The Miniature Conlangs Thread

[Tropylium]

One thing I want to try sometimes would be a small (2-4 classes), universal, nominally semantically rooted set of noun classes, but one that's neither gender nor animacy-based.
- Black/white/red is what I've been considering primarily. I imagine the last would be generalized to include e.g. most red-blooded animals, unless conspicuously colored (crows, swans).
- Flat/round/oblong/(rectangular?) seems to be a common division among noun classifiers. Might be difficult to extend to abstracts though.
- A highly dualistic yin/yang approach might be interesting but I'd need to look more into Daoism before attempting a riff-off of that.

[Click]

I was thinking about reworking my main conlang Kaıpó, when the following idea popped into my head: patientive nominalisation is done by means of suffixing the non-speaker pronoun on the verb that is to be nominalised, e.g. ano 'catch' → anoı 'catch; haul; prey'.
This makes zero sense on a first glance, but there's actually some pretty solid logic behind it if one's acquainted with passivisation in Kaıpó.
In essence, a clause is made passive by simply kicking the agent out of the sentence and leaving everything else in place, so that the former patient is understood to be promoted without any special marking, as shown below.

• to ano ı
  SP catch NSP
  'I catch it'

becomes

• ano ı
  catch NSP
  'it is caught'

Notice the non-speaker pronoun immediately after the verb?
As the language gradually changed over time and the ancestor language's patientive nominaliser eroded away, the pronoun came to be reinterpreted as a brand new patientive nominaliser of sorts while still remaining independent in passive constructions.

[gach]

There might be another layer of agreement, as well, where you get subject verb agreement, but then also subject - C agreement, so in a clause where the subject moves to Spec-CP, you get something that looks like this:

S Aux-Agr.S V-Agr.S O

But when the object moves to Spec-CP you get the following:

O Aux-Agr.O S V-Agr.S

This reminds me of a few examples I've seen from a language that has both subject and topic agreement on its verbs at the same time. I'm not by my bookshelf now so I can't find the name though.

[gach]

- Flat/round/oblong/(rectangular?) seems to be a common division among noun classifiers. Might be difficult to extend to abstracts though.

On the other hand, think of the possibilities in playing with the conceptual metaphors related to shape.

[cromulant]

Kuty will have "split-A" morphosyntactic alignment:

Code: Select all

1. nomino-dative
2. ergative
3. accusative

         1(S)
   1~2(A)   3(P)
2(D)     1(R)     3(T)

With some (most) verbs (call them "type 1"), the transitive agent aligns with the ditransitive recipient or the benefactive argument; with other ("type 2") verbs, the agent aligns with the ditransitive donor, or the causative argument. The distinction between types 1 and 2 verbs has to do with how affected the agent is, how much of an "experiencer" that agent is. Note that it has absolutely nothing to do with volition (although type 2 agents will always be volitional).

Kuty will also have a wacky system of personal deixis involving simple (singular) and compound (plural...and singular!) person distinctions. In other words, there will be 1+2.SG, 1+2.PL, 1+3.SG, 1+3.PL, 2+3.SG and 2+3.PL in addition to your basic 1/2/3.SG. Semantically, in the compound singulars, the higher-numbered member is the actual referent, e.g. 1+2.SG = 2.SG, but 1+2.SG is transparently related to 1+2.PL, while plain 2.SG is not.

Verbal person marking involves three different person slots in the verbal template filling in the grammatical role of that person (probably through tone; I don't yet know how this system arose). A 1+2.SG agent would have both the first and second person slot indicating an agentive role for those two persons.

The semantic difference between the simple and compound singular persons probably has to do with familiarity, or the importance of the referent in the speaker's life.

[KathTheDragon]

I'm thinking of making a language where the case of the arguments is marked on the noun, and tense-aspect information is marked on the arguments. I might even try and make it triconsonantal.

[Jess]

One lang of mine, Ainure, has adpositions that agree with their dependents in number and case. Sometimes this works just like Celtic's inflected prepositions, but the inflection is still obligatory when the object of the adposition is separately specified, which fits well with Ainure's obviative "fourth person" I think.

[Drydic]

I'm thinking of making a language where the case of the arguments is marked on the noun, and tense-aspect information is marked on the arguments. I might even try and make it triconsonantal.

If you do, please don't have root-internal vowel alternations distinguish person or number entirely by themselves ><

[KathTheDragon]

I won't. I've seen enough examples of how not to do it.

[Chagen]

I tried this idea out for Qeung Wo once, it completely died in the cradle because I honestly have no idea if it even makes sense or not:

-Completely isolating with almost nothing in the way of inflections, such duties being handled by post-positional particles

-Three genders: Masculine, Feminine, and Neuter

-Post-positional articles (def. indef. part. and neg.), each corresponding to each gender; articles are only used when introducing NEW information, at which point they're dropped

-Verbs can be nominalized by adding a definite article to them (this lets one do things like make relative clauses). You can add an indefinite one to ghetto-mark the subject of an action if needed

-Word is highly important; the language is mostly V2 and many different word orders are used, each one with specific semantic functions (subject and object particles are super-rare thanks to this)

-Adjectives (that aren't loanwords) are verbs--nominalized by articles they can be put with nouns (articles are clitics that attach to the end of a phrase)

Some examples:

Rao wa Fui
1SG speak 3SGF
I speak to her

Ruiin de-nga chuu Rao xuos nu.
boy=towards=DEF.MASC throw 1SG ball=DEF.NEU
I throw the ball to the boy.

(in the actual lang, stuff like "de-nga" with collapse into "dega")

Ruiin jhung nu.
boy sing=DEF.NEU
The singing boy/The boy who is singing

Ruiin jhung nga-qi leum nu.
boy sing DEF.MASC=GEN mother=DEF.NEU
The boy who is singing to his mother

Zuij kun mi
girl be.smart=DEF.FEM
The smart girl

Ruiin de kun wo liu.
boy=towards be.smart=NMLZ=NEG.MASC
A boy who is not smart (lit. "A boy to whom [there is] no state of being smart"--without the "wo" the sentence makes zero logical sense--the verb HAS to be nominalized in this case.)

Jhung de wo nu oj chis
sing=towards NMLZ DEF.NEU be nice
Being sung to is nice. (Literally "The state of something singing to [you] is nice"--there is no Passive, just saying something like "Something sang to me" for "I was sung to", while "I was sung to by Mira" would be "I receive Mira's singing")

[Buran]

My current project (Proto-Tovlisic) has relativisation marked on the verb, with two different suffixes for relativising the object and the subject (à la Welsh relative clauses). Common nouns that are the subjects/objects/indirect objects of verbs cannot appear without an accompanying pronoun. Infixes make everything extra fun. Here's a sample:

To k'ómongōnler ga hálsap hārólsk'e gep dlṓskap.
To k'-ó-mon-gōn-ler g-a h-á-lsa-p h-āró-lsk'e ge-p dlṓs-ka-p
tʰo ˈk'omongoːnler ga ˈhalsapʰ haːˈrolsk'e gepʰ ˈdloːskʰapʰ
1S cut.1S-PAST-REL.OBJ 3PL.ACC pine.ACC-PL mountain.ABL 3PL long-STAT-PL
"The trees that I cut down on the mountain are tall."

I'm thinking about having the pronoun system collapse into prefixes/suffixes in some daughters, merge with nouns into triliterals in another, and in yet another, get expanded and take over most noun inflection.

[Chagen]

So this is a system for a Dorishar language I came up with:


The main important thing is a telicity distinction. The atelic aspect indicates that an action has no real set goal, while the telic does (and that the goal was achieved).

For instance, the verb niṣ- "to see" is normally atelic (marked by a lowered-grade root suffixed with the pure root vowel and then personal endings):

neṣika: I see
neṣimi: you see
neṣiṭu: he/she/it sees

But the verb khuṣ- "to finish" is telic (marked with an ablauted root, suffix -yo, then personal endings), because the telic aspect indicates that an action is finished (and therefore obviously "to finish" would be telic):

khiṣyoka: I finish
khiṣyomi: you finish
khiṣyoṭu: he/she/it finishes

This is not a strict rule. Indeed, khuṣ- up there can appear in the atelic, where it has the mean "to try to finish" and through a metaphorical extension "to fail":

khoṣuka: I fail
khoṣumi: you fail
khoṣuṭu: he/she/it fails

And niṣ- can also be in the telic, which means "to look at (for the purpose of doing something").

Sometimes the distinction is more one of purpose. jov- "to throw" in the atelic javoka means "to throw something (away, aside, etc.)" but in the telic jevyoka it means "to throw something (for the purpose of hitting something).

Verbs which are more states like "hate", "love", "adore", etc take the habitual aspect, marked with reduplication of the root's initial consonant. For instance the root kī- "to love" forms the 1SG kikīyū "I love". Other verbs can take the habitual aspect (in which case the meaning is usually a little more unpredictable, but they also lower the root vowel in that case (nineṣū "I look at", jojavū "I pelt (something)/chuck at (something)....)

At times a verb may derive a whole telic/atelic verb from it by lowering the root vowel. For instance, khiṣ- up there can also just lower the root to kheṣ, to mean "to fail" which is then conjugated like a normal atelic verb (khaṣeka, khaṣemi, khaṣeṭu...)

All of these may take the perfect, a past tense, by prefixing the root vowel:

iniṣū: I saw
ukhiṣyoka: I finished
ukhoṣuka: I failed (or ekheṣuka)
ojavoka: I threw (something)
ikikīyū: I loved

There is no future marking as of now. Plurals are there too, with their own special endings. I'm thinking of a few more aspects than this, but not that many. I might have just these three and split the verbs up among them.

[DTheZombie]

I've been mulling around the idea of having a grammatical person contrast of 1st(ego) and 2nd/3rd(alter) for a minimalist conlang, an idea I got from Damin. I figure that alter would be used to refer to a non-first person argument that you already know the context of and maybe a construction of "alter-here/close" for 2nd person and "alt-there/far" for 3rd person when disambiguation is needed and/or in ambiguous writing. I'm fairly certain I've never run across a description of a natlang with that kind of person conflation, but I suppose its possible. Anyone know of any natlang analogues?

[Hallow XIII]

No but if your language has obviation marking it certainly seems possible, at least as a polite register thing.

[Yaali Annar]

I forgot what's the phenomenon called. But you can use intensifier to disambiguate. For example compare this two sentence:
- He hugged his puppies
- He hugged his own puppies.

The first sentence implies the pups belong to other person that is not the subject.

[Imralu]

I forgot what's the phenomenon called. But you can use intensifier to disambiguate. For example compare this two sentence:
- He hugged his puppies
- He hugged his own puppies.

The first sentence implies the pups belong to other person that is not the subject.

I wouldn't say it really implies that. It's certainly open to that interpretation, but when I read "He hugged his puppies", my first interpretation was that they were his own puppies. Stressing the his (and pointing at another guy) will indicate that someone else is meant, but with an unstressed his and without any other context, I would take it as meaning the same as his own.

[Thry]

I forgot what's the phenomenon called. But you can use intensifier to disambiguate. For example compare this two sentence:
- He hugged his puppies
- He hugged his own puppies.

The first sentence implies the pups belong to other person that is not the subject.

(Non-)reflexive?

I have come up with an idea for a synthetic passive voice coming about in a (weird) mostly-West Romance language spoken by a secret community in Greece, do you guys think it is plausible?

1) First, you had the archaic reflexes of Latin passive voice retained only in the most formal register of certain sayings and idioms. Take the verb άμω, <ámo>, to love:

Code: Select all

άμω     άμο(ρ)     <ámo     ámo(r)>
άμας    αμάρες     <ámas    amáres>
άμα     αμάτο(ρ)   <áma     amáto(r)>
αμάμες  αμάμο(ρ)   <amámes  amámo(r)>
αμάσες  αμάμην     <amáses  amámin>
άμαν    άμαντο(ρ)  <áman    ámanto(r)> 

These are discarded. The equivalent of "She is loved" is no longer Ίλλα αμάτο <Ílla amáto> but instead one of the following:
a) A synthetic form with "be" + the participle, which survives today as:
Ίλλα ης αμάρα. <Ílla is amára>
she be.3S love.PART
b) A pseudo-reflexive middle-voice form (cf. Sp "Se cierran las puertas", the doors are closed [in general]):
Άμασε ίλλα. <Ámase ílla>
love.3S-REFL she
The woman is loved (in general), The woman loves herself.

2) Let's ignore a) and analyze b): the trigger of the new passive comes here, as the reflexive pronoun from CL sē splits into two allomorphs first conditioned by emphatic stress: ση <si> and σε <se>. The first retains the reflexive meaning, while the second shifts from the pseudo-reflexive middle-voice to a true passive form. Thus:

Α δόμνα άμαση. <A dómna ámasi>
ART woman love.3S-REFL
The woman loves herself.
Α δόμνα άμασε. <A dómna ámase>
ART woman love.3S.PAS
The woman is loved (by whomever).

This distinction settles the woman as the semantically passive subject of the verb.

3) Before the distinction was settled, however, two competing syntactical structures were present, though, namely:

a) Semantical patient being the object, impersonal verb (cf. Spanish "se cierra las puertas > se las cierra" IMP* they.OBJ close.3S). *impersonal particle [the same as the reflexive particle].
b) Semantical patient being the subject, middle-voice verb (cf. Spanish "se cierran las puertas").

The splitting of the reflexive pronoun allowed for b) to make a genuine passive form out of the verb, with the semantical patient as the subject. Nevertheless, at the same time, structures with a) arose for the other pronouns:

Άμασεμη. <Ámasemi>
love.3S-IMP-me.OBJ
I am loved (in general).

Analogy with the other pronoun and a series of sound changes involving emphasis quickly leveled this form to άμασμε, άμαμμε and finally άμαμε, obscuring its relationship with the form μη <mi>. Other pronouns took other forms.

4) These new forms got their syntax reinterpreted in analogy with the third person forms, resulting in new synthetic passive forms for the first and second persons. This means that the patient could now be introduced as a grammatical subject. In addition, <ai> being homophone with <e>, the passive <-e> ending became spelled <-ai> out of orthographical influence from the Modern Greek passive forms.

Συ άμασθαι α τοία γήντε. <Sy ámasthai a toía gínte.>
you.NOM love.2S.PAS to all-ART people
You are loved by everyone.

What do you think?

Now that I think about it, I don't really know how real passives come about in natlangs.

[----]

One thing I want to try sometimes would be a small (2-4 classes), universal, nominally semantically rooted set of noun classes, but one that's neither gender nor animacy-based.
- Black/white/red is what I've been considering primarily. I imagine the last would be generalized to include e.g. most red-blooded animals, unless conspicuously colored (crows, swans).
- Flat/round/oblong/(rectangular?) seems to be a common division among noun classifiers. Might be difficult to extend to abstracts though.
- A highly dualistic yin/yang approach might be interesting but I'd need to look more into Daoism before attempting a riff-off of that.

Another interesting idea along these lines would be perhaps to have the classes based on rigidity. You could have hard solids, squishy solids, and then things that are melty or not solid at all. Navajo has a wide variety of verb bases for different shaped objects; a notable distinction is between slender, flexible objects; and slender, stiff objects. If you want more ideas you can check out the full list here on wikipedia: http://en.wikipedia.org/wiki/Navajo_lan ... tory_verbs

I read about a language, in Africa if I remember correctly, that assigned gender to non-human objects by whether they were long and thin (male), or roundish (female). There's a ton of different ways you could go with a noun class system but I do think your color-based system could be very interesting!

[Imralu]

Now that I think about it, I don't really know how real passives come about in natlangs.

Modern Scandinavian languages have a verbal suffix -s (in Swedish anyway, I think it might be -st/-sk in some dialects of Norwegian ... or maybe in Icelandic, I forget). It's equivalent to a kind of middle-voice/reflexive/passive thing. Sometimes verbs just have it, like Swedish hoppas 'to hope'. It's descended from the reflexive pronoun sig (or whatever its form was earlier on).

• Svenska talas i Sverige.
  Swedish speak.PRES.PASS in Sweden.
  Swedish is spoken in Sweden.

It's invariant for person and is distinct from the reflexive pronoun sig.

[Thry]

Okay so a grammaticalized stage of the pseudo-reflexive. Portuguese could have that in the future:

Fala-se sueco na Suécia.
speak.3S-REFL swedish in-ART Sweden

[Imralu]

Okay so a grammaticalized stage of the pseudo-reflexive. Portuguese could have that in the future:

Fala-se sueco na Suécia.
speak.3S-REFL swedish in-ART Sweden

As large parts of Brazilian Portuguese are losing the tu pronouns and verb forms (in favour of você and variants + third person verb forms), this kind of leaves it open for it to become simply falas. Of course, this is limited only to the third-person forms, but it could be like Finnish, where the 'passive' form essentially kind of means 'one speaks' (ignoring the fact that in informal Finnish it's also used for 'we') rather than Swedish, where it can be used with every grammatical person.

[Thry]

In colloquial registers of European Portuguese, it's already [ˈfa.lɐs], especially because tu falas is [tu ˈfa.lɐʃ].

[qiihoskeh]

Jan03 is oriented toward changes of state and the entities that undergo the change. Because of this, resultatives usually are the main predicate, with the action causing the result being secondary.

There are 2 main classes with regard to argument structure: univalent, whose argument is an unmarked patient, and bivalent, with an unmarked patient and an unmarked copatient immediately after the verb.

For most verbs where an agent affects a patient, the verb is univalent and the agent is an oblique argument taking the ergative-instrumental case. The main exceptions to this are verbs derived from body part words, which make the agent or perceiver (and possessor of the body part) the copatient. There are also verbs, such as "kick", where it's the agent that undergoes the change of state.

Spatial relations and a few other verbs (such as "own" and "know something") are also bivalent, with the location etc. being the copatient.

Verbs also distinguish originating states, current states or actions, and resulting states. For the last, completed result, attempted result, and failed result are distinguished (a partial result is treated as if a current action).

[clawgrip]

I was thinking about making one of my conlangs get rid of prepositional phrases completely. What I was thinking of having take over the function of preposition is a series of locational verbs and relative clauses.

Thoughts? Questions?

I've used variations of that idea for several conlangs. What I always have to think about is how to handle things like "into" and "away from" and "down from on top of".

My language Himmaswa does this. "into" is expressed with the verb "enter" e.g. "boy run enter house" and "down from on top of X" is "descend leave top belong.to X", e.g. "bird fly descend leave top belong.to tree"

[qiihoskeh]

I don't remember seeing this particular variation before.

In Jan13, a clause may have more than one predicate, for example, an action predicate and a result predicate. The tense (past, present, or future) may be marked on any predicate, the choice determining the aspect. If marked on an action predicate, the aspect would be durative, while if marked on a result predicate, the aspect would be perfect. For prospective aspect there would have to be an originating predicate. In some clauses, the action predicate might be just an ergative case marker, which could take tense marking. An ablative case marker might be used for an originating predicate.

table on-Prs book boy-Erg.
"The boy has put the book on the table."

table on book boy-Erg-Prs.
"The boy is putting the book on the table."

house at girl walk market-Abl-Pst.
"The girl was going to walk home from the market."