Naturalistic conlang: Where should a protolang start?

If you're new to these arts, this is the place to ask "stupid" questions and get directions!
Post Reply
WaterHat
rupestrian
rupestrian
Posts: 2
Joined: 11 Jul 2021 16:16

Naturalistic conlang: Where should a protolang start?

Post by WaterHat »

I've been researching to create a naturalistic conlang, starting to think up a conworld and a conculture, but one thing that has been eluding me is where do I actually start with the proto-language? How "developed" should it be? For every grammatical feature I think of, I can also think, "hey, this grammatical feature probably originated as something else, should I start it out as that instead?"

I'd assume that this kind of thing would be up to me, but I'm somewhat of a perfectionist, and want my language to be "properly" naturalistic. How far should I go with this and where do I start?
User avatar
Creyeditor
MVP
MVP
Posts: 4438
Joined: 14 Aug 2012 19:32

Re: Naturalistic conlang: Where should a protolang start?

Post by Creyeditor »

A proto-language is not substantially different from any other language. So you can start creating one naturalistic conlang that is not diachronically justified and the derive daughter languages from it, which are diachronically justified.

I usually work backwards. I create a naturalistic conlang, look for diachronic explanation for its quirks and then create the protolang by internal reconstruction. Only then I derive new daughterlangs, i.e. sisterlangs for the first conlang.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
dva_arla
cuneiform
cuneiform
Posts: 150
Joined: 25 Oct 2019 21:03
Location: Realm of Ideas

Re: Naturalistic conlang: Where should a protolang start?

Post by dva_arla »

One thing you'd want to do is to look into real-world proto-languages: PIE, Proto-Turkic, Proto-Bantu (or if you wish to go much earlier, Proto-Niger-Congo), Proto-Mayan, etc. This will possibly provide you a picture of what needs to be made.

Are you working at an a priori language, or an a posteriori one?
Salmoneus
MVP
MVP
Posts: 2500
Joined: 19 Sep 2011 19:37

Re: Naturalistic conlang: Where should a protolang start?

Post by Salmoneus »

Ah, the eternal question!

I think it's worth bearing in mind that this doesn't have to be dichotomous: as well as actually constructing a language, you can sketch it, or just have ideas about it. I guess the 'perfect' thing is to have gradually less and less idea about mother, grandmother, greatgrandmother (etc) languages as you go back. Often this isn't about knowing everything about the parent, but maybe having some key ideas about it.

If our conlang were English, how far back would we have to reconstruct? Well, it would be very helpful to have a fairly full sketch of Middle English, not least because that's the language our spelling chiefly reflects!

Before that? Well, I certainly wouldn't discourage reconstructing everything back to PIE - if you enjoy it and have the time. But you could probably create a fairly convincing English with just some particular notions about what had come before.

The big specific thing is probably the development of strong verbs. From the point of view of modern English, strong verbs are an irregular, nightmarish minefield... but diachronically, most of them can be developed straightforwardly from PIE ablaut. So you'd probably want to have an idea about how that happened and the sort of changes that were involved - and you could, say, sketch out the Proto-Germanic strong vowel patterns, which are much more predictable and regular than the modern English forms, while still being understandable in terms of how they could have been generated. And you could have some ideas about key developments after PGmc.

Syntactically, meanwhile, it would be great to have the idea of gradual shift from SOV toward SVO, via a fairly rigid V2.

But you wouldn't need to know every detail of Old English, let alone Proto-Germanic or PIE.


A good guide is probably to think: is this making things easier? Some diachronics does make things easier for you: it's actually easier (though it does't always seem like it) to, say, derive a massively complicated verb class system from a simple origin, than it is to just pull it entirely out of your head word-by-word. But a lot of diachronics is just making work for yourself: if something in a parent language will have absolutely no relevance in the daughter language, why bother creating it?



----

Of course, things get rather more difficult if you intend to create closely-related languages, rather than just one in isolation...
WaterHat
rupestrian
rupestrian
Posts: 2
Joined: 11 Jul 2021 16:16

Re: Naturalistic conlang: Where should a protolang start?

Post by WaterHat »

Thanks for the advice, everyone! I think I have a better idea of where to start now.

What I want to make is a priori, set in a fictional conworld (possibly with fantasy elements). The intention is, if I manage to keep the motivation, to eventually make this a large worldbuilding project with a couple language families. I also like the idea of creating a protolang without explicit goals for the daughter languages, and deciding on those later, so I will likely make the protolanguage a detailed language. I'll make sure to research real-life protolanguages for inspiration as well.
Last edited by WaterHat on 29 Jul 2021 00:53, edited 1 time in total.
User avatar
Dormouse559
moderator
moderator
Posts: 2902
Joined: 10 Nov 2012 20:52
Location: California

Re: Naturalistic conlang: Where should a protolang start?

Post by Dormouse559 »

WaterHat wrote: 12 Jul 2021 12:19 Thanks for the advice, everyone! I think I have a better idea of where to start now.

What I want to make is a posteriori, set in a fictional conworld (possibly with fantasy elements). The intention is, if I manage to keep the motivation, to eventually make this a large worldbuilding project with a couple language families. I also like the idea of creating a protolang without explicit goals for the daughter languages, and deciding on those later, so I will likely make the protolanguage a detailed language. I'll make sure to research real-life protolanguages for inspiration as well.
Hi WaterHat! I thought I'd reiterate what Creyeditor said about protolanguages, which is that they aren't fundamentally different from other languages — at least as the term is often used among conlangers, where it's a near synonym of "parent language". By that definition, Modern English will probably be a protolanguage a few centuries from now. Meanwhile, in linguistics, a protolanguage is specifically a reconstruction of an unattested common ancestor for various attested languages. In that case, the label just tells you that no writings or recordings of the language have survived and that it has modern descendants, nothing about the language itself. (But by all means, feel free to check out some protolanguages; they're quite interesting, just not the only way for a conlanger to inform a parent lang.)
Post Reply