[First post! ]
I have a conworld that consists of a single, small-ish continent and I want to create a proto-language for the whole world. I'd like to divide it up into ~5 subgroups and then derive fully fleshed-out languages from there. I have a few questions:
1) How developed would you recommend I make the proto-language?
2) Where could I find resources on common sound changes? Do certain changes tend to occur together?
3) This might be more of a conworlding question, but what time fram would be appropriate? I know PIE is hypothesized to have been spoken about 4500 years ago, but my proto-language isn't going to be as far-flung.
Thanks!
Creating a Proto-language
Re: Creating a Proto-language
I recommend that the proto-language be reasonably developed. You should be able to get most of your words from this languages. I think a reasonable inventory of words is maybe 300 words. You should also have variant forms that are reconstructible. There should be derivational morphology and some degree of nominal and verbal morphology.KatyaZ wrote:[First post! ]
I have a conworld that consists of a single, small-ish continent and I want to create a proto-language for the whole world. I'd like to divide it up into ~5 subgroups and then derive fully fleshed-out languages from there. I have a few questions:
1) How developed would you recommend I make the proto-language?
2) Where could I find resources on common sound changes? Do certain changes tend to occur together?
3) This might be more of a conworlding question, but what time frame would be appropriate? I know PIE is hypothesized to have been spoken about 4500 years ago, but my proto-language isn't going to be as far-flung.
Thanks!
We have a great document called the Index Diachronica. While, right now, the Index is offline, excepting for old mirrors, a person made a HTML page with the latest known version of the Index Diachronica. It's searchable, and is one of the best sources for diachronic linguistics. It was made through a large collaborative project headed by Linguifex. I at one point participated in adding information regarding a Romance language called Sardinian.
The oldest reconstructed language to my knowledge is Proto-Semitic which is dated as far back as 16000 years ago, to 9500 years ago. Dene-Yeniseian may be comparatively old dating to before the Beringia land bridge was submerged around 10000 years ago. If you want a family that's recognizably different, I'd highly suggest that an age of less than 2000 years be used. That's roughly the age of the modern Germanic languages, which, excepting English, are quite noticeably similar. To my knowledge, Romance is around the same age, maybe a couple hundred years younger, at maybe 1500 years old.
Spoiler:
- Frislander
- mayan
- Posts: 2088
- Joined: 14 May 2016 18:47
- Location: The North
Re: Creating a Proto-language
I'm pretty sure by that you're talking about Proto-Afro-Asiatic, of which Semitic is a much younger subbranch. I'm not sure it's been all that well reconstructed yet.qwed117 wrote:The oldest reconstructed language to my knowledge is Proto-Semitic which is dated as far back as 16000 years ago, to 9500 years ago.
Re: Creating a Proto-language
Ah yes. That was a mistake that I meant to correct. Thanks.Frislander wrote:I'm pretty sure by that you're talking about Proto-Afro-Asiatic, of which Semitic is a much younger subbranch. I'm not sure it's been all that well reconstructed yet.qwed117 wrote:The oldest reconstructed language to my knowledge is Proto-Semitic which is dated as far back as 16000 years ago, to 9500 years ago.
Spoiler:
Re: Creating a Proto-language
What exactly do you mean by variant forms? Like how PIE 1st person pronouns had two distinct roots?qwed117 wrote: I recommend that the proto-language be reasonably developed. You should be able to get most of your words from this languages. I think a reasonable inventory of words is maybe 300 words. You should also have variant forms that are reconstructible. There should be derivational morphology and some degree of nominal and verbal morphology.
Re: Creating a Proto-language
Yes, like how the 1st person pronouns had distinct roots, like the number 1 (Hoi-no-/*Hoi-wo-/*Hoi-k(ʷ)o-). They create the appearance that the proto-forms aren't 100% certain, and give it a more realistic dimension.KatyaZ wrote:What exactly do you mean by variant forms? Like how PIE 1st person pronouns had two distinct roots?qwed117 wrote: I recommend that the proto-language be reasonably developed. You should be able to get most of your words from this languages. I think a reasonable inventory of words is maybe 300 words. You should also have variant forms that are reconstructible. There should be derivational morphology and some degree of nominal and verbal morphology.
Spoiler:
Re: Creating a Proto-language
Even in reality the proto-languages have no assured existence ...
These are only a posteriori constructions (some sort of more serious conlang ...)
The synchronous creations do not really need to be doubled with diachronic cobstructions ... This new game can advantageously be played after ... to immerse itself even longer in the reassuring world of language-based creation ...
These are only a posteriori constructions (some sort of more serious conlang ...)
The synchronous creations do not really need to be doubled with diachronic cobstructions ... This new game can advantageously be played after ... to immerse itself even longer in the reassuring world of language-based creation ...
Re: Creating a Proto-language
I'd agree with Qwed on this one The majority of reconstructed proto-languages don't really seem to have many more than 1,000 words that are considered "securely" reconstructed.
The general idea is that the more time passes, the fewer and fewer words from the source language will survive, and which words drop out varies from language to language. This means that, for example, a Proto-Indo-European that survived solely into Proto-Germanic cannot be securely reconstructed as far back as PIE (we could propose a possible PIE form, but without cognates in other IE languages it would effectively be an educated guess). On the other extreme, I seem to recall some PIE roots that are only attested in Anatolian and Tocharian and the in Celtic, meaning that while we might be able to reconstruct a PIE form, it was lost in almost all IE languages.
The advantage here, as Qwed points out (if I'm reading things correctly), is that you then don't have to have a fully formed proto-language from which to derive all the vocabulary of daughter languages. You can throw words in at any stage of the family's evolution and just have it exist solely in one or a handful of branches.
Similarly, some morphemes might survive only in a handful of branches, or the same for some grammatical construction. So you don't need a fully fledged morphology or syntax either. You can flesh those out individually in each various branch as you go.
The main thing to remember, though, is that older, more closely related languages will likely be similar, so you can't just throw features or words in willy-nilly. The best option would probably be to start with the basics in the proto-language, then gradually add to them and shift them around as you work on different branches, always keeping in mind what you've done previously in other branches.
The general idea is that the more time passes, the fewer and fewer words from the source language will survive, and which words drop out varies from language to language. This means that, for example, a Proto-Indo-European that survived solely into Proto-Germanic cannot be securely reconstructed as far back as PIE (we could propose a possible PIE form, but without cognates in other IE languages it would effectively be an educated guess). On the other extreme, I seem to recall some PIE roots that are only attested in Anatolian and Tocharian and the in Celtic, meaning that while we might be able to reconstruct a PIE form, it was lost in almost all IE languages.
The advantage here, as Qwed points out (if I'm reading things correctly), is that you then don't have to have a fully formed proto-language from which to derive all the vocabulary of daughter languages. You can throw words in at any stage of the family's evolution and just have it exist solely in one or a handful of branches.
Similarly, some morphemes might survive only in a handful of branches, or the same for some grammatical construction. So you don't need a fully fledged morphology or syntax either. You can flesh those out individually in each various branch as you go.
The main thing to remember, though, is that older, more closely related languages will likely be similar, so you can't just throw features or words in willy-nilly. The best option would probably be to start with the basics in the proto-language, then gradually add to them and shift them around as you work on different branches, always keeping in mind what you've done previously in other branches.
You can tell the same lie a thousand times,
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.
Re: Creating a Proto-language
I don't think I said anything like that, but I think that's a great point to keep in mind.sangi39 wrote:I'd agree with Qwed on this one The majority of reconstructed proto-languages don't really seem to have many more than 1,000 words that are considered "securely" reconstructed.
The general idea is that the more time passes, the fewer and fewer words from the source language will survive, and which words drop out varies from language to language. This means that, for example, a Proto-Indo-European that survived solely into Proto-Germanic cannot be securely reconstructed as far back as PIE (we could propose a possible PIE form, but without cognates in other IE languages it would effectively be an educated guess). On the other extreme, I seem to recall some PIE roots that are only attested in Anatolian and Tocharian and the in Celtic, meaning that while we might be able to reconstruct a PIE form, it was lost in almost all IE languages.
The advantage here, as Qwed points out (if I'm reading things correctly), is that you then don't have to have a fully formed proto-language from which to derive all the vocabulary of daughter languages. You can throw words in at any stage of the family's evolution and just have it exist solely in one or a handful of branches.
Similarly, some morphemes might survive only in a handful of branches, or the same for some grammatical construction. So you don't need a fully fledged morphology or syntax either. You can flesh those out individually in each various branch as you go.
The main thing to remember, though, is that older, more closely related languages will likely be similar, so you can't just throw features or words in willy-nilly. The best option would probably be to start with the basics in the proto-language, then gradually add to them and shift them around as you work on different branches, always keeping in mind what you've done previously in other branches.
Funny off the side thing; both Germanic (South, North, Istvaeonic, Ingvaeonic, and Irminonic) and Romance (Southern, Eastern, Iberian, Gallic, Italo-Dalmatian) have 5 divisions in certain schemes. It's not 100% telling, and I believe Afro-Asiatic has a similar 5 part division (Cushitic, Berber, Egyptian, Semitic, Chadic).
Spoiler: