Aʻatun Scratchpad: Inventory questions

[yangfiretiger121]

The languages once barred gemination; instead, gemination will now be marked by what started as palatalization ([ik.nɑ'sʲɑ→ik.nɑ'ssɑ]) or an elided liquid ([ɰi'sol.tɑ→ɰi'so.ttɑ]).

While not gemination, the above conditions apply to /xʲ/, etc. as well ([in.tus'tr̥ʲɑ→in.tusʈ'ʈ͡ʂɑ]).

[yangfiretiger121]

As I've stated before, Altrunians would have devoiced /b→p/, etc. Later on, they respelled the devoiced consonants. Would a /j→ʝ→ç/ progression make sense in this context? /j/ can't, merely, vanish offhand because it would necessitate insertion of /ʔ/ as they'd go out of their way to retain a historical prevocalic consonant.

[Travis B.]

One thing to remember is that people do not know how past speakers spoke their language; people don't do things or not do things sound-change-wise because of how things are "supposed" to be or because people in the past did things a certain way. To invoke this requires that other varieties are spoken alongside a given variety that can influence it, or the existence of an orthographic tradition that can be a source of spelling pronunciations.

[yangfiretiger121]

Languages spoken on Altrunia, such as Common and Sphinx, coalesced into Altrunian around 100 years after the planet was unified. Sometime before Altrunians conquered Protewa, most of the voiced stops, affricates, and fricatives were devoiced (later to be respelled). Eventually, the religious figures created Pahus by removing most palatal consonants, eliminating all palatalizations from the language, changing /l/ to /w/ (word-initially or -medially) or [ʔV] (postvocalically) (coincidental respelling), and changing /j→ʝ/. Eventually, the final change migrated to the parent language. Is the /ʝ→ç/ needed to eliminate voiced consonants from Altrunian/Galactic Standard an attested change? The /ʝ→ʃ/ required to eliminate palatals from Pahus?

[yangfiretiger121]

Afwia is the current spelling of Alfia following /ɑ/ merging into /ɔ/ (later changed back to /ɑ/) and /l/ merging into /w/. Is ['ɑf.fwa] or ['ɑf.w:a] a more natural outcome, considering the new geminiation rule?

[Travis B.]

Afwia is the current spelling of Alfia following /ɑ/ merging into /ɔ/ (later changed back to /ɑ/) and /l/ merging into /w/. Is ['ɑf.fwa] or ['ɑf.w:a] a more natural outcome, considering the new geminiation rule?

['ɑf.fwa] is probably more likely than ['ɑf.w:a], considering it is common (but not universal, e.g. with Arabic) for languages that allow gemination to not allow gemination of /w/ or /j/,

[Vijay]

Malayalam has /jj/ and /ʋʋ/.

[yangfiretiger121]

Which of /p f p̪͡f ʋ̥ ɰ̊ ɥ̊ k͡p/ are likely outcomes for /ʍ/? They try to keep, at least, one place of articulation constant.

[Travis B.]

Which of /p f p̪͡f ʋ̥ ɰ̊ ɥ̊ k͡p/ are likely outcomes for /ʍ/? They try to keep, at least, one place of articulation constant.

None of those - look at what natlangs, e.g. English or North Germanic, have done with /ʍ/~/hʷ/~/xʷ/, i.e. turn it into [w], [ʋ], [v], [kv], or [k].

[yangfiretiger121]

According to Wikipedia, /ʍ→f/, while unattested, is quite possible (cf. /w→p/). As it stands, /ʍ/ is their only approximate because /l/ merged into /w/ and /j/ fortited to /ʝ/ before devoicing to /ç/.

[Travis B.]

IIRC ʍ > f does occur in some Scots varieties.

The thing is that it seems like you intentionally seek out rare or otherwise highly unlikely sound changes and have a poor understanding of how sound change happens (it does not happen because a language is "supposed to" be a certain way, and people have no memory of how people spoke any more than a few generations ago unless such is codified in a standard variety that is synchronically extant or otherwise reinforced by orthography). I would recommend that you work on acquiring a good working knowledge of linguistics at this point.

[yangfiretiger121]

I've decided to leave /ʍ/ alone. I'm rushed at this point because my collaborator went behind my back and scheduled one RP a year, with the one the languages are for planned for next year. Thus, I only have six-and-a-half months to get the basics down. Hells, I'd be much further along at this point if I'd used the current setting right away. Given this timeframe, I'm trying to keep the changes from Galactic Standard to Pahus simple because I can't construct a second language if the first is giving me this much trouble. Of course, this may be a lose-lose situation for me.

[Hallow XIII]

Aren't you spending a lot of time obsessing about sound change if it's for an RP setting? Or for any sort of setting, really. It's not really my business but if you're feeling pressed for time (which is almost always the case if you're creating languages for a purpose other than the simple fun of it), I would advise the following:

1. Concentrate on the languages you actually need. If you require, say, High Qaqsëq first, make that and then worry about Proto-Qaqsëq. Of course, if you won't have need to display older forms or sister languages, you can leave that bit out entirely.
2. In a similar vein, avoid creating IE-like situations where you would need to trace things back 6000 years, because that is ridiculously complicated. Most IE subfamilies, for example, are sufficiently distinct from one another that you can treat them as distinct families if the setting demands that you have such a large family; stick with certain areal effects and maybe try to sneak the occasional correspondence in. French and German's visible similarity is mostly down to convergence and mutual borrowing from Latin than any sort of inheritance.
3. Most fictional settings have some sort of huge dominant literary language. Simplify your life by creating a dictionary for it and then borrowing any sort of complicated term from that language instead of creating fifty derivational morphologies.
4. When creating any one language, focus on what you will need for the genre of text you will actually need to have produced in a language, and work by translation. Prepare short texts, maybe one literary prose, one spoken conversation, one scientific prose. You probably won't need to work on actual spoken discourse markers for your Conlatin, and you probably don't need to know how to make huge complicated sentences in a low-class patois.
5. Differentiate your languages by aesthetic. If they look sufficiently different nobody will care if you used the same complement clause syntax for each one. In fact, doing this (call it the "esoterogenic method") is highly recommended if for some reason other people are going to have to use your languages.

As an example for the last point, here are two conlangs I slapped together for my last D&D campaign. They use very similar syntax, but look very different, which is much more important in differentiating them to a non-technical audience as well.

• Deirxe hyranggun ninggun tasrag siersdiesden hyal galm xarbandsydwysyn. Galmgam ninggun waid yryn hyal deirx sireiun. Seir wyangulsmag tursewereweng isigun wied, isigun yrbag.
• Rhûn echayâsh ucham ngônas i-tthambrêkêk mahâlu pechkêk. Ani "Pechenyî na ngosam i-mhahâl, yesh kosam i-mhethênî hêk, ani uchyâke karash shîrinêk. Cholol indrây inna rhûwê na ngarikî" ôsh Rhûnu.

[yangfiretiger121]

Aren't you spending a lot of time obsessing about sound change if it's for an RP setting? Or for any sort of setting, really. It's not really my business but if you're feeling pressed for time (which is almost always the case if you're creating languages for a purpose other than the simple fun of it), I would advise the following:

1. Concentrate on the languages you actually need. If you require, say, High Qaqsëq first, make that and then worry about Proto-Qaqsëq. Of course, if you won't have need to display older forms or sister languages, you can leave that bit out entirely.
2. In a similar vein, avoid creating IE-like situations where you would need to trace things back 6000 years, because that is ridiculously complicated. Most IE subfamilies, for example, are sufficiently distinct from one another that you can treat them as distinct families if the setting demands that you have such a large family; stick with certain areal effects and maybe try to sneak the occasional correspondence in. French and German's visible similarity is mostly down to convergence and mutual borrowing from Latin than any sort of inheritance.
3. Most fictional settings have some sort of huge dominant literary language. Simplify your life by creating a dictionary for it and then borrowing any sort of complicated term from that language instead of creating fifty derivational morphologies.
4. When creating any one language, focus on what you will need for the genre of text you will actually need to have produced in a language, and work by translation. Prepare short texts, maybe one literary prose, one spoken conversation, one scientific prose. You probably won't need to work on actual spoken discourse markers for your Conlatin, and you probably don't need to know how to make huge complicated sentences in a low-class patois.
5. Differentiate your languages by aesthetic. If they look sufficiently different nobody will care if you used the same complement clause syntax for each one. In fact, doing this (call it the "esoterogenic method") is highly recommended if for some reason other people are going to have to use your languages.

As an example for the last point, here are two conlangs I slapped together for my last D&D campaign. They use very similar syntax, but look very different, which is much more important in differentiating them to a non-technical audience as well.

• Deirxe hyranggun ninggun tasrag siersdiesden hyal galm xarbandsydwysyn. Galmgam ninggun waid yryn hyal deirx sireiun. Seir wyangulsmag tursewereweng isigun wied, isigun yrbag.
• Rhûn echayâsh ucham ngônas i-tthambrêkêk mahâlu pechkêk. Ani "Pechenyî na ngosam i-mhahâl, yesh kosam i-mhethênî hêk, ani uchyâke karash shîrinêk. Cholol indrây inna rhûwê na ngarikî" ôsh Rhûnu.

Thank you, Hallow. Ironically, I'd started out solely focused on Galactic Standard and Outsider (should have focused on just GS). Then, my collaborator told me that GS' main ancestor, Sphinx, was "literally Ancient Egyptian." Thus, I got sidetracked on how Sphinx would have sounded, as if my plate wasn't full enough with two languages already. Also, the setting's something like Star Trek.

[yangfiretiger121]

Is /l→ʋ/ a likely change?

[Zaarin]

Is /l→ʋ/ a likely change?

/l > ɫ > w~ɰ > ʋ/ is plausible.

[Travis B.]

I should also state that often changes that are implausible as single steps are often plausible when broken up into a chain of intermediate steps which are each plausible.

[Vijay]

There are cases of /l/ changing to in some Dardic languages. Surely getting [ʋ] from there should be pretty trivial. Even as a direct change, I'm not sure /l/ > [ʋ] is really implausible.

[yangfiretiger121]

Okay. Thanks. What are possible endpoints for /fʲ θʲ ðʲ ʋʲ/, if any exist?

[Zaarin]

/ç ç ʝ ʝ/ or /ç ç ʝ j/ are perhaps the most obvious short of depalatalizing. /ʋʲ/ > /v/ is also not inconceivable. I'd buy /θʲ/ > /ɕ~ʃ/.

[Vijay]

I would expect either fronting of some sort or no change. (That's not to say there aren't necessarily other possible outcomes).

[yangfiretiger121]

Okay. I may end up with a four (/i e̞ u ɑ/ or /i e̞ o̞ ɑ/) or five (/i e̞ u o̞ ɑ/) system. If there's a four vowel system, is /u>o̞/ or /o̞>u/ more likely?

[Zaarin]

Okay. I may end up with a four (/i e̞ u ɑ/ or /i e̞ o̞ ɑ/) or five (/i e̞ u o̞ ɑ/) system. If there's a four vowel system, is /u>o̞/ or /o̞>u/ more likely?

I'd say equally likely.

[Vijay]

I find /i e̞ u ɑ/ more likely than /i e̞ o̞ ɑ/.

[Zaarin]

I find /i e̞ u ɑ/ more likely than /i e̞ o̞ ɑ/.

Some variation of /i e o a/ is pretty common in North America.