zompist bboard

Hi! I've been thinking of developing a naturalistic conlang derived from a historical proto-language, and I'd like some advice. I first wanted to use Proto-Italic (I thought about a Romlang, but that seemed overdone), but I could hardly find any sources on its vocabulary or grammar. (If anyone can help me with that, that'd be nice.) So I thought I'd use Proto-Indo-European.

I've been reading about Proto-Indo-European, and it just seems incredibly complex to me. Everything about it is quite hard to grasp. I found this wordlist for Proto-Indo-European, but it appears to be based on older sources and doesn't seem to line up with laryngeal theory. However, to be perfectly honest with myself, its level of complexity seems a bit more "within my reach".

So, what I'd like to know is: Would it be better for me to create a higher-quality derivation from a reconstruction that, while possibly out-of-date, I am able to work with more easily - or work with a more rigorous and complex reconstruction, but one that I am more likely to make mistakes with? Does it matter either way as long as my derived conlang is derived in a consistent manner?

Additionally, does anyone have any suggestions for good proto-languages to work with - that is, ones that are well-documented and aren't too difficult for someone with a hobby-level knowledge of linguistics to understand? (Clearly, I would not attempt to work with Old Chinese.)

Porphyrogenitos wrote:So, what I'd like to know is: Would it be better for me to create a higher-quality derivation from a reconstruction that, while possibly out-of-date, I am able to work with more easily - or work with a more rigorous and complex reconstruction, but one that I am more likely to make mistakes with? Does it matter either way as long as my derived conlang is derived in a consistent manner?

who cares

Srsly, just do whatever you enjoy more

Nothing matters.

All that's significant in choosing your starting material is what you want your result to be. If you're just randomly picking something to save you having to create a parent language de novo, sure, it doesn't matter how un-IE your source material is. If you want to make something that could pass for IE, or that impresses IEists, then clearly it does matter how IE your source material is.

[Do bear in mind, though, that all (I assume all?) of PIE's daughter languages have greatly simplified large parts of PIE's grammar.]

Porphyrogenitos wrote:Additionally, does anyone have any suggestions for good proto-languages to work with - that is, ones that are well-documented and aren't too difficult for someone with a hobby-level knowledge of linguistics to understand? (Clearly, I would not attempt to work with Old Chinese.)

Proto-Uralic is pretty well documented, and is nowhere near as batshit complex as PIE. Proto-Semitic is pretty well documented as well, but I guess it would be slightly more difficult starting material, particularly if you're not overly familiar with Semitic-style triconsonantal roots.

Alternatively, you could take Etruscan or Sumerian as your proto-lang? Both have reasonable documentation (the latter far more so than the former, of course)- and I don't think anyone has made a language derived from either.

Yng wrote:
who cares

Srsly, just do whatever you enjoy more

Well, clearly, I'll do what I enjoy most. In this case, part of what I enjoy is creating a plausible, realistic descendant of PIE or another language, so I was seeking advice on how to go about doing that.

Salmoneus wrote:Nothing matters.

All that's significant in choosing your starting material is what you want your result to be. If you're just randomly picking something to save you having to create a parent language de novo, sure, it doesn't matter how un-IE your source material is. If you want to make something that could pass for IE, or that impresses IEists, then clearly it does matter how IE your source material is.

[Do bear in mind, though, that all (I assume all?) of PIE's daughter languages have greatly simplified large parts of PIE's grammar.]

Hmm, you raise an interesting point there. Though I am aware that PIE's descendants didn't keep all of its grammatical complexities.

Dewrad wrote:Proto-Uralic is pretty well documented, and is nowhere near as batshit complex as PIE. Proto-Semitic is pretty well documented as well, but I guess it would be slightly more difficult starting material, particularly if you're not overly familiar with Semitic-style triconsonantal roots.

Alternatively, you could take Etruscan or Sumerian as your proto-lang? Both have reasonable documentation (the latter far more so than the former, of course)- and I don't think anyone has made a language derived from either.

Those are all good suggestions. In fact, I've also been considering Sumerian and Etruscan on the side, all throughout, and I think I'm going to try to look into Etruscan a bit more. I'll also take a look at Proto-Uralic and Proto-Semitic.

Porphyrogenitos wrote:
Well, clearly, I'll do what I enjoy most. In this case, part of what I enjoy is creating a plausible, realistic descendant of PIE or another language, so I was seeking advice on how to go about doing that.

I think the answer to this is implicit in your question. If your overriding desire is to make the most plausible (to IEists, nobody else knows enough to say if it's plausible or not from a diachronic perspective) and realistic descendant of PIE possible, then you'll probably want to go with the most up-to-date scholarship. If you'll find it more enjoyable working with outdated theories, then go for it. Nobody except you can really weigh up the aim of ****ultimate plausibility**** with the aim of ease of conlanging, and as you probably realise, nobody but you will probably care in the end either way.

Just note that you'd better have access to a university library if you want to do Uralic, because the material you'll need is all over the place. And a reading knowledge of Russian wouldn't hurt, either (I have the former, but not the latter). There's a thread in L&L on Uralic historical linguistics you should check out, though bear in mind it'll end up in L&L museum sometime within the next few days.

Actually, you'd better have access to a university library in any case, if you want to do a language derived from an existing one or a reconstructed protolanguage. I'm doing a language set in a not very obscure branch (Italo-Celtic) of IE, which is after all the best-documented and most-researched language family in the world, working with a six million-volume university library- and I have to special-order some pretty damn obscure books. If you don't have a university library, about the only family you can hope to make a reasonable member of is Romance, and even there you have to be careful about where in the family your project is located.

FearfulJesuit wrote:If you don't have a university library, about the only family you can hope to make a reasonable member of is Romance, and even there you have to be careful about where in the family your project is located.

And it is overdone. Most if not all of the more plausible possibilities are already taken, and most of the linguistically interesting ones are utterly implausible.

Does it matter if a possibility is 'taken'?

Salmoneus wrote:Does it matter if a possibility is 'taken'?

Not really, if you manage to get a different angle on it.

Wasn't there someone who already did Vampiric Sumerian? Or was that only a script?

Qwynegold wrote:Wasn't there someone who already did Vampiric Sumerian? Or was that only a script?

That was me - I haven't really had time to work on it lately, but it's been on my mind. I never got a firm enough grasp of Proto-Semitic to feel confident moving forward though, certainly not in comparison to my work with IE languages, or even East-Caucasic languages for that matter.

SOON


Porphyrogenitos:
If I get my server back up, I can get you a lot of material on IE, including a very up-to-date word-list, and a bunch of grammars. I can definitely advise you that nothing helped me learn about IE linguistics and the reconstructed grammar of PIE more than working on a conlang which preserved some of its more interesting features (namely, and active-stative typology, and fairly dynamic use of ablaut).

Goatface wrote:If I get my server back up, I can get you a lot of material on IE, including a very up-to-date word-list, and a bunch of grammars.

Ooh, yes please!

Thanks for all the additional responses, everyone! And I do have access to a university library system. I'll be sure to see what's available.

Goatface wrote:

Qwynegold wrote:Wasn't there someone who already did Vampiric Sumerian? Or was that only a script?

That was me - I haven't really had time to work on it lately, but it's been on my mind. I never got a firm enough grasp of Proto-Semitic to feel confident moving forward though, certainly not in comparison to my work with IE languages, or even East-Caucasic languages for that matter.

SOON


Porphyrogenitos:
If I get my server back up, I can get you a lot of material on IE, including a very up-to-date word-list, and a bunch of grammars. I can definitely advise you that nothing helped me learn about IE linguistics and the reconstructed grammar of PIE more than working on a conlang which preserved some of its more interesting features (namely, and active-stative typology, and fairly dynamic use of ablaut).

That would be very helpful; thank you!

KathAveara wrote:

Goatface wrote:If I get my server back up, I can get you a lot of material on IE, including a very up-to-date word-list, and a bunch of grammars.

Ooh, yes please!

Seconded.

Clıck wrote:

KathAveara wrote:

Goatface wrote:If I get my server back up, I can get you a lot of material on IE, including a very up-to-date word-list, and a bunch of grammars.

Ooh, yes please!

Seconded.

And thirded.

If I'm not missing something, folks are conflating two different senses of "proto-language" here.

Proto-language, in our conlang parlance, means "an older stage of the 'final' language I want to create". In the historical linguistics sense, it means "reconstructed common ancestor of two or more attested languages". Therefore, e.g. Sanskrit or Classical Latin, ancient as they are, are not proto-languages in the latter sense. While according to the former sense, Old English would be a "proto-language of English", or Old Japanese would be a "proto-language of Japanese".

If you're interested in doing a diachronic a posteriori language, there is no particular reason to start from a reconstructed example. If anything, that is going to greatly limit the amount of information at your disposal. You can well pick an attested language as your protolang and proceed from there. If you don't care about putting the language in a plausible modern-day setting (a "lostlang"), you could even start from a modern, well-described language, zap any vocabulary like "computer" sort out of it, and pretend to evolve it as if it were spoken thousands of years ago. Like start from Modern Turkish and apply a bunch of sound changes inspired by the development from Proto-Oceanic to Māori?

(Incidentally, the Proto-Polynesian / Proto-Oceanic / Proto-Malayo-Polynesian / Proto-Austronesian series is quite well reconstructed as well, IIUC.)

Another possible point against using reconstructed bottom-level protolangs is that a believable development of several millennia is goign to require working out half a dozen intermediate stages, a shifting palette of loanword influences, etc. But if the scenario is something like suddendly dropping Old Tupi into 17th century Korea, it's entirely OK to enforce some drastic sound and grammar changes that, at a more "natural" environment, would have taken a millennium.

Sanskrit might be one of the best compromises between ancientness, well-describedness, and sources-being-in-English-ness, actually. Yet I don't think I've ever seen Indic bogolangs?

I don't understand where you think that conflation is happening, or why it should matter.

Yes, handwaving a "Tupi gets dropped in 17th century Korea" is, in some ways, easier than evolving a realistic non-extant branch of Indo-European. But some people would prefer to do the latter than the former. Personally, I have no interest in the former because a) it requires the sort of conworld that doesn't much interest me, and b) it gives me little context to appreciate the language in. With an IE language, I can compare and contrast to real IE languages - with a Romlang, to other real Romance languages. But not being an expert in either Tupi or Korean, the context would be lost on me there, and I might as well just make a language from scratch.

Besides, resources for Indo-European are way more easily accessible than resources on Tupi or on 17th century Korean.

Tropylium wrote:If you're interested in doing a diachronic a posteriori language, there is no particular reason to start from a reconstructed example. If anything, that is going to greatly limit the amount of information at your disposal.

Surely, attested languages are easier to get sufficient data on than reconstructed ones, and major living languages are the best. But - the question of which source language to use depends on the setting the author is thinking about. AFAIK, most a posteriori diachronic conlangers choose the a posteriori solution not just because they want to avoid making up an a priori (proto-)language. They do so because they have an idea of what they could do in that family. And that informs their choice of the starting point language. Those of us who take the trouble of working with PIE do so because they want to do a new branch of IE, for whichever reason, and starting with some well-documented IE language such as Latin or Sanskrit is no option that makes sense in the intended scenario.

For instance, with Old Albic and its relatives, I work with a proto-language that is even more difficult to get information on than PIE, to the point that I have to reconstruct it all by myself - a language that is an ancestor of PIE and can be reached by internal reconstruction from PIE. I am doing this because I wish to explore a potential development of a structure I see buried in the prehistory of IE, and because it fits with my ideas about the linguistic landscape of prehistoric Europe.

Tropylium wrote:You can well pick an attested language as your protolang and proceed from there. If you don't care about putting the language in a plausible modern-day setting (a "lostlang"), you could even start from a modern, well-described language, zap any vocabulary like "computer" sort out of it, and pretend to evolve it as if it were spoken thousands of years ago. Like start from Modern Turkish and apply a bunch of sound changes inspired by the development from Proto-Oceanic to Māori?

Nobody says you can't do that - as an experiment. But most of us a posteriori diachronists are working on lostlangs or altlangs, where this doesn't make sense.

Tropylium wrote:Another possible point against using reconstructed bottom-level protolangs is that a believable development of several millennia is goign to require working out half a dozen intermediate stages, a shifting palette of loanword influences, etc.

Yes, all this has to be considered when developing a diachronic conlang from an ancient protolanguage. But that's all for the sake of realism, and therefore justified.

Tropylium wrote:But if the scenario is something like suddendly dropping Old Tupi into 17th century Korea, it's entirely OK to enforce some drastic sound and grammar changes that, at a more "natural" environment, would have taken a millennium.

Sanskrit might be one of the best compromises between ancientness, well-describedness, and sources-being-in-English-ness, actually. Yet I don't think I've ever seen Indic bogolangs?

Nor have I.

But I have grown out of bogolanging. It is extremely unlikely that the same sound changes happen twice in two different languages, and applying sound changes of one language to another always has to deal with the problem that the starting phonologies are virtually always different, so you have to bend your sound changes to make them fit.

WeepingElf wrote:

Tropylium wrote:Sanskrit might be one of the best compromises between ancientness, well-describedness, and sources-being-in-English-ness, actually. Yet I don't think I've ever seen Indic bogolangs?

Nor have I.

On a related note, I did start a New-Indo-Aryan Freaklang a year and a half ago. viewtopic.php?f=4&t=40459. Gotta love those first awkward posts . . .

If you want to mine the edges of IE-conlanging and specifically something other's haven't posted about much, then I agree. I think Indo-Aryan and Iranian is probably the way to go. Maybe Armenian and Albanian too?

WeepingElf wrote:. . . It is extremely unlikely that the same sound changes happen twice in two different languages, and applying sound changes of one language to another always has to deal with the problem that the starting phonologies are virtually always different, so you have to bend your sound changes to make them fit.

Maybe with broad 'sound laws,' but I think taking more common sound changes isn't that bad. For example, /coronal/ + /j/ -> palatalization, palatal POA, etc. To me at least, cherry picking sound changes isn't that bad if you are looking at them as a consistent whole and not trying to use them to shepherd a phonology towards X language's phonology or phonotactics (or worse, vocabulary).

I've found that with trying to get a particular flavor to a more or less a priori lang, it is often more productive to focus on the phonotactics and clusters than sound changes. I've definitely fallen in the trap of trying to get the appropriate word feel by stepping back to a proto-lang and deriving a few words with a sound change, only to find that it obliterates other words, clusters, phonemes, etc. in the targetLang that I liked previously.

I've found this board that doesn't seem to be active anymore but focused on IE altlangs. Maybe you could find some more resources through them if you can track users down?

kanejam wrote:I've found this board that doesn't seem to be active anymore but focused on IE altlangs. Maybe you could find some more resources through them if you can track users down?

We're still somewhat active. Ralph posted some new material just last week. I haven't posted at all since June and I've posted nothing of my conlangs (predominantly Dange) since late January. DePaw posted a few things in May/June and Zontas, Eldin and Rickspaghetti joined around that time as well, but the board has been lacking activity since late 2010 (if you look at the latest posts in the Languages of the World section, it's consistently either me or Ralph in every thread bar DePaw's since January 2011) and our time of greatest "diversity" was back in late 2009.

There are some good resources in the Resources thread in the General Board section and I've got some PDFs I don't mind emailing to people (The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World by Mallory and Adams for one) and I think Ralph and Thak (who you're probably likely to find on the CBB) might have some stuff as well

sangi39 wrote:

kanejam wrote:I've found this board that doesn't seem to be active anymore but focused on IE altlangs. Maybe you could find some more resources through them if you can track users down?

We're still somewhat active. Ralph posted some new material just last week. I haven't posted at all since June and I've posted nothing of my conlangs (predominantly Dange) since late January. DePaw posted a few things in May/June and Zontas and Rickspaghetti joined around that time as well, but the board has been lacking activity since late 2010 (if you look at the latest posts in the Languages of the World section, it's consistently either me or Ralph in every thread bar DePaw's since January 2011) and our time of greatest "diversity" was back in late 2009.

There are some good resources in the Resources thread in the General Board section and I've got some PDFs I don't mind emailing to people (The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World by Mallory and Adams for one) and I think Ralph and Thak (who you're probably likely to find on the CBB) might have some stuff as well

Wat. This is still active? I had a PIE lang on there, but I never got very far but for some sound changes...

I should go look this stuff up and do some comparison with my current PIE lang and maybe rip off myself a bit.

Bristel wrote:
Wat. This is still active? I had a PIE lang on there, but I never got very far but for some sound changes...

I should go look this stuff up and do some comparison with my current PIE lang and maybe rip off myself a bit.

Yeah, it never went away, it's the people that left. Dengweas (the one post on it anyway), is still there from 2010

sangi39 wrote:

Bristel wrote:
Wat. This is still active? I had a PIE lang on there, but I never got very far but for some sound changes...

I should go look this stuff up and do some comparison with my current PIE lang and maybe rip off myself a bit.

Yeah, it never went away, it's the people that left. Dengweas (the one post on it anyway), is still there from 2010

It was quite unfinished, in fact I am fiddling around with it now just to see what it looks like, and it only seems to be a few hundred years of sound changes, so I might do more, or just let Dengweas die out.