Linguistic Miscellany Thread

[Linguoboy]

Middle Irish > Modern Irish.

As mentioned elsewhere, the Middle Irish spelling <ḋ>/<dh> represents a lenition of earlier /d/ and is reconstructed as /ð/. At some point, this fell together with /ɣ/ from earlier /g/, which in non-initial position was either deleted or formed a diphthong with the neighbouring vowels.

So, for instance, Old Irish gadar "hunting dog" (< ON gagarr) > Modern Irish gadhar [ɡəiɾˠ], presumably via a Middle Irish */gaðˠəɾˠ/.

[Nortaneous]

And is ð > j an attested sound change anywhere? I recall seeing something about it having happened in Faroese.

j > ð is attested, so the reverse seems reasonable

[Linguoboy]

And is ð > j an attested sound change anywhere? I recall seeing something about it having happened in Faroese.

j > ð is attested, so the reverse seems reasonable

Except that there are a lot of changes which seem to be most commonly if not exclusively unidirectional.

Two that come immediately to mind are /r/ > /h/ (Brazilian Portuguese) and /lː/ > /ʃ/. Is the reverse of these attested anywhere?

[axolotl]

And is ð > j an attested sound change anywhere? I recall seeing something about it having happened in Faroese.

Correct, it is present in Faroese, and the change is well-reflected in its etymological orthography:

Faroese: gleði [klɛjɛ~klɛjɪ]
Icelandic: gleði [kle̞ðɪ]


There are also some dialects of Danish that realize the "soft d" as [j] or something close.

[MacAnDàil]

As this is a sound change common to Goidelic languages and Scandinavian ones, it is perhaps no concidence. Other features, such as the pitch accent in the north or the glottal stop further south seem to be common to these neighbouring language groups.

[akam chinjir]

If a language resolves vowel hiatus with epenthetic glides, and uses both j. or w for this, what's more likely (assuming the vowels are i e a o u)?

• You get w before o u, otherwise j.
• You get w after o u, otherwise j.
• You get w if either flanking vowel is o u, otherwise j.
• Something else.

(My instincts strongly favour the second choice, but I think I tend to syllabify intervocalic glides with the preceding consonant, and I'm worried that might be an Englishy habit.)

[Kuchigakatai]

I'd agree with your instinct really.

A couple examples of languages where it's not followed would be the occasional use of [w] in Arabic to solve [a:] + [i:]: faransā 'France' > faransāwī 'French' (the language otherwise uses [j] for [a:] + [a:]), and Old French which in at least one occasion used the w-like [v] to solve [ i] + [ə]: jūdaea > *[dʒuˈði.a] > *[dʒuˈi.ə] > Old French juive [dʒuˈivə] (normally this doesn't happen though: Marie [maˈri.ə], Germanic krītan(an) 'to shout ("cry")' > pre-French 3SG *[ˈkriðəθ] > Old French 3SG crie [ˈkri.ə]).

[akam chinjir]

Thanks! Maybe I'll just go with what sounds right to me, for the general pattern anyway.

I like the idea of having the result sometimes be an unpredictable w. (Standard Average Akam has lots of w.)

[Kuchigakatai]

Oh, I should've mentioned it, Arabic normally deletes the first vowel when dealing with [a:] + [i:] (I mentioned the use of [j] in [a:] + [a:], a different context), so e.g. the adjective faransāwī also exists as faransī.

[MacAnDàil]

I'd agree with your instinct really.

A couple examples of languages where it's not followed would be the occasional use of [w] in Arabic to solve [a:] + [i:]: faransā 'France' > faransāwī 'French' (the language otherwise uses [j] for [a:] + [a:]), and Old French which in at least one occasion used the w-like [v] to solve [ i] + [ə]: jūdaea > *[dʒuˈði.a] > *[dʒuˈi.ə] > Old French juive [dʒuˈivə] (normally this doesn't happen though: Marie [maˈri.ə], Germanic krītan(an) 'to shout ("cry")' > pre-French 3SG *[ˈkrīðəθ] > Old French 3SG crie [ˈkri.ə]).

Perhaps the possibility of having three vowels in hiatus induced the word 'juive' to act differently from 'Marie' and 'crie'.

[Kuchigakatai]

Perhaps the possibility of having three vowels in hiatus induced the word 'juive' to act differently from 'Marie' and 'crie'.

Oh yeah, maybe. I hadn't thought of that (places where I've seen juive mentioned don't give any explanation for the [v] either).

[quinterbeck]

A fun short video: Interview with an Anglo-Saxon in Old English

[Znex]

How common are rhotic trills vs. flaps crosslinguistically (where rhotics are distinct from laterals)? And where they pattern allophonically, in what environments does one tend to be more common than the other? I'm not personally familiar with any languages that prefer a rhotic trill over a flap.

[Kuchigakatai]

How common are rhotic trills vs. flaps crosslinguistically (where rhotics are distinct from laterals)? And where they pattern allophonically, in what environments does one tend to be more common than the other?

To really answer your question, we'd need to look into the phonetics of a large number of languages which is not at all an easy thing to do (good phonetic descriptions of this sort generally lacking, as opposed to phonemic ones). In general I'd say the distinction is common, with the caveat that it often surfaces as a geminate trill /r:/ vs. a simple flap /ɾ/, because such languages have geminates like /p:/ and /m:/ elsewhere.

Here is some data from some famous languages:

Spanish has /r/ vs. /ɾ/. The two only contrast in intervocalic position. In word-initial position, only the trill /r/ is admissible, and in word-final position, only the flap /ɾ/ is admissible. Combining these two facts, the use of the trill or flap can determine word boundaries: paremos /paˈɾemos/ [paˈɾemos] 'let's stop' vs. pa' remos [de cierto tipo] /paˈremos/ [paˈremos] 'for oars [of some sort]'. (Not knowing this is a classic, extremely common mistake among students of Spanish, who tend to say e.g. estar aquí with a trill, *[esˈtar aˈki], instead of the correct flap.)

Furthermore, in syllable-final position (once you include the next word's syllable too), only /ɾ/ is admissible. However, this /ɾ/ can be either [r] or [ɾ], and as far as I can informally tell, the trill [r] is preferred more often than not, although this trill is usually shorter than the word-initial one. Arpa /ˈaɾpa/ [ˈarpa ˈaɾpa] 'harp', por demás /poɾ deˈmas/ [pordeˈmas poɾdeˈmas] 'needless'.

Word-medially, they contrast in intervocalic position: era /ˈeɾa/ [ˈeɾa] 'era', erra /ˈera/ [ˈera] 'he/she errs'. As a glide-like approximant in onsets, only the flap /ɾ/ is allowed: premio /ˈpɾemjo/ [ˈpɾemjo] 'prize'. As the onset of a syllable after another consonant, only the trill [r] is admitted: honra [ˈonra] 'honour'.

(Please note that when discussing Spanish, as above, it is important to have it clear in mind that a rhotic can be word-final and syllable-final, as in tener poco /teˈneɾ ˈpoko/ [teˈner ˈpoko, teˈneɾ ˈpoko] 'to have it' (with its short trill or alternatively a flap), or word-final and syllable-initial, as in estar aquí /esˈtaɾ aˈki/ [esˈtaɾ aˈki] (with its flap).)

In Italian, there is /r:/ and /r/. While /r:/ is always a geminate trill, /r/ can be either a trill or a flap depending on whether any vowel next to it is stressed or not. If the /r/ is next to a stressed vowel, whether before or after, it may be pronounced either as a trill or a tap, the tap being more common in my experience. Andare a rotoli /anˈdare a ˈrotoli/ [anˈdaːre‿a ˈrɔtoli, anˈdaːɾe‿a ˈɾɔtoli] 'go up in flames, fall apart badly', prato /ˈprato/ [ˈpraːto ˈpɾaːto] 'lawn'. Next to an unstressed vowel it is generally pronounced as a tap, unless it's syllable-final in which case it's often a trill, e.g. deragliare /deɾaʎˈʎaːɾe/ [deɾaʎˈʎaːɾe deɾaʎˈʎaːre] 'to derail', dormiamo /dɔrˈmjamo/ [dɔrˈmjamo] 'we sleep'. The context where /r/ must be the trill [r] is as the onset of a syllable in utterance-initial position.

As you can see, unlike Spanish, Italian doesn't care whether /r/ is word-initial. This tends to confuse Spanish-speaking learners of Italian, because e.g. andare a ruba [anˈdaːɾe a ˈɾuːba] 'to sell like hot cakes' (with a word-initial flap!) may be easily misunderstood as "andare a Aruba" 'to go to Aruba (the island in the Caribbean)'.

In Japanese, there is a sociolect where the lateral flap /ɺ/ [ɺ] may be pronounced as a trill to express masculine roughness, coarseness, vulgarity, etc. Also, much to the surprise of Spanish speakers such as myself, in standard speech, the flap [ɺ] is used even after the coda /ɴ/, e.g. 人類 /dʑiɴꜜɾɯi/ [dʑĩnꜜɺɯi dʑĩꜜɺɯi] 'humanity, the human species'.

In Arabic, there is the geminate trill /r:/ [r:] and the simple flap /r/ [ɾ]. There's not much else to say here except that /r/ is often the trill [r] in syllable-final position, and furthermore among some speakers (a bit more likely if young and female) it may also be a somewhat curled alveolar approximant [ɹ], rather like the usual English rhotic, e.g. /arbaʕ/ [ˈɑɾbaʕ ˈɑrbaʕ ˈɑɹbaʕ] 'four'.

I'm not personally familiar with any languages that prefer a rhotic trill over a flap.

None of the languages I mentioned above prefers the trill over the flap, yeah.

However, I'd like to mention that 20th-century Montreal French for the most part had a single rhotic /r/, and it was realized as either the voiced trill [r] or its devoiced counterpart [r̥], something I find surprising because most languages with one alveolar rhotic have the alophone [ɾ], which they highly prefer, as you say. Examples would be prier [pr̥iˈje] 'to pray', rire [rɪ:jr] 'to laugh'. In the second half of that century this alveolar-trill-heavy speech started to be replaced by the uvular-trill-heavy speech you hear today, which uses lots of [ʀ] and [ʀ̥] (also the uvular fricatives [ʁ X], perhaps partially as an influence of European French), but you can still hear it in the speech of many older Montrealers today.

[Vijay]

My understanding has always been that rhotic trills are much less common than rhotic flaps.

[Kuchigakatai]

A curious passage from the Song of the Cid (3343-3350, begin here at the 7th line from the bottom, then continue here for the first two lines).

Pedro Bermúdez, an ally of the Cid who he affectionately calls Pero Mudo "Pedro the Mute" because of his speech impediment, is reproaching the Cid's treacherous sons-in-law in a furious long accusation, in spite of how difficult it is for him to speak. Near the end he says (Old Spanish on the left, with an orthographic update to modern morphology/spelling on the right):

Riebtot el cuerpo por malo ⁊ por t̃ẏdor 　　　(Rétote el cuerpo por malo y por traidor.)
Eſtot lidiare aq̃i antel Rey don alfonſſo 　　　(Esto te lidiaré aquí ante el rey don Alfonso,)
Por fijas del çid don eluira ⁊ dona ſol 　　　(por hijas del Cid, doña Elvira y doña Sol.)
Por q̃anto las dexaſtes menos valedes vos 　　　(Por cuanto las dejasteis, menos valéis vos.)
Ellas ſon mugieres ⁊ vos ſodes varones 　　　(Ellas son mujeres y vos sois varones,)
En todas guiſas mas valen q̃ vos 　　　(en todas guisas más valen que vos.)
Q̃ando fuere la lid ſi ploguiere al criadoR 　　　(Cuando fuere la lid, si pluguiere al creador,)
Tulo otorgaras a guiſa de t̃aẏdor 　　　(tú lo otorgarás a guisa de traidor.)
De q̃anto he dicho verdadero ſere ẏo 　　　(De cuanto he dicho, verdadero seré yo.)

I challenge you [= Fernando] to a duel for being an evil and treacherous man.
I will fight you for this in front of King Alphonse,
For the sake of the Cid's daughters, Elvira and Sol.
Because you left them, you're worth less [than them].
They are women and you are men, [note: plural "men" in the original]
So at any rate they're worth more than you are.
When the fight comes, if God wills, [lit. "if it pleases God"]
You will reach him as a traitor.
I will be found to be honest about everything I've said.

Two things:
- Latin reputō reputāre > Old Spanish riebto rebtar > modern Spanish reto retar. I hadn't noticed this before, but I think this is the one and only verb I know that has lost stem diphthongization in modern standard Spanish.
- But in particular, the line of "at any rate they're worth more than you are [because they're women]" is interesting...

If anybody was curious, guisa is a Germanic loan meaning "way, manner" (cognate with clockwise), so a guisa de traidor 'in the manner of a trator, like a traitor', en todas guisas 'anyway' lit. "in all ways".

[akam chinjir]

If anybody was curious, guisa is a Germanic loan meaning "way, manner" (cognate with clockwise), so a guisa de traidor 'in the manner of a trator, like a traitor', en todas guisas 'anyway' lit. "in all ways".

Also "guise"?

[Kuchigakatai]

Also "guise"?

Yeah, although "guise" got into English through Frankish and Old French.

[Nortaneous]

My understanding has always been that rhotic trills are much less common than rhotic flaps.

I don't know about that - I'd expect variation to be more common than either.

PHOIBLE has 1332 doculects with /r/ vs. 774 with /ɾ/, but surely there's a lot of variation elided, and some of those alleged trills will be flaps notated with the letter that's easier to type.

[Whimemsz]

Apparently one way of swearing in Unami Delaware was incorporation of scatological or similarly rude-ish roots into stems. For example:

/kpənihəla/ "you fall"

/hunti hu kpənʃːeːtːijehəla/! "pretty soon you'll fall the hell off!" (with incorporated -/ʃːeːtːije/- "anus")



(Also in honor of Thanksgiving, here is another Indian swearword FUN FACT: Choctaw speakers, who are basically all (all?) bilingual in English, normally avoid using the word for "turkey," because the native word for turkey is /fakit/ [fʌkɪt].)