Conlang Random Thread

[Vijay]

Acute accents? (After all, Hungarian uses them for length, and Spanish uses them for stress, so...).

[Raphael]

Acute accents? (After all, Hungarian uses them for length, and Spanish uses them for stress, so...).

Thank you! If I do that, what can I use for short stressed vowels, then?

[Xwtek]

Newbie question: When coming up with a romanization scheme for a conlang, is there an easy and not-too-confusing way of indicating that a vowel is both accented and long?

Can you double the vowel for the length and acute accent for the first letter for stress?

[Raphael]

Newbie question: When coming up with a romanization scheme for a conlang, is there an easy and not-too-confusing way of indicating that a vowel is both accented and long?

Can you double the vowel for the length and acute accent for the first letter for stress?

Thank you for your reply. I could do that, but if I ever get anything I might want to post, and post it somewhere, many of the people who see it will probably be English speakers, and I don't want to tempt them to interpret a long "e" as being pronounced like English "ee", or a long "o" as being pronounced like English "oo".



Right now, I'm tending towards graves for short stressed vowels and acutes for long stressed vowels.

[Xwtek]

Newbie question: When coming up with a romanization scheme for a conlang, is there an easy and not-too-confusing way of indicating that a vowel is both accented and long?

Can you double the vowel for the length and acute accent for the first letter for stress?

Thank you for your reply. I could do that, but if I ever get anything I might want to post, and post it somewhere, many of the people who see it will probably be English speakers, and I don't want to tempt them to interpret a long "e" as being pronounced like English "ee", or a long "o" as being pronounced like English "oo".



Right now, I'm tending towards graves for short stressed vowels and acutes for long stressed vowels.

How about no accent for unstressed short, acutes for stressed short, macron for unstressed long, and either macron+acutes and circumflex for stressed long. The circumflex evokes an accent while being similiar to macron.

[Raphael]

How about no accent for unstressed short, acutes for stressed short, macron for unstressed long, and either macron+acutes and circumflex for stressed long. The circumflex evokes an accent while being similiar to macron.

Yes, I think I'll do something like that. Thank you!

[bbbosborne]

Newbie question: When coming up with a romanization scheme for a conlang, is there an easy and not-too-confusing way of indicating that a vowel is both accented and long?

hungarian likes to use the double accent: <ő ű>

sometimes when long vowels are written as 2 symbols, both receive an accent: <áá>

[Raphael]

Another complete newbie question: if you divide languages into agglutinative, fusional, and isolating, how common is it for a language belonging to one of these types to eventually turn into another type over time?

[missals]

Another complete newbie question: if you divide languages into agglutinative, fusional, and isolating, how common is it for a language belonging to one of these types to eventually turn into another type over time?

Very common. For some changes, like completely isolating to completely fusional, we don't have a ton of evidence, probably because such a change would take a very long time. But we wouldn't necessarily expect to see that kind of thing very much anyways - most languages aren't purely agglutinative, or fusional, or isolating, but instead adhere partially to one morphological profile while having elements of the others. And it's very common for a language to gain or lose elements of one profile over time, or to broadly swing towards one profile while not entirely losing aspects of the others.

It was once assumed there was a natural typological cycle that languages went through, of isolating > agglutinative > fusional > and back to isolating again - taking perhaps 10,000 years, but who really knew - but this idea is rejected by many historical linguists today, perhaps even most. One big problem with this idea - aside from the fact that we've never observed a complete revolution of this cycle - is that we now know language contact has a lot to do with changes in the morphological profile of a language, and it's not at all clear that all fusional languages "naturally" become isolating/analytic if simply given enough time. Without contact effects, it is possible that a particular morphological profile will remain generally stable for a very, very long time - look at the persistence of the Semitic triconsonantal root system, for example. And Iroquoian has likely been polysynthetic ever since the time of Proto-Iroquoian 3,500 years ago.

Many of the changes in morphological type that we know about have been influenced by contact - it has been generally observed that languages that undergo a significant influx of adult learners often tend to lose inflection and become more isolating/analytic. Creoles may be an extreme example of this process - it was once widely thought that creole languages were entirely "new" languages built up from the detritus of other languages, but there is now an open debate about whether creole languages are simply the product of a language undergoing completely normal but very extensive processes of historical change - i.e. that Tok Pisin, for example, really is just a descendant of English that rapidly experienced a lot of changes. Many changes from a more fusional to a more agglutinative profile were also influenced by contact - Armenian and Ossetian are descneded from fusional Indo-European languages, but today are broadly agglutinative due to contact with the various Caucasian languages.

This is not to say that all changes in morphological profile are induced by contact. In most languages we can identify individual grammatical changes that result in a more agglutinative, fusional, or isolating profile that are clearly the product of entirely language-internal factors.

I guess I don't have one main point, just trying to give you some things to think about.

Also, on the subject of morphological profile in conlangs - to paraphrase something I once saw on here, or maybe the CBB - if you're evolving an agglutinative conlang, and you don't want sound changes to turn it into a synthetic language, then remember - there isn't some council that meets every 1000 years to "fix" all of the sound changes and restore an agglutinative profile. Agglutinative languages remain stable even through sound change due to constant, small, thoroughgoing analogical changes.

[Raphael]

Thank you!

[Raphael]

Continuing my string of newbie questions:

1) Would it be plausible to have a language in which verbs are divided into two groups, in such a way that for the verbs in the first group, the unmarked form is the present tense and the past tenses require some kind of declension, while for the verbs in the second group, the unmarked form is a past and/or perfect tense, and the present tense requires some kind of declension? The idea would be that verbs describing frequent everyday activities would be in the first group, while verbs describing relatively rare actions that people are more likely to talk about in the past tense would be in the second group.

2) Assuming that a language uses verb declensions, would it be plausible to distinguish between first, second, and third person, but not between singular and plural?

3) How plausible would it be to have a language in which adjectives routinely form compounds with the nouns they describe?

[Salmoneus]

Continuing my string of newbie questions:

1) Would it be plausible to have a language in which verbs are divided into two groups, in such a way that for the verbs in the first group, the unmarked form is the present tense and the past tenses require some kind of declension, while for the verbs in the second group, the unmarked form is a past and/or perfect tense, and the present tense requires some kind of declension? The idea would be that verbs describing frequent everyday activities would be in the first group, while verbs describing relatively rare actions that people are more likely to talk about in the past tense would be in the second group.

That's... an interesting question. Conceptually, it seems to make sense - in the way that some verbs inherently have one aspect or another, or the way that some verbs have different alignments from others. However, I've never heard of it happening in a real language and my instinct is that it probably doesn't, no. Not on a large scale, anyway. In English, there are a few verbs that are basically only found in the past tense (more that are mostly only found as past participles), but these are very rare, and I'm not sure a language could have a large class like this. Indeed, "only found in the past tense" seems more plausible to me than "requires an affix outside the past tense" - maybe because, if these are verbs that don't apply to present actions, how often do you use those affixes anyway? But what might happen maybe is having some verbs only found in the past tense, and then derivation into other verbs that can appear in either tense, but usually used as present tenses of the defective verbs, with the same derivative affix also used in other contexts?
For example, I think the only English verb that is officially only found in the past tense (rather than those where the present tense is just more obscure) is "quoth". So maybe if people wanted a present-tense version of "quoth", they might derive a verb like "quothify" - officially this would have present and past tenses (I quothify today, you quothified yesterday), but in practice it could act as a suppletive, pseudo-inflectional present tense (I quothify today, you quoth yesterdy). The situation where this might arise would be where the verb was irregular and hence it was possible to 'forget' some present tense forms (as with 'quoth' - the present tense is/was 'quethe', but only philologists know that). I'll have to remember this idea for a future conlang. But I think it's only likely, even then, to be a way of forming a small class of irregular verbs, rather than something that could cover half the verbs in a language.

2) Assuming that a language uses verb declensions, would it be plausible to distinguish between first, second, and third person, but not between singular and plural?

Absolutely. Verbs can distinguish two numbers, or three (often a dual or a paucal), or four, or maybe more, or, absolutely, none at all.

3) How plausible would it be to have a language in which adjectives routinely form compounds with the nouns they describe?

What do you mean by 'routinely'? Do you mean it's a common way to derive new words, or do you mean each adjective and each noun always form their own compound? If the former: sure, English does it ('blackbird'). If the latter: I doubt it. If the 'compound' is entirely predictable from its parts with no sandhi, I'm not sure why you'd describe it as a compound (if there were no uncompounded noun phrases to compare it to); but if each one of the billion possible adjective-noun combinations has to get its own derived form, then that's too many words for people to realistically remember.

[Raphael]

Ah, bummer. Thank you, anyway!

[Raphael]

Next question: what about having the same word for "to give birth" and "to be born"? That is, one word for the act of participating in a birth as either the mother or the baby?

[mèþru]

Hebrew has the same root but not the same verb. I use the same word in some of my conlangs, but affixation is used to help disambiguate sometimes. That may be how Hebrew's situation evolved.

[Xwtek]

Next question: what about having the same word for "to give birth" and "to be born"? That is, one word for the act of participating in a birth as either the mother or the baby?

Ergative language?

[Raphael]

Next question: what about having the same word for "to give birth" and "to be born"? That is, one word for the act of participating in a birth as either the mother or the baby?

Ergative language?

Not sure if that's necessary. There's no law of nature saying that "to be born" has to be a passive construction the way it is in English. Even in English, you can, if you want, use the construction "to emerge from the womb", even though people usually don't use that construction. What I was thinking of would be a language where

1) the standard way to describe was the baby does during a birth is a verb usually used in the active voice, that might be literally translated as something like "to emerge from the womb", and

2) that same verb also means "to give birth".

[akam chinjir]

About the verbs: it wouldn't be too strange to have some verbs that are imperfective by default, and others that are perfective; and it also wouldn't be too strange (I think) to have the imperfective turn into a nonpast and the perfective into a past. So that'd give you something like the split you're suggesting, though with the line falling in a different place. Anyway it's the closest I've come up with.

Next question: what about having the same word for "to give birth" and "to be born"? That is, one word for the act of participating in a birth as either the mother or the baby?

Ergative language?

You don't need an ergative language to get that sort of alternation, just (what have been called) ergative verbs (not everyone likes the terminology, though). "Give birth"/"be born" is a pretty good candidate; Chinese 生 (shēng, sāang, etc) is an example.

[Salmoneus]

So the sentence "She XANG in 1976" could mean either "she had a child in 1976" or "she was born in 1976"? That seems... kind of an important distinction to make?
A compromise might be having a verb with one meaning when univalent, and another meaning when bivalent? So "she birthed" = "she was born", but "she birthed a son" = "she gave birth to a son". In English, this is a zero-derived verb derivation, indicated only grammatically. If you have both zero-derivation and optional object-dropping, you could have it LOOK like they're the same verb, but they'd underlyingly have a different valency, which could be used to disambiguate? That seems to me much more likely than having a single, univalent verb with both meanings (or a single bivalent verb with both meanings, for that matter). But I don't think I can absolutely rule out the latter either - someone else would have to weigh in. .

[akam chinjir]

Yeah, I was assuming that with the to give birth meaning, it'd be transitive. (So you could only find that sort of an alternation where transitivity isn't somehow indicated on the verb.)