(Conlangs) Q&A Thread - Quick questions go here [2010-2020]

[Salmoneus]

Epistemic modality conveys truth judgements. What is the speaker willing to commit to claiming about the truth of the proposition, and on what basis?

Deontic modality conveys value judgements. What is the speaker's reaction to and attitude toward the facts presented by the proposition?

A third category lacks (to my knowledge) a clear name. This conveys the relation between the world described by the proposition and the real world. Some of the important possibilities here are already implied by epistemic or deontic modalities - the optative, for example, (deontic, expressing the judgement that something happening is to be wished for) implies that the proposition is not currently true. However, there are potential important categories here not covered by the above, including the indicative (this is the world the proposition describes), the counterfactual (this is not that world), the hypothetical (the proposition describes a world we are imagining), and the conditional (the proposition describes a possible world that will only become actual if certain things happen).


--------

Here are some uses of "should":

a) You should eat more fruit.

b) [pointing a gun at someone] You should put the knife down now.

c) You should help old people cross the road.

d) Class, you should now turn to page five.

e) Oh wow, you should totally tie his shoelaces together!

f) I should be on holiday already.

g) It's 9:05? Shit, I should be in the office!

h) The train should arrive at 9:30.

i) Your origami butterfly should now be complete.

j) By my reckoning, he should be passing Norway around now.

k) I should hardly think so!

l) Your answers to questions four and five should match.

m) If someone said that to me, I should give them a piece of my mind!

n) Should you find yourself in Botswana unexpectedly, head immediately for the Latvian embassy.

o) It should be here!


I strongly suspect this list is not exhaustive, but each of these fifteen instances of 'should' conveys a different modality. Let's look at them:

a) is giving advice - it's identifying fruit-eating as advisable or wise.

b) is on the surface also giving advice, but is actually issuing a threat.

c) could be giving advice just like a) - it could be saying that, say, the old people might reward you if you do that. But what it's actually probably doing is conveying a moral judgement - it's identifying old-people-helping as morally good.

d) is a bit like b), because it's also an attempt to directly provoke an action. But in this case it's not a threat, it's an instruction.

e) this masquerades as advice, but isn't. Because if you follow it, the speaker will say something like "oh shit, I didn't mean you should really do it!" - instead, 'should' here is being used to fantasise. It's not necessarily saying that something would be good if it really happened, but it's saying that it's enjoyable to imagine. Another example would be the angry worker muttering under their breath "I should resign, and I should tell that idiot exactly what I think about them!" - sometimes they're giving themselves advice, but more often they're actually fantasising. Or lamenting (see next item).

f) is very different from all the above. All the above deal with future actions, but this deals with the present - it's a type of counterfactual that specifically identifies something as not being true. Specifically, it conveys (often irritated) regret for something that seemed likely to happen but did not happen.

g) looks a bit like a moral claim, but usually isn't (you're not necessarily saying it's morally wrong for you to be late). It looks a bit like a regret (for a start, it's a counterfactual), but again, it isn't really (you may be happy not to be in the office). What it actually is is usually an acknowledgement of a rule. In this sense, it's a bit like the instruction in d), which issues a rule - this one instead admits that the rule has not been followed.

h) this is also about rules (in this case, the rules set by the train timetable). But unlike d), it's not issuing a rule, and unlike h) it's not saying the rule has been broken. It's simply describing a rule.

i) this is quite like the last example, with two differences. First, it's present-tense, unlike h). It's neither indicative nor counterfactual - it's a little bit epistemic, because the speaker indicates an estimation. And specifically, this is an estimation based on advice or instruction - if you did what you were told, you will now have an origami butterfly (so it's also a bit conditional, although the protasis is only implied).

j) this is similar in some ways - again, you're dealing with an estimation about what's currently true. But this time the estimation is based not on the protasis of you having followed the advice, but on logical deduction.

k) I don't really know what this does. It's basically a way of strengthening claims, which perhaps derives from the advisory sense (it would hardly be wise of me to think so), but it's developed a less literal meaning. Or, it may be from the conditional sense (see below). Is this the same modality as in "I should hate to disturb you", or are they different?

l) this goes back to i) - again, it's a conditional based on you following advice or instruction. But it's not a present-tense estimation - or it needn't be. Instead, it can be a prediction, or, more interestingly, a piece of guidance (like "you should pass a hospital on your left" - it's not really predicting that you will, so much as giving you some reassuring infomation for later use).

m) this is a conditional. In traditional, middle-class English, "should" and "shall" are first-person suppletive forms of "would" and "will", just as th latter are suppletive first-person forms of the former (i.e. their meanings swap in the first person*). Many contemporary speakers no longer observe this shibboleth, but some have instead constructed other should vs would nuances, or else use them interchangeably, or only one and not the other.

n) this is also about conditionals, but the other way around: this is a protasic or hypothetical.

o) this is similar to f), in that it indicates a counterfactual (it isn't there!) and does so in a negative way (shit, it's not there!). But where f) indicates a regret, o) indicates surprise, shock, sometimes even outrage. It also isn't fully counterfactual (maybe it IS here!), whereas f) was fully resigned.









*A man falls into a river. "I shall drown, I shall drown!" he cries, "no one will save me!" - the crowd of good middle-class people naturally throw him a rope and haul him out of the river. The next day, another man falls into the same river. "I will drown, I will drown!" he exclaims. "No one shall save me!" - the onlookers shake their heads sadly, say "oh, the poor fellow", and respect his wishes by leaving him to drown.

[Shemtov]

Yeah, sorry, I should have specified what "shall" I meant. It's your f), g) and h), which in my English, is the most likely usages that I would use "shall" as a modal aux, ie, they're my "prototypical" usages of "shall" as a modal aux.

[KaiTheHomoSapien]

I think Boral qualifies for this, although it was less of a revamp and more of a Ship of Theseus style rebuilding. Here's a comparison almost 4 years apart:

August 2015
Jo ped te pardon por ri scrir i orthograph d'i langua - jo s'a un somer camp a Cambridge avant joudi e j'aif mal creð c'j'avlas Internet aver. A su pugnt, veðeit un bel long phraz por mangear pouc e pouc!

May 2019
Jo vos peð pardon d'ayent rien scaut sull'ig façon a dreçar y parol – jo's apart cas mey an cours stival a Cambridge fin je Jouf, e j'au mal creit connexion aïr, pu non, par nafranç. All'oc pognt, veyað eç bel e lonc protas a pascr a pieç minuscr !

I apologise for not writing anything about the language's spelling system - I'm on a summer course in Cambridge until Thursday, and I thought I'd have Internet, but alas, no. Instead, here are a nice long couple of sentences to pick over!

Well, I think I like the newer version better (without knowing anything about it!). Either way, I like to see eth being used.

Thanks for your answers, guys. The changes I've made to Lihmelinyan are far more minor than what most of you are describing. I guess it's normal to "update" your conlang every now and then.

[yangfiretiger121]

As part of New Common's evolution into Imperial Standard Common, I'm having intervocalic and final {ħ, l, ɾ → ʔ} and preconsonantal {l, ɾ → ∅}. While I'd like initial {ħ, l, ɾ → w}, {ħ, ɾ → w}'s unattested, and [l → w] seems to depend on adjacency to rounded vowels. Is a full, unconditioned merger plausible?

[Zekoslav]

As part of New Common's evolution into Imperial Standard Common, I'm having intervocalic and final {ħ, l, ɾ → ʔ} and preconsonantal {l, ɾ → ∅}. While I'd like initial {ħ, l, ɾ → w}, {ħ, ɾ → w}'s unattested, and [l → w] seems to depend on adjacency to rounded vowels. Is a full, unconditioned merger plausible?

/l/ > /w/ might happen unconditionally if /l/ was dark, i.e. [ɫ]. It happened in Polish and is happening in Bulgarian. Likewise, /ɾ/ > /w/ might happen unconditionally if /ɾ/ was rounded. Something similar happens in some varieties of English, namely /ɹ/ > /ʋ/, and English /ɹ/ is often (if not exclusively?) rounded.

Other than that, I think pure [l] and [ɾ] are more likely to merge into [j]. The latter happened in Mwotlap (an Oceanian language) and both often happen in young children's speech everywhere. [ħ] > [w] is quite weird.

[holbuzvala]

Does it seem reasonable that in a lang with largely (C)C-C roots that there be a rule disallowing /ʕ/ and /ɣ/ to both be in a root? Note that either can occur as plain,palatilised, or labialised.

Spoiler:

So really the rule would disallow any two of the six /ʕ ʕʷ ʕʲ ɣ ɣʷ ɣʲ/ from co-occuring in a root.
[ /spoiler]

[yangfiretiger121]

As part of New Common's evolution into Imperial Standard Common, I'm having intervocalic and final {ħ, l, ɾ → ʔ} and preconsonantal {l, ɾ → ∅}. While I'd like initial {ħ, l, ɾ → w}, {ħ, ɾ → w}'s unattested, and [l → w] seems to depend on adjacency to rounded vowels. Is a full, unconditioned merger plausible?

/l/ > /w/ might happen unconditionally if /l/ was dark, i.e. [ɫ]. It happened in Polish and is happening in Bulgarian. Likewise, /ɾ/ > /w/ might happen unconditionally if /ɾ/ was rounded. Something similar happens in some varieties of English, namely /ɹ/ > /ʋ/, and English /ɹ/ is often (if not exclusively?) rounded.

Other than that, I think pure [l] and [ɾ] are more likely to merge into [j]. The latter happened in Mwotlap (an Oceanian language) and both often happen in young children's speech everywhere. [ħ] > [w] is quite weird.

Okay. Ironically, my original thought was {ħ, l, ɾ → ʝ}. But, I shunned it because I wanted to keep [ʝ] as having a single genesis. Recently, I added [x] into the change because, rather than a simultaneous monolithic change, I've decided the changes'll be {ħ, x → h} and a tiered {l, ɾ → ɦ} in place of the above {ħ, l, ɾ → w} as described below. Do they look fine?

{ħ, x → h}, [h] being a true fricative
{l, ɾ → j → ɦ}, [ɦ] being a true fricative and preventing [j → ʝ] as I don't think [j, ʝ] can coexist

Additionally, ISC will have uvulars, [ɢ, ɴ, q, X], adjacent to [ɑ]. Do the front vowels, [æ æː e eː i iː], have to go all the way to centrals when adjacent to the uvulars? Their [æ, æː] are IPA [a, aː]. They didn't change the transcription because each planet transcribed its language's low front vowel as [æ], no matter its IPA value. Logically for us [X], would be [ʁ] because it's from [ɣ]. But, the language is non-rhotic.

[Creyeditor]

Does it seem reasonable that in a lang with largely (C)C-C roots that there be a rule disallowing /ʕ/ and /ɣ/ to both be in a root? Note that either can occur as plain,palatilised, or labialised.

Spoiler:

So really the rule would disallow any two of the six /ʕ ʕʷ ʕʲ ɣ ɣʷ ɣʲ/ from co-occuring in a root.
[ /spoiler]

Generally, this sounds reasonable, but it also depends on the rest of your inventory. Can other consonanants freely cooccur. Do /ʕ/ and /ɣ/ form a natural class of voiced `back' continuants?

[wintiver]

I just wanted a gut check on realism of the breadth and complexity of one of my morphemes, or rather CV-reduplication processes.

In my as-of-yet unnamed conlang, I am utilizing initial CV- reduplication to indicate the following:

• On verbs, CV- reduplication indicates increased intensity, decreased control and distributed (chaotic) distribution. I took this (and tweaked it a little) from Mountain Arapesh you can find this under section 2 on this page: https://wals.info/chapter/27
• On nouns, CV- reduplication indicates a few different things:
  • For Class 1: Humans and Class 2: Sentients & Gods it indicates untrustworthiness, villainy, perceived danger etc. This would turn "man" into "criminal" or something thereabouts. It could also turn "gods" into "demons" or something akin to that. Some caveats here though, for familial terms, especially monosyllabic ones, CV-reduplication is use endearingly - yá "older sister" > yayá "beloved/darling older sister". It can be used on polysyllabic familial terms to indicate either/or, so typically CV- reduplication is eschewed altogether, ísa "older brother" would be irísa but that is considered awkward.
  • For Class 3: Large Animals and Class 4: (Gen.) Animals this indicates danger, wildness, fierceness. This is very akin to the last bullet point with the caveat that Class 1 and 2 can have more abstract flavors of "danger" such as the aforementioned "untrustworthiness"
  • For Class 5: Supernatural (Animate) Phenomena and Class 6: Natural (Animate) Phenomena the term is applied to mean "foreboding", "ominous" and/or "dangerous"
  • For Class 7: Implements it is usually taken to mean "crappy, rusted, ruined, useless" and as always "potentially dangerous"
  • For Class 8: Places and Large Inanimates and Class 9: (Gen.) Inaninmates it usually means "dangerous, latent with power, ominous, holy or sacred"
  • For Class 10: Abstract & Mass Nouns this can mean "dangerous, toxic/inedible/non-potable, ominous, (rightfully) provoking fear"
• On attributive adjectives, CV- reduplication means "all over, all over haphazardly, sloppily/poorly/chaotically, inconsistency of the adjective/adverb itself"

[Creyeditor]

The verbal/adjectival ones sounds sensible, it is similar in meaning to some uses of Indonesian ter-.The nominal one sounds a lot like what augmentatives are used for in some languages.

[holbuzvala]

Does it seem reasonable that in a lang with largely (C)C-C roots that there be a rule disallowing /ʕ/ and /ɣ/ to both be in a root? Note that either can occur as plain,palatilised, or labialised.

Spoiler:

So really the rule would disallow any two of the six /ʕ ʕʷ ʕʲ ɣ ɣʷ ɣʲ/ from co-occuring in a root.
[ /spoiler]

Generally, this sounds reasonable, but it also depends on the rest of your inventory. Can other consonanants freely cooccur. Do /ʕ/ and /ɣ/ form a natural class of voiced `back' continuants?

The whole inventory is this:

Spoiler:

p t k pʰ tʰ kʰ
pʲ tʲ kʲ pʰʲ tʰʲ kʰʲ
tʷ kʷ
tʰʷ kʰʷ
tᶣ kᶣ
s sʲ sʷ sᶣ
m n ŋ
mʲ nʲ ŋʲ
mʷ nʷ ŋʷ
r l
w j ɥ
ɣ ɣʲ ɣʷ
ʕ ʕʲ ʕʷ

a i u y aː iː uː yː

So looking at that I'd say that /ʕ/ and /ɣ/ do form a class of voiced back continuants.
Regarding what other consonants can freely co-occur within a root, it's pretty flexible. Roots have a structure a bit more complex than just (C)C-C as mentioned above, and actually look like this:

(A)B-((D(s))E)

In short:
- Roots must begin with a consonant, but need not end in one
- The 'A' of the AB cluster must be /p t k pʰ tʰ kʰ pʲ tʲ kʲ tʷ kʷ s sʲ sʷ/, while 'B' can be any consonant except /tᶣ kᶣ/, provided that not both consonants have either palatilisation or labilalisation. I.e. an AB cluster can only have either one palatilised or one labialised consonant in it. E.g. /tkʷ/, /tʷk/, /sʲpʰ/ are fine; but /tʷkʷ/ is not allowed.
- Only one labialised element may exist in the AB cluster, such that things like /tʷm/ and /pŋʷ/ are disallowed.
- The AB cluster cannot have a nasal after a stop where both are at the same point of articulation: /kŋ/, /tn/, /pm/
- /ʕ ʕʲ ʕʷ/ cannot be adjacent to any aspirates.
- The DE cluster is essentially the linear reverse of the AB cluster, but can also be two voiced consonants provided they follow the sonority hierarchy: w j > r > l > m n ŋ > ɣ. However, /ʕ/ can form the 'D' or 'E' element of the cluster.

[yangfiretiger121]

My setting's Proto-Common language is reconstructed with [*pV, *kʝV, *pʝV, *tʝV, *ɾʝV]. While [*kʝV, *tʝV, *ɾʝV → kʰV, tʰV, rV] are unconditioned, [*pV, *pʝV → fV] conditioned by vowel fronting and rounding. The vowel system is [*a, *ɑ, *e, *ɛ, *ɜ, *i, *ɪ, *o, *ɔ, *u, *ʊ, *y, *ɑɪ̯, *ɑu̯, *ɔɪ̯], with all but [*ɜ] nasalizing after [*m, *n, *ŋ → b, d, g]. Under these circumstances, is [f~f͆] phonemic or allophonic to [p~pʰ] in Old Common?

[Dormouse559]

Under these circumstances, is [f~f͆] phonemic or allophonic to [p~pʰ] in Old Common?

If none of those phones were present in the language before the sound change, that's pretty simple to determine. The change doesn't create any contrasts, so I'd say [f f͆ p pʰ] are all allophones of the same phoneme. If any of the phones were already there, the situation gets more complex, but the odds of representing a single phoneme go way down.

[yangfiretiger121]

Under these circumstances, is [f~f͆] phonemic or allophonic to [p~pʰ] in Old Common?

If none of those phones were present in the language before the sound change, that's pretty simple to determine. The change doesn't create any contrasts, so I'd say [f f͆ p pʰ] are all allophones of the same phoneme. If any of the phones were already there, the situation gets more complex, but the odds of representing a single phoneme go way down.

Okay. [*p, *pʰ, *pʝ] are reconstructed from [p, ħ, pʰ], but [f~f͆]'s origin is split between [*p, *pʝ]. [*pʰ, *kʰ, *tʰ] collapsed into [ħ]. Old Common's aspirants and trill descend from [*Cʝ] unfailingly. They were made interchangeable with the tenuis plosives or flap because of their relative rarity. How can I describe the allophony as [Cʝ] is forbidden due to the sound changes?

[Dormouse559]

They were made interchangeable with the tenuis plosives or flap because of their relative rarity. How can I describe the allophony as [Cʝ] is forbidden due to the sound changes?

When phones are interchangeable, that means they are in free variation, which is represented by a tilde, like you were using in your previous post. The prohibition on [Cʝ] sequences is a matter of syllable structure, not allophony.

[yangfiretiger121]

They were made interchangeable with the tenuis plosives or flap because of their relative rarity. How can I describe the allophony as [Cʝ] is forbidden due to the sound changes?

When phones are interchangeable, that means they are in free variation, which is represented by a tilde, like you were using in your previous post. The prohibition on [Cʝ] sequences is a matter of syllable structure, not allophony.

Point on [Cʝ]. I've just refined this so [f~f͆] only appears before front rounded vowels; {*py, *pʝu → fy~f͆y} and [*pʝo, *pʝɔ, *pʝʊ → fø~f͆ø, fœ~f͆œ, fʏ~f͆ʏ]. How does this interact with [*pʝɑu̯], which goes from unrounded to rounded, and [*pʝɔɪ̯], which goes from rounded to unrounded?

[Dormouse559]

Point on [Cʝ]. I've just refined this so [f~f͆] only appears before front rounded vowels; {*py, *pʝu → fy~f͆y} and [*pʝo, *pʝɔ, *pʝʊ → fø~f͆ø, fœ~f͆œ, fʏ~f͆ʏ]. How does this interact with [*pʝɑu̯], which goes from unrounded to rounded, and [*pʝɔɪ̯], which goes from rounded to unrounded?

In this case, the syllabic part of the diphthong will be most likely to determine the result. You've shown with monophthongs that the change is based on phones in contact with each other. But I gather from the [*pʝo … → fø~f͆ø …] examples that there's a step before [f] appears where some back vowels are fronted. The same thing could happen to [ɑu̯] and [ɔɪ̯].

[yangfiretiger121]

Point on [Cʝ]. I've just refined this so [f~f͆] only appears before front rounded vowels; {*py, *pʝu → fy~f͆y} and [*pʝo, *pʝɔ, *pʝʊ → fø~f͆ø, fœ~f͆œ, fʏ~f͆ʏ]. How does this interact with [*pʝɑu̯], which goes from unrounded to rounded, and [*pʝɔɪ̯], which goes from rounded to unrounded?

In this case, the syllabic part of the diphthong will be most likely to determine the result. You've shown with monophthongs that the change is based on phones in contact with each other. But I gather from the [*pʝo … → fø~f͆ø …] examples that there's a step before [f] appears where some back vowels are fronted. The same thing could happen to [ɑu̯] and [ɔɪ̯].

[*pʝ → f~f͆] happens simultaneously with the fronting of the back rounded vowel, whereas [*p → f~f͆] happens because of the front rounded vowel. The changes are based on a combination of Baxter's reconstruction ([*pʝ]) and Chan's reconstruction ([*p]) of Middle Chinese altered to fit the transition between Proto-Common and Old Common. Combining that with what you said indicates [*pʝɔɪ̯ → fœɪ̯~f͆œɪ̯] as a certainty. Is it reasonable of the non-adjacency of [*pʝɑu̯]'s [*ʝ] and [*u̯] to arrest the fronting process at [fäʉ̯]?

[Dormouse559]

Is it reasonable of the non-adjacency of [*pʝɑu̯]'s [*ʝ] and [*u̯] to arrest the fronting process at [fäʉ̯]?

That's maybe overthinking things. The difference between [ä] and [a], the fully fronted version of [ɑ], isn't that important. For example, it's extremely unusual for a language to contrast them. So, unless your conlang does make that distinction, whether you transcribe the sound as [a] or [ä] really depends on preference, how you visualize the phonology of the language on an abstract level.

[yangfiretiger121]

Is it reasonable of the non-adjacency of [*pʝɑu̯]'s [*ʝ] and [*u̯] to arrest the fronting process at [fäʉ̯]?

That's maybe overthinking things. The difference between [ä] and [a], the fully fronted version of [ɑ], isn't that important. For example, it's extremely unusual for a language to contrast them. So, unless your conlang does make that distinction, whether you transcribe the sound as [a] or [ä] really depends on preference, how you visualize the phonology of the language on an abstract level.

Old common's phonology has [a]. But, fronting this [ɑu̯] to [ay̑] would bring the sound to [ɑː] in Middle Common due to an already finalized sound change. They'd see it as a partially cyclic chance, which they don't like, because of the non-syllabic back vowel. The idea is having this merge back into its parent sound, [ɑu̯], as [æ̰ː] in Middle Common.