Linguistic Miscellany Thread

[bradrn]

Are these kinds of vowels attested in natlangs, and if so, what do people usually call them, and how do they get analysed?

IIRC Estonian has something similar - cf. here. Though I'm not aware if those gen. sg. stem-final vowels are denoted by any special term.

Thanks! I’ll look into this further when I get time.

[Glenn]

The term I use is stem vowel, which I think has less theoretical baggage than thematic vowel. It fits nicely into the same metaphor as "root". Though I guess we don't call the endings "leaves".

Now I want to see a conculture whose linguists do call the endings "leaves," and where all of the remaining grammatical terminology is plant-based as well. (Maybe phonemes could be "seeds," and morphemes could be "sprouts"? )

[chris_notts]

The term I use is stem vowel, which I think has less theoretical baggage than thematic vowel. It fits nicely into the same metaphor as "root". Though I guess we don't call the endings "leaves".

Now I want to see a conculture whose linguists do call the endings "leaves," and where all of the remaining grammatical terminology is plant-based as well. (Maybe phonemes could be "seeds," and morphemes could be "sprouts"? )

And sentences are thickets?

[Travis B.]

I like this terminology.

[Rounin Ryuuji]

This might actually be just what I needed...

[abahot]

Quick question that I've been wondering about. So there's the concept of a "principal part" in Latin or Ancient Greek grammar, for example, where each verb has multiple stems that need to be known to conjugate the verb in all forms. Is this feature of a grammar, where each word has multiple unpredictable stems, known in many non-Indo-European languages? The only example I can think of is Navajo where verbs seem to have some insane alternations, but where else is this found?

[bradrn]

Is this feature of a grammar, where each word has multiple unpredictable stems, known in many non-Indo-European languages?

Oh, yes. The best example I can think of off the top of my head is Komnzo, but there are many others — stem alternation isn’t rare.

[Kuchigakatai]

There is apparently a [x] vs. [ç] minimal pair in German for many speakers: Aachen vs. Archen [aːxn̩ aːçn̩].

Quick question that I've been wondering about. So there's the concept of a "principal part" in Latin or Ancient Greek grammar, for example, where each verb has multiple stems that need to be known to conjugate the verb in all forms. Is this feature of a grammar, where each word has multiple unpredictable stems, known in many non-Indo-European languages? The only example I can think of is Navajo where verbs seem to have some insane alternations, but where else is this found?

Learners of Standard Arabic learn three principal parts for every verb, perfect (~past), imperfect (~nonpast) and the verbal noun. Learners of Ge'ez learn two per verb, perfect and jussive. Learners of Classical Tibetan learn four per verb. They just don't call them "principal parts".

[hwhatting]

There is apparently a [x] vs. [ç] minimal pair in German for many speakers: Aachen vs. Archen [aːxn̩ aːçn̩].

Well, phonemicity always depends on the speech variety you look at; so yes, for non-rhotic German speakers, [x] vs. [ç] is a minimal pair.
For rhotics, here is also Kuchen "cake" vs. Kuhchen "diminutive of Kuh (cow)" [kuːxn̩] vs. [kuːçn̩], and with some imagination you can come up with more, probably. This is an edge case, as the diminutive suffixes -chen and -lein normally trigger umlaut (i.e. Kühchen), but the umlaut-less form exists in some registers, and for those [x] vs. [ç] would again be minimal pairs.

[Zju]

Isn't [ç] marginally phonemic regardless, since it appears -chen even after back vowels?

[WeepingElf]

Isn't [ç] marginally phonemic regardless, since it appears -chen even after back vowels?

Read hwhatting's post. He has answered your question: the suffix causes umlaut, i.e. back vowels are fronted, such that it never follows a back vowel. However, this rule is on its way out with colloquial forms such as Frauchen, and that means that [ç] and [x] are on their way to becoming distinct phonemes.

[Moose-tache]

Is this feature of a grammar, where each word has multiple unpredictable stems, known in many non-Indo-European languages?

Oh, yes. The best example I can think of off the top of my head is Komnzo, but there are many others — stem alternation isn’t rare.

Navajo should be a clue here. All the Na Dene languages are like this, as are many other Native American languages in the Iroquoian and Muskogean families.

[bradrn]

I’ve been meaning to share this very interesting thesis: A Concatenative Analysis Of Diachronic Afro-Asiatic Morphology. I don’t know enough about Afro-Asiatic to completely assess it, but it’s certainly fascinating.

[chris_notts]

I’ve been meaning to share this very interesting thesis: A Concatenative Analysis Of Diachronic Afro-Asiatic Morphology. I don’t know enough about Afro-Asiatic to completely assess it, but it’s certainly fascinating.

I enjoyed it!

Another fun language fact: A cluster of languages in North Africa have a "no case before the verb" rule, which means that ergative or marked nominative marking is lost when an NP is fronted for topic and/or focus marking purposes. The origin in many cases seems to be rigid verb initial + clefts with gapping in relative clauses for focus, with the copula or relative clause marker lost. I.e. the following with acc being zero marked:

Unmarked:
V S-nom/erg O-acc

Contrastive A/S focus:
COP S-acc REL V O-acc -> S-acc V O-acc

For the topic case I guess it's clear, since detached topics are not really part of the clause they're associated with, so a lack of case marking in that case makes some kind of sense even if the construction becomes more grammaticalised.

[chris_notts]

Another fun language fact: A cluster of languages in North Africa have a "no case before the verb" rule, which means that ergative or marked nominative marking is lost when an NP is fronted for topic and/or focus marking purposes. The origin in many cases seems to be rigid verb initial + clefts with gapping in relative clauses for focus, with the copula or relative clause marker lost. I.e. the following with acc being zero marked:

Unmarked:
V S-nom/erg O-acc

Contrastive A/S focus:
COP S-acc REL V O-acc -> S-acc V O-acc

For the topic case I guess it's clear, since detached topics are not really part of the clause they're associated with, so a lack of case marking in that case makes some kind of sense even if the construction becomes more grammaticalised.

I found another example. Semelai, a Mon-Khmer language, fronts NPs for topic marking and core case marking (e.g. recipient, agent) is always lost in this position. Obliques retain their prepositions.

[Raphael]

A recent short discussion in conlangernoob's Proto-Langs thread in the Conlangery forum made me wonder: does the term "proto-language" mean only a reconstructed ancestral language, or are historically documented languages that became ancestor languages of later language families, such as Latin or Sanskrit, also seen as proto-languages?

[WeepingElf]

I tend to avoid the term "proto-language" in this sense because it means something completely different in language origin studies (which many non-specialists don't realize that it is a different discipline from historical linguistics, which results in various misunderstandings). I prefer to speak of "common ancestor languages", which also covers attested languages of this kind (e.g. Old Irish, an attested language, is the common ancestor of Modern Irish and Scots Gaelic). However, using "Proto-X" for the reconstructed common ancestor language of a family is an established custom, though saying "Common X" would be nicer. (Attested common ancestors usually have names of their own, so such naming conventions are unnecessary there and are not used, though some scholars say such things as "Proto-Romance" to emphasize that the common ancestor of Romance was a different register of the Latin language than the kind of Latin that is taught in grammar schools.)

[Raphael]

Thank you!

[hwhatting]

However, using "Proto-X" for the reconstructed common ancestor language of a family is an established custom, though saying "Common X" would be nicer.

"Common X" is often used differently, for developments that are common to all languages of a family, but have spread after the proto-stage. "Common Slavic" is regularly used that way.

[WeepingElf]

However, using "Proto-X" for the reconstructed common ancestor language of a family is an established custom, though saying "Common X" would be nicer.

"Common X" is often used differently, for developments that are common to all languages of a family, but have spread after the proto-stage. "Common Slavic" is regularly used that way.

I understand; you are right.