strong" camp, and you're forced to say something that perhaps isn't so satisfactory.

I'm glad we agree on this, though I'd note that in form these sentences are identical to "It must be raining, but perhaps it's not." Classical logic at least doesn't let us peer into a proposition P to see if it's inherently difficult to test.

Yeah, exactly—this is precisely why von Fintel and Gillies have to appeal to pragmatics and the context-sensitivity of assertion in order to deal with this data.

I'm not sure if I understand what the disagreement is then—neither I nor von Fintel and Gillies (nor Lassiter, nor Mandelkern, etc.) are challenging the idea that must indicates deductive grounds/indirect evidence for our belief---I think everyone (now) acknowledges that this is probably true. But the question that people are arguing about in the literature is whether must p entails p, or if it doesn't.

It seems to me that you want to divide up "must" into two things, one of which we agree on.

I don't see that there's a pragmatic or deductive part that you can take away, leaving some sort of "essence of must" where must(p) -> p. To me the parts you want to set aside-- the things we both acknowledge weaken its meaning-- are key to its meaning. But perhaps I just don't understand you... what is the "other part" of "must"?

Ah, no, I think you're misunderstanding the position. The position is that must has, at its core, the meaning of the necessity operator ☐ from modal logic—i.e. ☐p is true at a world w iff for every world v that is "accessible" from w by some relation R (i.e. wRv), p is true in v.

Now, what is R? What are these "accessible worlds"? Here, there's some freedom and context-sensitivity to play with. Making use of this context-sensitivity, von Fintel and Gillies formalize for us some ways to restrict R and these accessible worlds v in order to capture the deductive character of must (they don't characterize it in exactly this way, but this is essentially what they're doing). And of course, the set of worlds that are accessible via R is context-sensitive—intuitively, you'd only consider the relevant/salient possibilities. According to them, it's the combination of those two things that gets you the right characterization of the felt weakness of must.

They also want the R you get with epistemic modals to be "reflexive"—i.e. for wRw to be true in all worlds in the model. This is supposed to be a characteristic of epistemic modals—they quantify over worlds that are open possibilities for what the actual world could be—so the "accessibility relation" R that you get with an epistemic modal needs to be reflexive, in order to capture this intuition. (Actually, strictly speaking, all you need is that the actual world w* be accessible from itself. But general reflexivity is equivalent for our purposes here.) A consequence of the reflexivity of R is that if you universally quantify over all worlds v such that wRv, then you'll also end up quantifying over w as well. Thus, ☐p ends up entailing p (i.e. ☐p being true in a world w entails that p is true in w).

Von Fintel and Gillies think that this entailment is fine, and that we can capture the data by appealing the indirectness of must to capture its felt weakness, plus some pragmatic principles. Lassiter, in contrast, thinks that this entailment is bad, and he weakens the meaning of must to instead quantify over almost all of the accessible worlds (a kind of probabilistic semantics), with the threshold of what counts as "enough" worlds being contextually variable. (And then he can also do some extra stuff to capture the deductive character of must, just like von Fintel and Gillies).

So at its core, I think the disagreement can be boiled down to the following: von Fintel and Gillies (and others) think that the entailment from must p to p is fine, and that we can capture the weakness of must with a combination of indirect evidentiality and the context-sensitivity of the relevant accessibility relation R. In contrast, Lassiter (and others) think that this entailment is terrible, and that we can't capture the weakness of must with just indirect evidentiality and context-sensitivity of R, and that we have to instead weaken must to quantifying over only most of the relevant worlds, rather than all of them (with the appropriate context-sensitivity for how many worlds count as "most" worlds, of course).

I hope that all makes sense...

When you say "emphasizes where language does not work like logic", I wonder what kind of logic you mean? I think most semanticists well understand that human language doesn't work on the principles of classical logic—this is why we make use of the tools of modal logic, or sometimes multivalued logics, etc.

It goes beyond "if", which he discusses. The words "and, or, each, every", and more, don't exactly match their logical counterparts.

This doesn't mean that there's something wrong with logic, just that we can't just say that e.g. "and" means AND and that's it. And conjunctions are relatively simple.

Tell me about it! You'd hope that conjunctions are simple, but then you see that people have written 1000-page books on the semantics of natural-language "and"...(though, funnily enough, Schein's argument is actually that "and" is always Boolean conjunction, and he has to spend 1000 pages defending that thesis and addressing all the counterarguments...)

[priscianic]

For instance, what exactly are you ordering? When people say things like SVO, what is S, what is V, and what is O? Those aren't theoretically-innocent letters: they're syntactic notions that people take to be in some sense "fundamental" to how languages are structured. That's syntax too! (and there's plenty of disagreement and quibbling to be had about what S, V and O mean and whether they're actually coherent, useful concepts.)

That last statement is pretty surprising — to me, V/S/O (or V/S/A/O if you want to split up S) have always seemed like pretty useful and coherent concepts. Could you expand a bit on why this may not be?

Ah, there are hundreds (if not thousands) of pages you could write about "subjecthood" alone—you can see McCloskey (1997) and Poole (2016) for just a small taste.

The issue boils down to the fact that subjecthood is very slippery—it's extremely hard (if not impossible) to define reliably, and it's extremely hard to come up with core "subjecthood diagnostics". In the Chomskyan generative tradition (though not other generative traditions, as Poole notes, and not in a lot of the typological literature), the standard answer to this issue is to deny that subjecthood is a unitary primitive in the grammar, but rather that each subjecthood diagnostics potentially diagnoses a separate property, and that constituents that satisfy multiple subjecthood diagnostic actually have several of these properties. (Though we do use the term "subject", for better or for worse, but if the author is doing their due diligence they will have specifically defined what exactly they mean by "subject".)

As for objecthood, I think similar kinds of ideas apply as well, though people haven't written so much about the difficulty of objecthood specifically. But sometimes certain objects in one construction might show similar properties as certain subjects in other constructions, and the whole thing can end up looking like a mess if you try to go digging deep into what exactly "objecthood" is, in a similar sort of way.

As for verbs, there's less controversy over whether verbs exist. But if you're one of the people that buy that some languages don't have N/V distinctions, then that calls into question what "verbhood" really is, and how we can properly define it, and in particular how Vs are distinct from Ss and Os (if Ss and Os are also Vs...).

This is also a separate question to whether S, V, and O are generalizations/crude abstractions that are useful and helpful for certain purposes (e.g. large-scale word order typology). It may well be true that these concepts start to break down if you stare at them hard enough, but are still useful for some things. (Perhaps a useful analogy might be to how Newtonian mechanics starts to break down when you get small or big enough, but it suffices for a lot of purposes.)

[Kuchigakatai]

Yet, if I see a ‘syntax’ section in a reference grammar, it’s usually all about word order. So, if that’s not ‘syntax’, then what is syntax actually about, and where does word order fit into it?

A clarification: when I talked about ‘word order’, I was including stuff like relative clauses and questions — I consider those to be fundamentally a matter of word order as well.

I find this... interesting. I'm guessing this means that for you the selection of word forms is part of morphology, and then you also see changes in phrases as mostly word order, without minding structures or changes in semantic roles except as something of an aside...? So, say, in something like "He really wants something, what does he really want?", you see the "want" > "wants" change as "morphology", and the movement of "something" > "what" to the beginning as word order. But then what would you say about the insertion of "does" as the new main verb, to carry the inflection? It seems strange to me to not think much of the change from active to passive voice from Marcus amat 'Mark loves' to Marcus amātur 'Mark is loved' except in terms of morphology (active amat > passive amātur) because the words don't change in order. In the latter, you could insert ā Terentiā 'by Terentia' as the agent, an addition invalid in the former.

I'd like to ask if you can tell how you've arrived at this view, which, please note, I'm not imputing you particularly at fault of. It's not uncommon among conlangers...

It also isn't true in my experience that when I see a reference grammar, the syntax section is all about word order, including in your wider definition... What I usually see is that there is no section labelled "Syntax" at all.

Sometimes it's because all the morphology and syntax bits get treated together at the same time —an example of this would be Donaldson's German: An Essential Grammar (2007), in which every chapter first presents word forms and then how those forms are used.

But sometimes it's because, although the linguists are clearly dividing the grammar into Morphology vs. Syntax, the syntax section is so, so much longer than the morphology section that it's likely felt pointless to group most of the chapters under a "Syntax" chapter group. An example would be Badawi et al.'s Arabic: A Comprehensive Grammar (2002) covering inflectional morphology in chapter 1, syntax from chapters 2 to 11, finishing with derivational morphology in chapter 12. Or for a more radical example using a language with little morphology, Li and Thompson's Mandarin Chinese: A Functional Reference Grammar (1981), which starts with two overview chapters, goes on with a chapter on morphology, and then covers syntax topics from chapters 4 to 24.

For an example where the label "Syntax" does get used as a chapter group, there's Allen & Greenough's New Latin Grammar for Schools and Colleges (1903 edition, 1st ed. in 1872), which is divided into three parts, one labelled "Words and Forms" (=inflectional and derivational morphology, 8 chapters), one labelled "Syntax", which covers things like when each of the cases is used, or how main clauses are turned into noun clauses (4 chapters grouping 34 sections), and one labelled "Prosody", which deals with poetic metres (3 sections), plus a final separate short section on measures. The "Syntax" chapter group ends up taking 55% of the content (vs. the 38% of the "Words and Forms" part).

[bradrn]

Yet, if I see a ‘syntax’ section in a reference grammar, it’s usually all about word order. So, if that’s not ‘syntax’, then what is syntax actually about, and where does word order fit into it?

A clarification: when I talked about ‘word order’, I was including stuff like relative clauses and questions — I consider those to be fundamentally a matter of word order as well.

I find this... interesting. I'm guessing this means that for you the selection of word forms is part of morphology, and then you also see changes in phrases as mostly word order, without minding structures or changes in semantic roles except as something of an aside...? So, say, in something like "He really wants something, what does he really want?", you see the "want" > "wants" change as "morphology", and the movement of "something" > "what" to the beginning as word order. But then what would you say about the insertion of "does" as the new main verb, to carry the inflection? It seems strange to me to not think much of the change from active to passive voice from Marcus amat 'Mark loves' to Marcus amātur 'Mark is loved' except in terms of morphology (active amat > passive amātur) because the words don't change in order. In the latter, you could insert ā Terentiā 'by Terentia' as the agent, an addition invalid in the former.

I think I maybe was a bit unclear here. I have a very broad definition of ‘word order’: it includes anything where the order of words is important. e.g. in English, that includes questions and relative clauses — which in English are formed by reordering words.

Note that I have not said that structures, changes in semantic roles etc. are unimportant — they have their place as well! But, fundamentally, I see syntax as a matter of putting words in the correct order (or moving them there, of course).

I also think that what I’m saying is not too dissimilar to your ‘naïve syntax theory’ which you mentioned here. Fundamentally, both my conception of syntax and that one are just talking about words being assembled and moving around — not saying that everything else is unimportant.

I'd like to ask if you can tell how you've arrived at this view, which, please note, I'm not imputing you particularly at fault of. It's not uncommon among conlangers...

I’m not sure I actually ‘arrived’ at it as such — it’s just always seemed obvious to me.

It also isn't true in my experience that when I see a reference grammar, the syntax section is all about word order, including in your wider definition... What I usually see is that there is no section labelled "Syntax" at all.

Well, yes — I was talking in general about all the sections which cover syntax.

But, for an example of what I meant, take Hyslop’s grammar of North-East Ambae (it’s freely available; Wikipedia should have a link). Most of this grammar is, as I see it, about word order. In fact, I’m not quite sure what else it would be about! For instance, the chapter on the noun phrase goes into detail about which words can go before the noun, which words and phrases can go after the noun, how all those are ordered, and of course how each of those noun modifiers are constructed. All of those are, as I see it, matters of word order. (I’m honestly not sure what else they could be, but feel free to correct me if you think they’re something else!)

[Qwynegold]

I'm losing my mind over Chinese stroke order! For some characters it's different than in Japanese, so I'm trying to set straight what's correct in Chinese. But different source disagree with one another. I have a handout from when I went to a Chinese class that says that the order for 北 is this:

1. The left vertical
2. The NW horizontal
3. The SW horizontal
4. The right leg
5. The NE horizontal

But some source I found online said it was:

1. The left vertical
2. The NW horizontal
3. The SW horizontal
4. The NE horizontal
5. The right leg

And now I have found a source that says it's:

1. The NW horizontal
2. The left vertical
3. The SW horizontal
4. The NE horizontal
5. The right leg

Which one is it? And if you know of an authorative correct source, can you tell me what are the orders for 比 and 区?

[Kuchigakatai]

I think I maybe was a bit unclear here. I have a very broad definition of ‘word order’: it includes anything where the order of words is important. e.g. in English, that includes questions and relative clauses — which in English are formed by reordering words.

Note that I have not said that structures, changes in semantic roles etc. are unimportant — they have their place as well! But, fundamentally, I see syntax as a matter of putting words in the correct order (or moving them there, of course).

I also think that what I’m saying is not too dissimilar to your ‘naïve syntax theory’ which you mentioned here. Fundamentally, both my conception of syntax and that one are just talking about words being assembled and moving around — not saying that everything else is unimportant.

Well, yes — I was talking in general about all the sections which cover syntax.

But, for an example of what I meant, take Hyslop’s grammar of North-East Ambae (it’s freely available; Wikipedia should have a link). Most of this grammar is, as I see it, about word order. In fact, I’m not quite sure what else it would be about! For instance, the chapter on the noun phrase goes into detail about which words can go before the noun, which words and phrases can go after the noun, how all those are ordered, and of course how each of those noun modifiers are constructed. All of those are, as I see it, matters of word order. (I’m honestly not sure what else they could be, but feel free to correct me if you think they’re something else!)

I can't help but find this a very strange way to look at syntax... Also, it's not what I had in mind when I mentioned people who naively discuss syntax in that post, but anyway...

I mean, the common way to look at morphology and syntax, in linguistics at least, sees morphology as contextless. Linguists really just look at a word and the various meanings it has to determine its synchronic morphological structure, its inflections and derivations. So here you are using a different definition of "morphology" and "syntax" altogether. For example, English "user" in the sense of 'person who uses' involves the verb use + agent -er, but in the legalese sense of 'long-term continuous usage' the whole word is a root as it is a borrowing of the French infinitive user /yze/. What matters then, too, is that in the first sense it's a countable noun, with plural -s, and in the latter it can only be plural in the sense of usage types, and then that this plural marking in absent as a modifier in compounds (username, usernames not *usersnames). Whether the word would be singular "user" or plural "users" in a given sentence, it doesn't matter in "morphology", as that is a matter left to "syntax".

Wikipedia didn't have a link to Hyslop's grammar, but I found it on the website of the Australian National University, so I just added the link to the Wikipedia article now. An example of something syntactic (in the usual definition) that is unrelated to word order in that grammar would be the choice of nominative a vs. accusative na as discussed in section 5.6.

I'm losing my mind over Chinese stroke order! For some characters it's different than in Japanese, so I'm trying to set straight what's correct in Chinese. But different source disagree with one another.

Yes. It's common for native speakers to disagree in a number of characters such as 北 běi 'north', 火 huǒ 'fire' or 母 mǔ 'mother' (the last one regarding the order of the dots vs. the middle horizontal line). I'd say that in such cases it doesn't matter much as long as you broadly follow larger patterns, say, writing the left half of 北 before the right half, or in the case of 母, leaving the dots and the middle horizontal line to the end (regardless of which order you end up using for these three final strokes).

It is also common for native speakers to disregard a number of things about stroke order when doing regular handwriting (informal 行書 xíngshū or Semi-cursive Script) as opposed to fully standard 楷書 kǎishū (Regular Script), even when they are perfectly aware of the standard order. For example, 生 shēng 'be born' is written in its standard Regular Script form as:

1. NW slant
2. top horizontal
3. mid horizontal
4. vertical
5. bottom horizontal

But in regular handwriting it is very common to use instead:

1. NW slant into top horizontal (looking like a kind-of-L)
2. vertical
3. mid horizontal
4. bottom horizontal (forming a "Z" with the previous stroke)

Or if you don't mind a crude Paint drawing made with a mouse:

Which one is it? And if you know of an authorative correct source, can you tell me what are the orders for 比 and 区?

The 现代汉语通用字笔顺规范 Xiàndài Hànyǔ Tōngyòngzì Bǐshùn Guīfàn ("Stroke Order Standards for Modern Chinese Characters in Use", 1997) made by the Chinese government's 国家语言文字工作委员会标准化工作委员会 Guójiā Yǔyán Wénzì Gōngzuò-wěiyuán Biāozhǔnhuà Gōngzuò-wěiyuán ("Standardization Working Committee of the National Language Script Working Committee") says the following:







Note this dictionary does not distinguish between the 提 tí (slanted from SW to NE) and 撇 piě (slanted from NE to SW) strokes (this is because it lists everything by its Wubi composition). However, the greater thickness at the NE end in this dictionary's font, as well as the various Simplified Chinese fonts I have (including Google's Noto Serif SC), show it's a 撇 piě stroke (so drawn from the NE end towards the centre).

China's Baidupedia mentions a difference between Simplified Chinese, Japanese and Korean (hanja) vs. Traditional Chinese here (Taiwan and Hong Kong are second and third from the left):



In the Traditional Chinese form, the SW stroke typically reaches the bottom of the left vertical, instead of touching it at a point about 70% the way down. The Taiwanese form is also shown to be near horizontal in the NE stroke previously discussed to be 撇 piě. The website strokeorder.com.tw (naturally a website from Taiwan) confirms this (also while using the same stroke order as China), and shows China's 撇 piě stroke is actually a 橫 héng (horizontal) stroke in Taiwanese Traditional. I asked someone from Taiwan and they mentioned this horizontal stroke is in fact about 10 degrees upwards, as you can correctly observe on that website, although sometimes a zero-degree horizontal is seen as in Baidupedia's font sample for Taiwan.

[bradrn]

I can't help but find this a very strange way to look at syntax... Also, it's not what I had in mind when I mentioned people who naively discuss syntax in that post, but anyway...

Oh, sorry for misinterpreting you then!

I mean, the common way to look at morphology and syntax, in linguistics at least, sees morphology as contextless. Linguists really just look at a word and the various meanings it has to determine its synchronic morphological structure, its inflections and derivations. So here you are using a different definition of "morphology" and "syntax" altogether. For example, English "user" in the sense of 'person who uses' involves the verb use + agent -er, but in the legalese sense of 'long-term continuous usage' the whole word is a root as it is a borrowing of the French infinitive user /yze/. What matters then, too, is that in the first sense it's a countable noun, with plural -s, and in the latter it can only be plural in the sense of usage types, and then that this plural marking in absent as a modifier in compounds (username, usernames not *usersnames). Whether the word would be singular "user" or plural "users" in a given sentence, it doesn't matter in "morphology", as that is a matter left to "syntax".

I must admit that I’m not quite sure what you’re trying to say here.

Wikipedia didn't have a link to Hyslop's grammar, but I found it on the website of the Australian National University, so I just added the link to the Wikipedia article now. An example of something syntactic (in the usual definition) that is unrelated to word order in that grammar would be the choice of nominative a vs. accusative na as discussed in section 5.6.

How is that a matter of syntax? I wouldn’t normally consider that to be syntactic (although I must admit I’m not quite sure where it fits if not syntax).

[priscianic]

Wikipedia didn't have a link to Hyslop's grammar, but I found it on the website of the Australian National University, so I just added the link to the Wikipedia article now. An example of something syntactic (in the usual definition) that is unrelated to word order in that grammar would be the choice of nominative a vs. accusative na as discussed in section 5.6.

How is that a matter of syntax? I wouldn’t normally consider that to be syntactic (although I must admit I’m not quite sure where it fits if not syntax).

I think this is why thinking of syntax as the question "what do you need to know in order to put together a well-formed sentence from its individual bits?" is useful. Making sure to choose the right case form is something you need to know in order to make a well-formed sentence, and most (if not all) of the time case assignment is not wholly determined by semantics or morphology or phonology. So it's (partly) syntactic. (And I think in most cases, case assignment isn't determined by semantics/morphology/phonology, so most of the time case assignment is wholly syntactic.)

[Kuchigakatai]

Guys, next time someone mentions being marvelled that in some language or other the word for "forehead" is used as the adposition 'in front of', please remember to not only mention that English "front" itself comes from a French word meaning "forehead", but also that English "down" comes from adūne, a reduced form of Old English ofdūne 'down', which comes from the phrase "off the sanddune/hill", used metaphorically as 'away from the height'.

[Travis B.]

Guys, next time someone mentions being marvelled that in some language or other the word for "forehead" is used as the adposition 'in front of', please remember to not only mention that English "front" itself comes from a French word meaning "forehead", but also that English "down" comes from adūne, a reduced form of Old English ofdūne 'down', which comes from the phrase "off the sanddune/hill", used metaphorically as 'away from the height'.

That makes me wonder about the names of groups such as the Front de la liberation nationale and the military use of front, where the meaning of front seemingly was either borrowed back into French after having undergone secondary development in English or borrowed a second time into English after having undergone secondary development in French. (I checked Etymonline, and it did not answer this question.)

[Kuchigakatai]

That makes me wonder about the names of groups such as the Front de la liberation nationale and the military use of front, where the meaning of front seemingly was either borrowed back into French after having undergone secondary development in English or borrowed a second time into English after having undergone secondary development in French. (I checked Etymonline, and it did not answer this question.)

Classical Latin frōns (acc. frontem) is attested as 'forehead' and also 'front (opposed to back)' and the military sense of 'front of an army, battle line' too (and amusingly also 'face (showing a certain feeling)', as a noun for '(someone's) feeling', and 'modesty, shyness, propriety'). Old French just happened to conserve all three meanings when English borrowed the word in the sense of 'front (opposed to back)', and modern French still has them.

The Oxford Latin Dictionary gives an attestation of the military sense in Cato the Elder's Dē Rē Mīlitārī (early 2nd century BC). The Trésor de la langue française gives an attestation in Benoît de Sainte-Maure's Troie, from around 1165, as the oldest attestation of the military sense in Old French.

[Raphael]

Classical Latin frōns (acc. frontem) is attested as 'forehead' and also 'front (opposed to back)' and the military sense of 'front of an army, battle line' too (and amusingly also 'face (showing a certain feeling)', as a noun for '(someone's) feeling', and 'modesty, shyness, propriety'). Old French just happened to conserve all three meanings when English borrowed the word in the sense of 'front (opposed to back)',

As far as I can tell, English borrowed the military sense, too. "They're sending more troops to the front" sounds like a perfectly fine English sentence to me.

[Kuchigakatai]

Funny how English has "fluency" but "influence", "confluence", "affluence". It should really be *fluence, from Latin fluentia. It's as if English speakers looked at "fluentia" and figured it had the Greek suffix -ía that forms abstract nouns, rather than Latin -tia as in perseverantia > perseverance. But then that should be *fluenty, wouldn't it? Maybe it's fluence reinforced by -y after all.

[KathTheDragon]

Note also the pattern of e.g. "profficient", "profficiency".

[Linguoboy]

Funny how English has "fluency" but "influence", "confluence", "affluence". It should really be *fluence, from Latin fluentia.

You can remove the asterisk; fluence exists as a term of art in physics.

The OED has this to say about the usage of the two variants:

< Latin -entia, the termination of abstract nouns formed upon present participles (participial adjectives or nouns) in -ent- by means of the suffix -ia (whence English -y in modesty, fallacy, etc.: see -y suffix³, -cy suffix). The Latin nouns in -entia (like those in -ia generally) denoted primarily qualities or states; but some of them came by development of sense to be nouns of action or process, and in late Latin and in Romanic the formation of nouns of action became the normal function of the suffix. Consequently the English nouns in -ence suffix (which are adaptations of Latin types in -entia either through French or according to French analogies) have very frequently the sense of action or process, either in addition to, or to the exclusion of, that of quality or state. The nouns in -ency, on the other hand, being purely English adaptations of the Latin types, have properly only the sense of quality or state, and concrete senses thence developed. As exemplifying this difference of use between the two suffixes, cf. recurrence and currency, confluence and fluency, residence and presidency. When the same word exists in both the -ence and the -ency forms, the tendency is (where the sense of the verbal etymon permits) to restrict the former to action or process (i.e. to connect its meaning rather with that of the verb than with that of the adjective), while the latter is used to express quality; cf. coherence and coherency, persistence and persistency. In a few instances both forms of a word have equally the sense of quality or condition; in most of these cases the one or the other of the forms has become obsolete or archaic; where they are both in current use, the distinction usually is that -ency has a more distinct reference to the sense of the related adj. or n. in -ent, considered as the predicate of some particular subject; cf. for example, ‘sentience is an attribute of animals’ with ‘some maintain the sentiency of plants’. See -ancy suffix.

With all that in mind, it's really affluence which stands out in this set, since affluent is in common use. But affluence was borrowed directly from Middle French rather than derived from the corresponding adjective.

[bradrn]

Are there any languages with just a simple past/non-past tense distinction? I know Finnish has this system, but I’d be interested to know if there are any others. (And no, I don’t count English as having a non-past tense — English has a future tense, it just isn’t affixed to the verb.)

[Moose-tache]

It's pretty common. The obvious example is Japanese.