Linguistic Miscellany Thread

[Xwtek]

1. abso-fucking-lutely like in English
2. It's good thing I changed my username. Since Akangka means chicken in Chocktaw language.

[Pabappa]

bodacious might have been originally with /ɔ/.... apparently in 1840's rural southern US english, it was seen as a variant of bodyacious.

http://www.as.ua.edu/lavis/handouts/davis.pdf

[Kuchigakatai]

I recently finished reading Steven Dworkin's A Guide to Old Spanish (2018). I think it was a great read, particularly in terms of his attention to where particular phenomena is attested (i.e. in Castile proper, or the more easterly la Rioja, or Aragon, or the Leonese-speaking region...).

One thing that surprised me was that apparently 13th century prose, like that of the General Estoria ("General History"), has a particularly high proportion of VS and VSO sentences, and this might have been an influence of Arabic and maybe also Hebrew on the language.

Another surprise was that he says some uses of ir 'go' + infinitive in 13th and 14th-century Old Spanish might have been perfective pasts, i.e. the meaning Catalan uses for this construction (as opposed to the future tense meaning it has in many Romance languages). I had read the likes of la manol ban besar (Poem of the Cid 298) before, but I had (wrongly?) thought this sort of thing meant 'they go on and kiss his [the Cid's] hand', not the more Catalan-like 'they kissed his hand'.

Another one was that some Latin words with [l:] survived in non-Castilian dialects of medieval Hispano-Romance written as <ll>, e.g. rebellem 'rebel' > reuelle, cellam 'small room' > cella, but Castilian borrowed them as /ld/ [ld] (rebelde, celda). Apparently there's still some evidence in modern Asturian, Leonese and Aragonese (not specified in the book) that such dialects may have had a retroflex consonant here ([ɭ]?, [ɭ:]?, a Sardinian-like [ɖ]!?). Note that the normal Castilian outcome would've been CiellV > CillV, so *rebille and *cilla, cf. sellam 'chair' > OSp. siella > modern silla.



As he says in the introduction, this is the first book published in English introducing Old Spanish, even if it's a mere guide of 100 pages plus some (untranslated) sample texts. I think it would've been desirable to specify that the notation in the phonology section uses the RFE phonetic alphabet, where "[ʂ]" is a dental sibilant (IPA [s̪]) and "[ṡ]" is an alveolar sibilant (IPA [s̠]), instead of assuming English-speaking readers even know what the RFE notation actually is.

He also uses slashes in a lot of places where it would've been better to use square brackets, notably in the "inventory of consonant phonemes" in page 18, where he lists all five of "/dʒ ʒ f φ h/" (the fourth one is IPA "ɸ" of course), even though [φ], [h] and early [f] stand for the same phoneme, and so do [dʒ] and [ʒ]. He also uses slashes when discussing different people's phonetic reconstructions, even if they're of the same phoneme.


(The book contains quite a number of words I didn't know managed to survive from Latin into Old Spanish. I'll post some next time.)

[Darren]

Another one was that some Latin words with [l:] survived in non-Castilian dialects of medieval Hispano-Romance written as <ll>, e.g. rebellem 'rebel' > reuelle, cellam 'small room' > cella, but Castilian borrowed them as /ld/ [ld] (rebelde, celda). Apparently there's still some evidence in modern Asturian, Leonese and Aragonese (not specified in the book) that such dialects may have had a retroflex consonant here ([ɭ]?, [ɭ:]?, a Sardinian-like [ɖ]!?). Note that the normal Castilian outcome would've been CiellV > CillV, so *rebille and *cilla, cf. sellam 'chair' > OSp. siella > modern silla.

In Asturian-Leonese, ḷḷ is used to represent /ts ~ ɖʐ ~ ɖ ~ ʈʂ/, which occurs in the western dialects instead of palatal /ʎ/ (e.g. ḷḷingua, parḷḷar for standard llingua, parllar).

[Pabappa]

Do we know why Spanish abolir was defective for so long? (Apparently, it no longer is considered such even by Real Academia Española ... they say just that it is more commonly used in the forms with /i/). Is it just because the conjugated forms would have collided with abuelo, abuelas, abuela etc? Im guessing thats not even it, because the forms that are used today all have a pure /o/, never /ue/. Also i just noticed the verb ending is -ir, so that makes even less sense. Are there any other defective verbs that arose for no clear reason ... or at least a reason that isnt obvious to us today?

[Whimemsz]

abso-fucking-lutely like in English

Yeah, I actually considered mentioning the English "fucking" [etc.] infix construction. I will note that in English it's an infix and in Unami it's just an incorporated noun that goes into the available slot for incorporated nouns within Unami verbs. (Which is interesting, but...honestly I guess the English case is actually more interesting since English rarely uses any kind of infixation, and infixation is rare in general...?)

It's good thing I changed my username. Since Akangka means chicken in Chocktaw language.

Heh. Did you already know that or just look it up now?

[äreo]

Do we know why Spanish abolir was defective for so long? (Apparently, it no longer is considered such even by Real Academia Española ... they say just that it is more commonly used in the forms with /i/). Is it just because the conjugated forms would have collided with abuelo, abuelas, abuela etc? Im guessing thats not even it, because the forms that are used today all have a pure /o/, never /ue/. Also i just noticed the verb ending is -ir, so that makes even less sense. Are there any other defective verbs that arose for no clear reason ... or at least a reason that isnt obvious to us today?

On a related note, do we know quite how the deponent verbs of ancient IE languages, particularly of Latin, came about?

[Pabappa]

Wiktionary seems to assign verbs of the type that only use forms with /i/ to the "aterir" class. I wonder if it was once called the "abolir" class. I dont have my old paperback Spanish dictionary which listed all those verbs, but Wiktionary only lists 2 such verbs, along with nine other defectives that seem mostly to be semantic (e.g. *diluvio "I rain", etc). Anyway this doesnt give me much of a clue as to what caused the non-semantic defectives to become this way.

[Xwtek]

It's good thing I changed my username. Since Akangka means chicken in Chocktaw language.

Heh. Did you already know that or just look it up now?

I just looked it up when you mention that the Chocktaw for turkey was fakit.

[Kuchigakatai]

In Asturian-Leonese, ḷḷ is used to represent /ts ~ ɖʐ ~ ɖ ~ ʈʂ/, which occurs in the western dialects instead of palatal /ʎ/ (e.g. ḷḷingua, parḷḷar for standard llingua, parllar).

!

Do we know why Spanish abolir was defective for so long? (Apparently, it no longer is considered such even by Real Academia Española ... they say just that it is more commonly used in the forms with /i/). Is it just because the conjugated forms would have collided with abuelo, abuelas, abuela etc? Im guessing thats not even it, because the forms that are used today all have a pure /o/, never /ue/. Also i just noticed the verb ending is -ir, so that makes even less sense. Are there any other defective verbs that arose for no clear reason ... or at least a reason that isnt obvious to us today?

Okay, this post I wrote below got long, but hey, at least I had the decency to put all the off-topic stuff at the bottom in a footnote.

Spanish, French and probably the rest of Romance languages have a number of uncommon verbs, some inherited and some learnèd, that are not used aside from the infinitive (whether as intransitives or transitives, whether as absolutes or reflexives) and the past participle.

Sometimes, these verbs become so uncommon the infinitive use goes away, and as the verb survives purely as a past participle, it's best to now consider it transformed into an adjective or sometimes a noun. Rarely, it's the infinitive that survives as a noun. This happened recently to empedernir(se) (< in-petrĭn-īre, from petrĭnus 'made of stone'), which in the 19th century some people still used with the meaning '(tr.) to make sth solid; (intr. refl.) to get stubborn, insist' with the i-defective pattern of abolir/aterir that you mention, other speakers using it only in the infinitive and the past participle. Nowadays it's just an adjective meaning "hopeless/incurable (person with a bad habit)", especially in the phrase borracho empedernido 'heavy incurable drunkard'. Note that the DRAE lists this verb with the i-defective pattern, but I've only encountered it as an adjective.

Sometimes, these verbs actually increase in frequency, so they gain normal behaviour. This has happened recently to agredir, which many 20th century works list with the i-defective pattern, and which is still defective like that for some people, but I personally have no problem with the likes of yo agredo, yo agrederé, etc. The RAE has no problem with these new forms either.

The i-defective pattern is basically a behaviour that uncommon -ir verbs sometimes get among some speakers in the middle between these two poles (death and normal vitality), where speakers remember the canonical unstressed stem but are not comfortable putting the stress on the stem.* This has a reasonable phonological origin and motivation in -e-/-o- stems because speakers don't know the stressed stem ("does agredir become yo agredo or yo agriedo?"), but the pattern is also applied to stems with other vowels (as in the speakers who use garantir as nosotros garantimos, but are not comfortable with yo garanto).

The fact that all such i-defective verbs are uncommon and can be easily replaced by some other verb is important. Derogar takes over abolir, atacar and asaltar do so over agredir, and garantizar over garantir.

So, for -ir verbs, we can thus posit a spectrum between the two poles:
1. normal vitality
2. i-defective but used in all tenses
3. i-defective and used only in the infinitive, past participle and the imperfect past
4. used only in the infinitive and the past participle
5. death (may or may not survive with the adjective converted into an adjective or noun, or rarely the infinitive into a noun)

The thing that happens in abolir 'abolish' and other defective verbs is that not all Spanish speakers agree on their location in the spectrum. I think it's fine if any given native speaker is willing to say yo abolo, but I don't say that, because for me it's in position #2. Maybe if I started hearing this verb used a lot as it were a normal verb in position #1 (yo abolo, tú aboles...) then I'd get comfortable with the stressed stem, but so far I'm not.

On the other hand, as I said, agredir 'attack' is in position #1 for me, but is still in #2 for other people. Desvaír is in #3 for me, but I'm sure it's in #4 for many speakers. Some man on WordReference thinks compungir is in #4, but for me it's in #5, pretty much dead. The same man was responding to a post where he was told some conjugation guide says manir is in #2, but for him (and me) it's in #5. He also thinks aterirse is in #3, but for me it's in #4 very much verging on #5.


* In contrast, -ar and -er verbs tend to be entirely lost at once, leaving only the infinitive or a participle behind. This happened to manjar 'feast; culinary delicacy', originally the infinitive of a verb meaning 'to eat' (cognate to French manger). The old verb redar 'to throw nets (to catch fish)' (< rēt-āre, from Latin rēte 'net') has now only left its past participle as the abstract action noun redada 'police raid'.

As a total off-topic digression, I imagine redada was used and kept due to its similarity to English raid < revived Scottish archaism from Old English rād. Compare this with the use of the common word arreglo 'a fix, repair; beautiful arrangement (of flowers, furniture, etc.)' (< ad-regul-um_{4th decl.} < Latin regula 'ruler, rule') with the meaning 'array' when talking about C, C++, Java, PHP and other programming languages, due to its similarity in sound with English "array" (< Anglo-Norman arraier < medieval Latin ad- + "red" + -iāre, where the stem "red" < Germanic raidaz).

This kind of semantic expansion influenced by the sound of a foreign word is rare and actively avoided in Spanish. There are many programmers who hate this use of arreglo and insist in saying vector/matrix or even a straight-forwardly borrowed el array, and they do it arguing that arreglo has been merely chosen for its similarity to English array.

Funnily, Mandarin speakers have no problem with this kind of strategy, e.g. 酷 kù 'strong (said of wine)' > 'cool', 粉丝 fěnsī 'mungbean noodles' > '(a pop idol's) fans'. Although it's true that more commonly they create phonetically-similar meaningful novel compounds, like 愛 ài 'love' + 豆 dòu 'bean' > 愛豆 àidòu '(pop) idol', or 基 jī 'base, foundational' + 因 yīn 'cause, reason' > 基因 jīyīn 'gene'.

[Kuchigakatai]

Gyawdzss, and I just noticed I didn't say anything about how that pattern even came about. It has to do with the analogy that -ir verbs with i/u in the stem caused in verbs with any of e/ie/o/ue in the stem. This effectively merged all those stem vowels into an alternation of e/i or simply the vowel -u-.

In other words, verbs like these:

dīcere dīcō, rīdēre rīdeō > OSp. dezir digo, reír río
condūcere condūcō, lūcēre lūceō > OSp. conducir condusco, OSp. lucir, lusco

influenced verbs like these:

mētīrī mētior > medir mido
servīre serviō > OSp. servir siervo > servir sirvo
fugīre fugiō > OSp. foír fuyo > huir [u.ˈiɾ] huyo
complēre compleō > OSp. complir cumplo > cumplir cumplo
cooperīre cooperiō > OSp. cobrir cubro > cubrir cubro

So the modern language has those forms instead of the expected *medo, *siervo, *hoyo, *complo and *cuebro. As you can see, Old Spanish still retained some expected vowels, but this process was complete by the 17th century. Only five extremely common verbs managed to escape this: venir 'to come', sentir 'to feel', oír 'to hear', dormir 'to sleep' and morir 'to die', all of which kept stem diphthongs (or a weird stem in the case of oír). Podrir/pudrir 'to go bad, rot' optionally retains an unstressed -o-.

This effectively means that when there were more recent attempts to borrow Latin verbs like agredir and abolir, these new verbs had no conjugation model to hold onto. Verbs borrowed from Latin generally keep their stem vowel unmodified, but Spanish speakers unconsciously felt and still feel very uneasy about having -ir verbs with patterns like agredir agredo and especially abolir abolo. And so we get the i-defective pattern we all know and love.

"Urgh, agredir, abolir. These words are very uncommon and look learnèd, so they should keep the thing intact like all good uncommon verbs, so... agredo/abolo. But maaan, I don't know of any verb that does this. Should I say agriedo/abuelo like I say siento/duermo, or agrido/abulo like I say mido/cumplo? Maaan, this is sooo weird..."

[Travis B.]

One question I have is how do the modern orthographies for standard German and Dutch represent final "voiced" versus "voiceless" consonants in positions where there is no alternation of voicing (i.e. no vowel-initial affixes attached finally), when neither written Middle High German nor written Middle Dutch represented these consonants as such (i.e. they always marked them as voiceless except when vowel-initial affixes wre attached)? Did they compare the consonants against other Germanic languages without final devoicing to determine what these consonants "should" be? (I presume they did not compare against Old High German or Old Dutch, as the former has only limited attestation and the latter has almost no attestation at all.)

(In other words, why is und "und" and Kant "Kant", and not "unt" or "Kand"?)

[Pabappa]

I think in most cases the original voiced stops survived in inflections and thus the speakers knew which ones were which. But there are at least some words where the change was generalized, perhaps most commonly in names where there were no commonly used inflected forms. und might have retained a voiced stop when used word-internally.

edit:sorry, i misread your post.... you more or less already know what I wrote above, but .... even so, there may be some cases where vowels came to follow the word in question that did not involve the traditional -es/-er/-en type of inflections. Consider, for example, diminutives attached to names.

[Linguoboy]

(In other words, why is und "und" and Kant "Kant", and not "unt" or "Kand"?)

The most common Middle High German form of und was unde.

Kant has the variant form Kante; it's a Low German word meaning "edge, coast". (The ultimate etymon is Proto-Celtic *kantos; the progression was most likely PC > Latin > Old French > Middle Dutch > MLG > MHG.)

[Travis B.]

(In other words, why is und "und" and Kant "Kant", and not "unt" or "Kand"?)

The most common Middle High German form of und was unde.

I should have suspected this considering its cognate in Midde Dutch was ende.

[Zju]

Points all taken. I think the conservative orthographies of modern Germanic or Romance languages obscure the extent to which they've really diverged (much easier to connect <thought> and <dachte> than [θɔʔ] and [däxtɛ]). Let's stick to two millennia.

Is [θɔʔ] really one of the more common realisations of <thought>? I θɔt θɔt was way more common.

[Vijay]

Points all taken. I think the conservative orthographies of modern Germanic or Romance languages obscure the extent to which they've really diverged (much easier to connect <thought> and <dachte> than [θɔʔ] and [däxtɛ]). Let's stick to two millennia.

Is [θɔʔ] really one of the more common realisations of <thought>? I θɔt θɔt was way more common.

Yes. Releasing the stop is not so common.

[Zju]

Isn't an unreleased stop just [t̚]?

[Vijay]

Sure, but a lot of people have [ʔ] instead of [t̚].

[Travis B.]

Sure, but a lot of people have [ʔ] instead of [t̚].

I am used to [ʔ] as the realization of /t/ here unless followed by a vowel, where then it becomes [ɾ] or even ∅.