Conlang Random Thread

[akam chinjir]

Seems reasonable. Maybe more like the construct state in Semitic languages, though the colloquial version is at least on the way to possessor agreement. (Could lasson occur with an overt indepentent pronoun? Like Turkish ben-im ev-im 1s-GEN house-1s.POSS 'my house' (evim by itself would also work).)

[Jonlang]

Seems reasonable. Maybe more like the construct state in Semitic languages, though the colloquial version is at least on the way to possessor agreement. (Could lasson occur with an overt indepentent pronoun? Like Turkish ben-im ev-im 1s-GEN house-1s.POSS 'my house' (evim by itself would also work).)

I believe my plan is to simply have the suffixes show the possession, but due to having strict rules about what consonants can occur finally many nouns will look different when nominative vs declined; /k/ isn't allowed word-finally, it becomes /t/, but with a suffix the /k/ is restored: est 'child' becomes eskel(le) in the dative (the letters in the parantheses are found in the literary form), and so 'my child' would be eskon or eskó manta.

[Richard W]

SE Asian scripts were often written on banana leaves, which tear with straight lines. Hence the extreme curviness of Malayalam, Tamil and other related scripts.

1. Why do people keep calling South Asia "Southeast Asia"? Thai, Khmer, etc. got their scripts as a result of South Asian (mainly Tamil) influence there, not because of writing on leaves. Indians wrote on leaves.
2. Malayalam is curvy. Tamil has way more straight lines.
3. Palm leaves, not banana - we'd write on palm leaves and (still to this day in some contexts) eat on banana leaves.

Much of SE Asia used palm leaves for writings to be kept long term. I get the impression that it was rather expensive. (They use cardboard for the performative roles nowadays.) There seems to be a fairly natural tension between straight lines for carving and curves for writing on leaves, though it's straight lines in certain directions that are bad.

Banana leaves do get used for writing on - even in the Roman alphabet. It seems that they can be used as a cheap substitute for paper - or at least as something that doesn't have to be bought in. i don't know how well they keep. I can imagine writing on a dried steamed banana leaf, but I've never tried it.

[Vijay]

There seems to be a fairly natural tension between straight lines for carving and curves for writing on leaves

Why "tension"?

[bradrn]

There seems to be a fairly natural tension between straight lines for carving and curves for writing on leaves

Why "tension"?

I presume he means that the original Brahmi script was carved, and had many straight lines, but then as writing moved to palm leaves, these strokes wouldn’t work any more as they tore the leaves. So there was a tension between keeping the straight lines of the original script and adapting them to curves which worked better with the palm leaves. Presumably this then means that the palm-leaf scripts are on a continuum ranging from very angular (e.g. Lontara) to very rounded (e.g. Malayalam)

…but of course I’m not Richard W, and as we saw above I’m no expert on Brahmic scripts either, so I could be wrong.

[leanancailin]

I'm trying to decide how to handle compounds that would end up violating phonological rules in my new conlang. The language allows word final nasals, but no finals in internal syllables, so /talam/ is a fine, but /!tamla/ isn't. I want a lot of compounding for word derivation, so /siro/ + /talam/ can become /sirotalam/, but /talam/ + /siro/ would become the disallowed /!talamsiro/.

There are some pretty obvious solutions here (drop the nasal from the first word, add an intrusive vowel, or just disallow words ending in a nasal from forming the first element in compounds). What I'm wondering is: do any of these strategies have further implications, or surprising correlations? And are there any less obvious solutions attested in natlangs (or conlangs)?

[Nortaneous]

Is it realistic to turn rhotacized vowel into creaky voiced vowel?

no

Really?

yes

you could have them *correspond*, by some convoluted series of sound changes

- *Vr > *Vr in one language and *Vʔ in another + simplification of finals...
- *Vt > *Vr in one language (cf. the Chinese dialect underlying Korean readings of hanja, IIRC) and *Vʔ in another
- or if your language has preinitials, copy some Tibetosphere developments, *r-CV > CVr in one language and *r-CV > *C:V > CV̰ in another - you don't *necessarily* need *Vr and *Vʔ as intermediates!

but there's no motivation for rhotacized vowels to become creaky-voiced

I'm trying to decide how to handle compounds that would end up violating phonological rules in my new conlang. The language allows word final nasals, but no finals in internal syllables, so /talam/ is a fine, but /!tamla/ isn't. I want a lot of compounding for word derivation, so /siro/ + /talam/ can become /sirotalam/, but /talam/ + /siro/ would become the disallowed /!talamsiro/.

There are some pretty obvious solutions here (drop the nasal from the first word, add an intrusive vowel, or just disallow words ending in a nasal from forming the first element in compounds). What I'm wondering is: do any of these strategies have further implications, or surprising correlations? And are there any less obvious solutions attested in natlangs (or conlangs)?

why does the language allow word-final nasals but no finals in internal syllables? what historical developments led it to this state?

[malloc]

I'm trying to decide how to handle compounds that would end up violating phonological rules in my new conlang. The language allows word final nasals, but no finals in internal syllables, so /talam/ is a fine, but /!tamla/ isn't. I want a lot of compounding for word derivation, so /siro/ + /talam/ can become /sirotalam/, but /talam/ + /siro/ would become the disallowed /!talamsiro/.

There are some pretty obvious solutions here (drop the nasal from the first word, add an intrusive vowel, or just disallow words ending in a nasal from forming the first element in compounds). What I'm wondering is: do any of these strategies have further implications, or surprising correlations? And are there any less obvious solutions attested in natlangs (or conlangs)?

The solution might depend on the history of the language. Final consonants generally arise when final vowels are dropped. Presumably the form /talam/ derives from some older form like /talamu/ which lost its final vowel. Perhaps this final vowel returns in compounds (or more precisely never dropped in such environments). Thus the compound /talam/ + /siro/ might yield /talamusiro/. You could also look at natural languages with phonotactics like your conlang and see how they solve this problem.

[Pedant]

I'm trying to decide how to handle compounds that would end up violating phonological rules in my new conlang. The language allows word final nasals, but no finals in internal syllables, so /talam/ is a fine, but /!tamla/ isn't. I want a lot of compounding for word derivation, so /siro/ + /talam/ can become /sirotalam/, but /talam/ + /siro/ would become the disallowed /!talamsiro/.

There are some pretty obvious solutions here (drop the nasal from the first word, add an intrusive vowel, or just disallow words ending in a nasal from forming the first element in compounds). What I'm wondering is: do any of these strategies have further implications, or surprising correlations? And are there any less obvious solutions attested in natlangs (or conlangs)?

There's also the possibility of consonant lenition in a new environment. I'm not sure how many actual consonants your conlang possesses, but a form of cluster-internal sandhi might be a good way to get both new forms in place. For example, /talam/+/siro/ would be /!talamsiro/ unchanged, but perhaps there's a rule in there that says /s/ plus a nasal becomes a geminate consonant (if those are allowed), so the result would be **/talassiro, talas:iro/. Or maybe the consonant clusters went through a couple of sound changes so that the resulting consonant has the place of the former (so a labial) and the manner of the latter (so a fricative), and the combined form (a labial fricative) is the one used, thus **/talaɸiro/. Or maybe the preceding vowel is marked in some way, lengthened or stressed or given a heightened tone or even nasalized, so for example **/ta'lasiro/ or **/talaasiro>tala:siro/ or **/tàlásīrò/ or **/talãsiro/.
The choice is yours, of course. malloc's suggestion of dropped final vowels isn't a bad one either, although you'd have to come up with a reason for the vowels to have been dropped while others stayed (final consonant restrictions--/m/ is permitted but /r/ is not, final short vowels dropped while later the distinction between long and short vowels is lost, lax vowels dropped wherever in a word and then reformed to prevent consonant clusters--**/talɐmʊ>talm>talam/, the rare but possible transformation of vowels into semivowels into fricatives or nasals--for example **ãu>aũ>aw̥̃>am̥>am, or some other means). Whatever you choose, do let us all know; it's always fun to see how these things turn out.

EDIT: the **/talaɸiro/ form would be derived **/talamsiro>talamɸiro>talaɸiro/. "No consonant clusters allowed" rule followed, and a whole new phoneme for the compound!

[Kuchigakatai]

leanancailin:

The use of compounding doesn't mean that word-internal phonotactics have to be respected. It is perfectly naturalistic to have normally unallowed clusters across the morphemes of a compound word. So you could allow "talamsiro" while forbidding morphemes like *wamsa. English and German have plenty of examples of this, where compounds such as "ragtime" or "lip balm" allow clusters like /gt/ and /pb/ that are normally forbidden within a morpheme.

Burmese is an interesting language in terms of compounding because compounded morphemes undergo various changes that clearly identify them as being part of a compound, allowing this highly isolating/analytic language to reuse the same word order unambiguously (unchanged morphemes indicate a syntactic construction, altered morphemes indicate a compound). The initial consonant of the head morpheme often undergoes lenition (voicing, sometimes nasalization after a nasal) while modifier morphemes undergo vowel shortening (with accompanying phonation/tone change, as the phonations/tones have associated vowel length), vowel reduction, assimilation of the coda consonant, and sometimes voicing of the initial consonant. An example:

kula [kəˈla˨] 'Indian'
htuing: [tʰãːi˥] 'to sit'
kula htuing: [kəˈla˨ ˈtʰãːi˥] 'The Indian person sits.'
kula htuing: [kələˈtʰãːi˥] 'chair' (literally "Indian-sitting-thing")

As you can see, however, the writing system does not distinguish syntactic constructions from compounds, even though these are distinguished in speech.

[Qwynegold]

I have a highly inflecting conlang (currently named 'L'). Nouns decline for genitive and possessed, i.e. the thing possessed by the noun in the genitive takes a "possessed" case marking. I don't think this is a thing in European langs (as far as I can tell from a quick search) and need a better name than "possessed" for it. I thought of "possessive" but this could be confused with genitive.

Maybe this is late now, but have you seen this?: https://en.wikipedia.org/wiki/Possessive_affix

[Qwynegold]

malloc's suggestion of dropped final vowels isn't a bad one either, although you'd have to come up with a reason for the vowels to have been dropped while others stayed

IIRC, Japanese got its coda nasal from an earlier /mu/.

leanancailin: Which nasals do you have?

[Richard W]

There seems to be a fairly natural tension between straight lines for carving and curves for writing on leaves

Why "tension"?

I presume he means that the original Brahmi script was carved, and had many straight lines, but then as writing moved to palm leaves, these strokes wouldn’t work any more as they tore the leaves. So there was a tension between keeping the straight lines of the original script and adapting them to curves which worked better with the palm leaves. Presumably this then means that the palm-leaf scripts are on a continuum ranging from very angular (e.g. Lontara) to very rounded (e.g. Malayalam)

…but of course I’m not Richard W, and as we saw above I’m no expert on Brahmic scripts either, so I could be wrong.

You've understood me, except that the same script may be used for carving and writing on leaves. And people don't just carve in stone - they also carve on wood - in the abstract, think runes, and as a matter of Indic history, think of the Philippine scripts, which are traditionally carved in wood.

[Xwtek]

The term possessed is also used for describing Nahuatl noun. Although, it's not considered a case marking. However, given the other form is named absolutive, it's just waiting a unwitting person to call it "case". And in fact, I thought that as Nahuatl case until I read further.

[bradrn]

The term possessed is also used for describing Nahuatl noun. Although, it's not considered a case marking. However, given the other form is named absolutive, it's just waiting a unwitting person to call it "case". And in fact, I thought that as Nahuatl case until I read further.

Is there any particular reason why this isn’t considered to be a case system?

[akam chinjir]

The term possessed is also used for describing Nahuatl noun. Although, it's not considered a case marking. However, given the other form is named absolutive, it's just waiting a unwitting person to call it "case". And in fact, I thought that as Nahuatl case until I read further.

Is there any particular reason why this isn’t considered to be a case system?

Because it's head-marking.

[Vijay]

It doesn't have anything to do with case as far as I can tell. One form of the noun is used when a possessive prefix is used, e.g. the equivalents of 'my house', 'your house', etc. The other is used in all other cases.

[leanancailin]

Thanks everyone for your replies! I hadn't worked out any of the background before; thinking about it more and looking through the different strategies, I think I'm going to say any old N before consonants turned into length (and maybe some impact on tone), and have old~fossilized compounds lose their nasals to length. I'll play around with how I like the look and sound of final nasals inside compounds and decide whether the loss is productive and ongoing or if newer or ad hoc compounds also use lengthening or not.

[Xwtek]

Is 1DU.INCL > 1SG.S.2SG.O and 1PL.INCL > 1.S.2.O.PL (either/both the subject and the object are plural) realistic, because I saw the Tagalog word for "I ... you" is "kita", which means 1PL.INCL in Indonesian.

[bradrn]

Is 1DL.INCL > 1SG.S.2SG.O and 1PL.INCL > 1.S.2.O.P realistic, because I saw the Tagalog word for "I ... you" is "kita", which means 1PL.INCL in Indonesian.

What do these abbreviations mean?