How quickly can you sketch out a grammar?

Substantial postings about constructed languages and constructed worlds in general. Good place to mention your own or evaluate someone else's. Put quick questions in C&C Quickies instead.
Post Reply
User avatar
Jonlang
Avisaru
Avisaru
Posts: 305
Joined: Tue Mar 14, 2006 11:21 am
Location: Cymru
Contact:

How quickly can you sketch out a grammar?

Post by Jonlang »

I feel as if I've been working on Quebric for a rather long time, and I'm kinda thinking that some of it may be over-complicated, maybe it isn't... I dunno. Now I have revisited some things more than once to revise them or completely redo them, so that's taken some time, but I feel like it's never ending. So I'm wondering, how long would it take you to sketch out a grammar for a new conlang (or family) where you'd be comfortable with stringing a few basic sentences together?
My conlangery Twitter: @Jonlang_
Me? I'm just a lawn-mower; you can tell me by the way I walk.

User avatar
Ser
Smeric
Smeric
Posts: 1542
Joined: Sat Jul 19, 2008 1:55 am
Location: Vancouver, British Columbia / Colombie Britannique, Canada

Re: How quickly can you sketch out a grammar?

Post by Ser »

dyolf wrote:I feel as if I've been working on Quebric for a rather long time, and I'm kinda thinking that some of it may be over-complicated, maybe it isn't... I dunno. Now I have revisited some things more than once to revise them or completely redo them, so that's taken some time, but I feel like it's never ending. So I'm wondering, how long would it take you to sketch out a grammar for a new conlang (or family) where you'd be comfortable with stringing a few basic sentences together?
If the immediate goal is stringing basic sentences together, technically speaking, it can take me (and you too) nothing.

As H13 says in his first sermon, even though all the conlanging manuals recommend some procedure where you start with a phonological inventory and continue onto inflectional (and some derivational) morphology followed by basic syntax, you don't have to follow that procedure. H13 proposes starting with the details that you yourself like about your language, whether phonological or syntactical or semantic etc, working out more details later as you go. A procedure where you start with whole nonce paragraphs iteratively working out something you like is mentioned by Rhetorica under that sermon, too.

Let me illustrate this with a language I've never worked on until this post. I'm imagining a language with heavy use of articles, cases like nominative and accusative, with a singular-plural distinction, but where noun incorporation is productively used instead of independent nouns for plural nouns of very indefinite reference. There are three articles distinguishing three levels of definiteness (u "a thing", ir "a certain thing", dan "the thing").
  • 1. ui dan drambr-a palan
    1SG DEF.ART apple-SG.ACC eat
    'I'm eating the apple'

    2. ui u drambr-a palan
    1SG INDEF.ART apple-SG.ACC eat
    'I'm eating an apple.'

    3. ui dan drambr-u palan
    1SG DEF.ART apple-PL.ACC eat
    'I'm eating the apples.'

    4. ui drambr-i-palan
    1SG apple-INCORPORATION-eat
    'I'm eating apples.'
    (No article before drambr-i- even though we normally see articles used with direct objects, as it's incorporated onto the verb. To further strengthen my case that this is incorporation into one grammatical word, and not the presence a strangely word-final article -i, this is also pronounced as one phonological word: [dɾam.bɾi.paˈlaŋ].)
Noun incorporation of generic, dummy nouns is also commonly used to make intransitive versions of some verbs that are often transitive (such as "eat-bread" for intransitive "eat", "hit-skin" for a use of "hit" that skips the direct object).

All this together could make something relatively interesting if further worked out, and I haven't even thought of what phonemes I'll have yet, what their allophones are (other than /n#/ [ŋ]), or what the full declension of drambr-e 'apple-SG.NOM' is.

User avatar
kusuri
Lebom
Lebom
Posts: 107
Joined: Wed Dec 04, 2013 3:57 pm
Location: 此の市、其の州、彼の国

Re: How quickly can you sketch out a grammar?

Post by kusuri »

It depends on the conlang. It takes me a while because I like word creation more than text creation, so I tend to have a dictionary ready before a grammar, and the only time I care to make sentences is usually when I want to test out a script. A lot of my conlangs are remakes and variations, and one particular strain has been active for years, so I don't know if it's fair to say if my latest version is one month old, or dozens.

User avatar
Hallow XIII
Avisaru
Avisaru
Posts: 846
Joined: Sun Nov 04, 2012 3:40 pm
Location: Under Heaven

Re: How quickly can you sketch out a grammar?

Post by Hallow XIII »

My revised & extended advice is to just make shit up and integrate it into a coherent system later. As a worked example, let's say I wish to make a language spoken by a people who live in a very mountainous area. So now I can decide some interesting cultural tidbits already, without knowing anything else, that can already make a large difference in how my language feels. Since they live high in the mountains, their language could have a pervasive distinction between doing things uphill and doing things downhill, so let's say they just mark all verbs for that. So I already have a nice little detail about the language that significantly changes the way everything is expressed, because it adds an obligatory distinction.

Since I'm already thinking about this, I might just as well apply it to actually producing sentences. So let's say my conpeople have a children's story about one of the tallest mountains in their territory. Let's call it something like "Why North Father Faces North". So I already know two things: the theme of the story, and the fact that there is a tall mountain called North Father. So what should I call north? Well, maybe the land they live on has a higher average altitude in the south than in the north, so I could relate "north" to "down". So let's give the north a name, and call it afɯn. You'll note I'm just making up words -- a systematization of the phonology can come later.

I could go write the story first and then translate it, but that takes time and I want to make up a sentence now so I can get a feel for the language. But since I already have the title of the story, I can guess that the last sentence might simply be "and thus it is that North Father faces north". Okay, so I can already do a sentence, and it's interesting because it has backreference to what was said before and a subordinating structure. Okay, so what might this be in my language? Let's say it's something like "North Father TOP always face_north CMPL its_reason this they_say". Nice and basic sentence template I can fill my morphology into.

The first thing to do is decide on how to do possession and compounding, since we need to be able to express both "its reason" and "North Father". I'll say possession is head marking, and the same structure is used for what in English is compounding and more loose coupling. So then we have "north its_father". Let's say "father" is lɯm, and the third-person possessive prefix is du-, and I have afɯn du-lɯm. I might as well do the other simple words now, so I'll say "its reason" is du-ʔudɯk, "always" is hone and "this" is ea.

So now I need to look at the verbs. I need to decide how I want to indicate that one sentence is the complement of another. I could do a nonfinite verb form, but I don't want to make a language that has a lot of those right now, so I'll just make the complementizer a separate word. What do I use for that? There are lots of different grammaticalization paths attested, such as a demonstrative (English), the verb "say" (Uzbek) or "thing" (Korean, Chintang). Since I'm lazy and like to copy Asian languages, I'll just use the latter, and make a word for "thing": mul. So what about the verb? Who cares, let's just say the form is foɲaʔefi, and I'll try to find a nice segmentation of this. I could make it a long root and a person affix, like foɲaʔe-fi or something, but I'm feeling like synthesis right now, so let's make that a bit bigger and throw some funky morphophonology in: fa-w-ni-a-ʔwi-fi. Nice.

Okay, now I have a really long string of nonce morphemes, so let's try to have this make sense. I already said that all verbs are marked for the upward/downward axis, and since "north" is identified with "down", let's say one of these affixes means that. "North" is afɯn, and we have an a- in here, so I'm going to be utterly shameless and just say those are related, so that means "downward" now. Okay, what with the rest? Maybe I could fix the root. A nice way to obscure verbal paradigms is to have bipartite stems, so let's say fa-ʔwi is the stem. If I have a complicated stem, I might as well complicate the derivation, so let's say that instead of "face north" the verb literally means "look downward". I'll liberally steal from Algonquian and Athabaskan and say that the stem on its own is "light up" and somewhere in there there's an affix that means "by eye", so "light up by eye" > "look at". In fact, fuck it, make the morpheme more general and just have it mean "with a round object". Let's say that's ni. So now I have two morphemes left, and I suppose I should be doing person marking and TAM. For person marking, let's say third-person is zero-marked, and have that w- be an Antipassive marker, since North Father isn't looking at anything in particular. That leaves the -fi, which could be a straightforward imperfective or present marker, but that's boring, and since I'm already doing weird templatic morphology, I'll say that TA(M) is organized in a georgian-like screeve system and -fi is a meaningless suffix that appears in some TAM categories for this verb class. So I have: Σ-3-ANTIP-by_eye-down-light-THM. Nice.

So the other verb we have is "they say it", or perhaps rather "it is said". This time the structure is already pretty clear, since the TAM is broadly similar to the last verb. Maybe one thing I can do is not have a thematic ending in the passive, and have something like Σ-3-by_mouth-PASS-down-say, with "down" as a default direction for "say" (since all verbs have to have a direction marker). Let's say this is asaroe, with segmentation a-∅-sa-ra-a-we. And actually, while I'm at it, I might as well turn "reason" into some sort of verbal derivative too: "that which makes it so". Let's just have that be a headless relative clause, so no thematic again: du-u-∅-∅-de-ɯk = 3POSS-thus-3-3-up-make, with the glottal stop epenthetic. "Up", again, is a default direction for "make". Nice.

So in the end, I get:
afɯndulɯm e hone foɲaʔefi mul duʔudɯk ea asaroe
afɯn
north
du-
3POSS-
lɯm
father
e
TOP
hone
always
fa-
PVB-
∅-
3-
w-
ANTIP-
ni-
with_round_object-
a-
down-
ʔwi
light_up
-fi
-THM
mul
thing
du-
3POSS-
u-
thus-
∅-
3-
∅-
3-
de-
up-
ɯk
make
ea
this
a-
PVB-
∅-
3-
sa-
by_mouth-
ra-
PASS-
a-
down-
we
say

"And so it is that North Father faces north."

Apart from the time I took to write this post, that took maybe 15 minutes to come up with. Making shit up and going with the flow is a p powerful method.

EDIT: apparently the gloss tag doesn't like capital sigmas in glosses. terrible
陳第 wrote:蓋時有古今,地有南北;字有更革,音有轉移,亦勢所必至。
R.Rusanov wrote:seks istiyorum
sex want-PRS-1sg
Read all about my excellent conlangs
Basic Conlanging Advice

User avatar
Dama Diwan
Sanci
Sanci
Posts: 37
Joined: Wed Jan 13, 2016 8:42 am

Re: How quickly can you sketch out a grammar?

Post by Dama Diwan »

I have made a description of Dama Diwan grammar in 4 lines on an A4 page, and by using conventional abbreviations it can be shorter. E.g.
-o =NMLZ/PASS, -o-n =ACC, -a-n=GEN, -e=ACT, -e-n=ACT TRANS, -a=ADV; head-last direction.
But, to make it easily and well understood by giving examples, a good description has taken 12 lines of a fb post, in https://www.facebook.com/groups/omado.s ... 776113744/ (pinned post of the group).
But, sorry to say, i find that the universal glossing system https://www.eva.mpg.de/lingua/resources ... -rules.php is not adequate for Dama Diwan, because -o-n is not simply accusative: it is object placed BEFORE the verb. And -e-n is not simply ACT TRANS: the suffix means that the object will be AFTER the -EN verb (whatever the object is, a noun, a secondary verb, or a whole sentence). Do you think that the Leipzig Glossing Rules are sufficient for all languages, even for all conlangs made by this forum members? If so, how can i gloss the Dama Diwan suffix -N after a verb (-E) and after a noun (-O)?
i kiwa jenon naje, nake, nibe, sake ka:
https://www.facebook.com/groups/omado.sosti.matiko

Post Reply