Neo-grammarianism

[Travis B.]

I was recently reading a paper on the vocalic historical phonology of Hebrew, and one thing I noted was that the author of the paper went out of their way to explicate with Neo-grammarianism was Right, and how all sound changes were through either sound laws or processes such as analogy, borrowing including dialect borrowing, and contamination. In cases where these do not seem to firmly apply, supposedly the reason is Not Enough Sound Laws. They did this to make a point of how the traditional analyses of Semitic languages were Wrong, where traditionally sound change is not seen as in terms of firm sound laws and like.

This got me thinking, and some points came to my mind:

• How stressed sounds, morphemes, and words typically are can affect their sound change.
• How frequently sounds, morphemes, and words typically are used can affect their sound change.
• What environments sounds, morphemes, and words typically are used in can affect their sound change.
• What morphological and syntactic contexts sounds, morphemes, and words typically are used in can affect their sound change. (Of course, some of these cases could be chalked up to analogy.)

With these in the mind, the idea that all sound change can be chalked up to sound laws, analogy, borrowing, and contamination quickly breaks down, and the "solution" of adding more sound laws ends up creating a wide menagerie of sound laws to cover not just every sound and every environment, but every stress level, every frequency, and every morphological and syntactic context, which of course is non-viable in practice.

Any thoughts?

[Raphael]

I was recently reading a paper on the vocalic historical phonology of Hebrew, and one thing I noted was that the author of the paper went out of their way to explicate with Neo-grammarianism was Right,

Sounds to me like the author made a presumably informed guess about the positions on linguistic theory of the people who would decide whether to publish the paper.

[Travis B.]

Oh, I missed one thing that the author specified as a source of sound change - folk etymology.

[zompist]

This stuff has been talked about for over a hundred years, and even elementary textbooks on historical linguistics will cover some of the difficulties.

The lesson of the Neo-Grammarians shouldn't be taken as "100% of sound changes are regular and complete", but "Don't give up on regularity too easily." There are plenty of examples where apparent irregularities were explained by a more careful analysis.

The Neo-Grammarians hadn't read Labov. We now know a lot more about how sound changes happen, and they do not proceed all at once, affecting every word in every environment. They start with certain words and spread. Historically we see completed or mostly-completed sound changes, so they look regular. But it's not guaranteed that every word that could be affected actually was.

[Travis B.]

This stuff has been talked about for over a hundred years, and even elementary textbooks on historical linguistics will cover some of the difficulties.

The lesson of the Neo-Grammarians shouldn't be taken as "100% of sound changes are regular and complete", but "Don't give up on regularity too easily." There are plenty of examples where apparent irregularities were explained by a more careful analysis.

The Neo-Grammarians hadn't read Labov. We now know a lot more about how sound changes happen, and they do not proceed all at once, affecting every word in every environment. They start with certain words and spread. Historically we see completed or mostly-completed sound changes, so they look regular. But it's not guaranteed that every word that could be affected actually was.

This is exactly what I thought of, and that is why I thought it was remarkable that a recent author would choose Neo-grammarianism as their hill to die on.

[Travis B.]

The lesson of the Neo-Grammarians shouldn't be taken as "100% of sound changes are regular and complete", but "Don't give up on regularity too easily." There are plenty of examples where apparent irregularities were explained by a more careful analysis.

A good example that comes to mind is θ > ð / #_V in English. This is a sound change that simply cannot be analyzed as a regular sound law. Rather, it is sensitive to stress and word frequency, which is why it is rare in English from a dictionary standpoint and extremely common in English from a corpus standpoint, as you must know.

[Ahzoh]

Historically we see completed or mostly-completed sound changes, so they look regular. But it's not guaranteed that every word that could be affected actually was.

So sound changes can have a patient zero?

[Emily]

The lesson of the Neo-Grammarians shouldn't be taken as "100% of sound changes are regular and complete", but "Don't give up on regularity too easily." There are plenty of examples where apparent irregularities were explained by a more careful analysis.

A good example that comes to mind is θ > ð / #_V in English. This is a sound change that simply cannot be analyzed as a regular sound law. Rather, it is sensitive to stress and word frequency, which is why it is rare in English from a dictionary standpoint and extremely common in English from a corpus standpoint, as you must know.

but stress is a phonological characteristic that is very frequently a factor in sound changes even under neogrammarian frameworks, and the general explanation around these words is that they were typically found in an unstressed position, so their change to /ð/ can indeed be demonstrated by essentially regular sound change laws. i think a clearer example is the stress shift pattern in words like record, project, etc. i don't remember which way it went, but it was a sound change that was morphologically conditioned rather than on any particular phonetic environment

[Travis B.]

The lesson of the Neo-Grammarians shouldn't be taken as "100% of sound changes are regular and complete", but "Don't give up on regularity too easily." There are plenty of examples where apparent irregularities were explained by a more careful analysis.

A good example that comes to mind is θ > ð / #_V in English. This is a sound change that simply cannot be analyzed as a regular sound law. Rather, it is sensitive to stress and word frequency, which is why it is rare in English from a dictionary standpoint and extremely common in English from a corpus standpoint, as you must know.

but stress is a phonological characteristic that is very frequently a factor in sound changes even under neogrammarian frameworks, and the general explanation around these words is that they were typically found in an unstressed position, so their change to /ð/ can indeed be demonstrated by essentially regular sound change laws. i think a clearer example is the stress shift pattern in words like record, project, etc. i don't remember which way it went, but it was a sound change that was morphologically conditioned rather than on any particular phonetic environment

The thing, though, is that sound changes like the one I referenced aren't simply a matter of "this sound change happens in an unstressed syllable", because the words that are affected, aside from the, commonly are stressed -- they just are slightly less stressed on average than unaffected words, they are more common on average than unaffected words, and they are almost invariably grammar words as opposed to unaffected content words. Hence you cannot form a clear rule like one you could feed into an SCA that could generate the output of this sound change across the entire vocabulary.

[zompist]

Historically we see completed or mostly-completed sound changes, so they look regular. But it's not guaranteed that every word that could be affected actually was.

So sound changes can have a patient zero?

Patients zero, but yeah, pretty much. Labov even identified which sociological groups are likely to spread sound changes.

[jcb]

Historically we see completed or mostly-completed sound changes, so they look regular. But it's not guaranteed that every word that could be affected actually was.

So sound changes can have a patient zero?

Patients zero, but yeah, pretty much. Labov even identified which sociological groups are likely to spread sound changes.

Which groups did he identify?

[zompist]

Labov even identified which sociological groups are likely to spread sound changes.

Which groups did he identify?

Young women, lower working class, with high social connectivity.

Here's an article that mentions Labov but focuses on more modern research.

(You might ask: does that mean young men pick up new speech habits from young women? No... but young boys pick them up from their mothers. This has been traced in Argentina: in the 1960s /j/ (orthographic y or ll) was pronounced [ʃ] by women, [ʒ] by men. A generation later, both sexes had [ʃ].)

[xxx]

Young women, lower working class, with high social connectivity.

does this mean that conlangers don't influence languages...

[zompist]

Young women, lower working class, with high social connectivity.

does this mean that conlangers don't influence languages...

First, the question was about sound changes, not "influencing languages."

Second, you seem to be making some strange assumptions about the demographics of conlangers.

[dɮ the phoneme]

A good example that comes to mind is θ > ð / #_V in English. This is a sound change that simply cannot be analyzed as a regular sound law. Rather, it is sensitive to stress and word frequency, which is why it is rare in English from a dictionary standpoint and extremely common in English from a corpus standpoint, as you must know.

but stress is a phonological characteristic that is very frequently a factor in sound changes even under neogrammarian frameworks, and the general explanation around these words is that they were typically found in an unstressed position, so their change to /ð/ can indeed be demonstrated by essentially regular sound change laws. i think a clearer example is the stress shift pattern in words like record, project, etc. i don't remember which way it went, but it was a sound change that was morphologically conditioned rather than on any particular phonetic environment

The thing, though, is that sound changes like the one I referenced aren't simply a matter of "this sound change happens in an unstressed syllable", because the words that are affected, aside from the, commonly are stressed -- they just are slightly less stressed on average than unaffected words, they are more common on average than unaffected words, and they are almost invariably grammar words as opposed to unaffected content words. Hence you cannot form a clear rule like one you could feed into an SCA that could generate the output of this sound change across the entire vocabulary.

This could be accounted for by distinguishing word stress from syllable stress—which one should do anyway, because word stress is an intonational phenomenon and syllable stress isn't—and analogy. θ > ð / #_[+syl, -cons, -word stress] (or whatever) and then an analogical shift of stressed words to match their unstressed counterparts. That latter stage might be expected to be irregular. In fact, something like that is currently happening with the vowel in the. There is an unstressed form /ðə/ (/ðiː/ before a vowel) and a stressed form /ðiː/, which both coexist in the language. And I believe that vowel reduction in unstressed words is a regular process, although I might be wrong about this. Either way, people are starting to use /ðə/ in all circumstances for the, and this is certainly an analogical and not a phonological shift.

So, what you're talking about can be explained in a Neogrammarian framework. Should it? I don't know. But it can.

[xxx]

First, the question was about sound changes, not "influencing languages."
Second, you seem to be making some strange assumptions about the demographics of conlangers.

their lack of influence would only be proven in phonology...
as for the sociology of conlangers, a real study would be needed
to refute the perception of a community of young male graduates...