Where are the analytic and nonconcatenative conlangs?

[Vardelm]

I'm not sure how helpful the concept of "nonconcatenative language" is in the first place. We can identify items of nonconcatenative morphology, but I don't think any real language is purely nonconcatenative... and a truly vast number of languages, surely the majority, have at least some nonconcatenative morphology (between ablaut, harmony, gradation, irregular sandhi, suppletion, etc...). So there isn't really a binary division to be made here.

Isn't the same true for ergativity? No language is purely ergative, and there are different elements of ergativity. And yet, we still talk of "ergative languages" in a binary manner. Is that useful? Probably, since it indicates there are at least some significant features of ergativity, but it also tends confuse people because they think ergative languages all work in a unified manner.

[Salmoneus]

I'm not sure how helpful the concept of "nonconcatenative language" is in the first place. We can identify items of nonconcatenative morphology, but I don't think any real language is purely nonconcatenative... and a truly vast number of languages, surely the majority, have at least some nonconcatenative morphology (between ablaut, harmony, gradation, irregular sandhi, suppletion, etc...). So there isn't really a binary division to be made here.

Isn't the same true for ergativity? No language is purely ergative, and there are different elements of ergativity. And yet, we still talk of "ergative languages" in a binary manner. Is that useful? Probably, since it indicates there are at least some significant features of ergativity, but it also tends confuse people because they think ergative languages all work in a unified manner.

I don't think it is useful, no, unless we're actually talking about an (almost) completely syntactically as well as morphologically ergative language. Otherwise I would say something like "highly ergative language" or "language with many ergative constructions".

[Neon Fox]

And one day, I would like to develop Tolkein's Dwarvish to give it enough words and grammar to say anything. Dwarvish works like a Semitic language. But we only have like 20 words...

Actually, it's ~30 names, place names, or short sentences, and ~50-55 words or word forms, depending on how you count. But still, not much.

You might want to check out Neo Khuzdul (she says, a month and a half later...)

[Pabappa]

Ive had conlang fatigue lately, but I did come with a new idea recently for something that could be developed into a nonconcatenative language that is also unlike Semitic: indeclinable suffixes fusing to the nouns, pushing inflections inward and then developing paradigms that spread to new forms.

For example, there was a classifier yog denoting all handheld objects. In most languages of the family, the classifiers took case markings, but in Highland Pabappa they did not. proto-Highland Pabappa meiḳ yog "soap" ---> meiso, but meḳis yog "of the soap" ---> mekìso. This appears to be a simple infix of -k-, and since this same pattern would occur for all other nouns originally ending in -k, it could be generalized. Furthermore, since the reflexes of /kj/ and /sj/ merge, the /k/ could spread to nouns that originally ended in -s. (Or, perhaps an /s/ could spread to the -k nouns instead, but -k was more common.)

I dont plan to take the language beyond this step, but from here someone could develop an array of more complicated paradigms. On the other hand, for this to work, it helps to have a language where even the roots change form for cases, such as the previously existing meiḳ~meḳis alternation.

[Nortaneous]

You can do a lot of interesting things with analytic syntax, though, like having coverbs and relational nouns rather than adpositions, or playing with word order, or not really using pronouns.

Hathic is isolating, and adpositions are generally absent:

suka da mwu c!hagh
1S go get_closer market
I go to the market

cihw n!agh dde nl!ay lea t!any n!agh cwi
bird CLF.REL be_at be_over SBJZ tree CLF.REL chirp
The bird above the tree is chirping

cihw cwi dde lea nl!ay t!any
bird chirp be_at be_over tree
The bird chirps above the tree

Vengic is less isolating than Hathic, and generally uses relational nouns instead of adpositions:

hë ca brö es të vʼo tog
COLL JUSS walled_city POSS inside 1-make.MAJ tower
Let us build a tower in the city (lit. "...[at] the inside of the city")

But some Vengic languages have enough contact with Hathic and Amqolic to pick up verbs of location. For example, Hlu:

yae xwi Ggejjaehldduela
house EGO.IPFV Gejaehl-be_in-3
My house is on Gejaehl

dding li Gejjaehlddue li wuya nyoemha
person REL Gejaehl-be_in REL beer drink-COLL-3
The people on Gejaehl drink beer

Cf. Amqoli:

bguli qaKechashex buzhashi tselya
man GER-Gejaehl-be_in beer-PAR 2/3T.T-drink_cold_stuff
The people on Gejaehl drink beer

There's also a dative case, but only in Gejaehl:

yahlyah i xi ddoe hnuexhmun ji choenila / roahlyoah xi ddoe hnuex ji dduela
nothing POSS man for heaven=DAT meet_obligation NEG-COP.HABIT-3 / nothing man for heaven meet_obligation be_at-3
Man has nothing with which to recompense Heaven

[k1234567890y]

several of my conlangs, including Lonmai Luna, Nevotak, Classical Uraki and Ame, are isolating morphologically...

[Pedant]

The Wenglanese languages (and the Hūam languages on which they are based) are isolating, while Salvian (and its earlier form Sóßtèlaña), the Peninsular languages, and the Rocklander and Wačtik languages are non-concatenative. Actually, Old Hūam Pōu (and Archaic Wenglanese like it) are somewhat non-concatenative on top of being isolating; internal fixtures (metathesized consonants, ablaut, and umlaut) are used in derivational morphology (and, to a certain extent, in verb conjugation--although this dropped by the time the daughter languages really took off).