Proto-Søkkli scratchpad

[Travis B.]

.

[Travis B.]

.

[Travis B.]

Hmm... how much precedent is there for just about double-marking everything?

[cromulant]

There is plenty precedent for double marking both the clause (case marking on nouns + polypersonal agreement on verb) and possessive NPs (both the possessor and the possessum). Burushashki is one example. What else is included in "everything?"

[Travis B.]

Double-marking relationships between nouns which are not possession.

[Travis B.]

.

[Linguist Wannabe]

Isn't it highly typologically unusual for a language to have phonemic voicing in fricatives but not in stops?

[GBR]

By the looks of things, the voiced fricatives are lenited stops. I don't know if that makes it any more likely. Perhaps they could be on a continuum /b~β/ e.t.c?

[Nortaneous]

Isn't it highly typologically unusual for a language to have phonemic voicing in fricatives but not in stops?

http://wals.info/chapter/4

tl;dr: yes, but it happens

[Travis B.]

By the looks of things, the voiced fricatives are lenited stops. I don't know if that makes it any more likely. Perhaps they could be on a continuum /b~β/ e.t.c?

The voiced fricatives are supposed to go together with the stops, forming the series /β p pː/, /ð t tː/, and /ɣ k kː/; this is also why the voiced fricatives cannot be long (but the voiceless fricatives can), and why there is no voiced alveolopalatal fricative.

(This was in part inspired by Spanish, Greek, and older Finnish (which used to have a similar system before its voiced fricatives underwent sound change to other things).)

[Travis B.]

What I am now wondering about is how to, with regard to split-ergativity, map ergative and absolutive roles in ergative-absolutive alignment and nominative and accusative roles in nominative-accusative alignment onto less than four cases (and less than four markings on the verb). The unification of the two that I am thinking of is, as syntactic pivot, uniting the absolutive role in ergative-absolutive alignment with the nominative role in nominative-accusative alignment. This leaves the ergative role and the accusative role separate, resulting in a absolutive-nominative case, an ergative case, and an accusative case.

[Travis B.]

.

[Travis B.]

From how it seems, I am either going to have to create a lot of paradigms, or go back to my previous idea at least partially and have paradigms shared between multiple markers (e.g. have onsets specific to particular markers but have vowels and codas that are shared).

[Linguist Wannabe]

By the looks of things, the voiced fricatives are lenited stops. I don't know if that makes it any more likely. Perhaps they could be on a continuum /b~β/ e.t.c?

The voiced fricatives are supposed to go together with the stops, forming the series /β p pː/, /ð t tː/, and /ɣ k kː/; this is also why the voiced fricatives cannot be long (but the voiceless fricatives can), and why there is no voiced alveolopalatal fricative.

(This was in part inspired by Spanish, Greek, and older Finnish (which used to have a similar system before its voiced fricatives underwent sound change to other things).)

Maybe this is slightly pedantic, but doesn't Spanish have voiced stops word-initially and after a nasal, which become approximants when between vowels?

[Travis B.]

Maybe this is slightly pedantic, but doesn't Spanish have voiced stops word-initially and after a nasal, which become approximants when between vowels?

I am pretty sure that is utterance-initially and after a nasal, and I am not sure I would analyze the underlying form as being a stop as opposed to undergoing fortition in those circumstances.

[Travis B.]

.

[Travis B.]

.

[gach]

Do you have any diachronics for the consonant alternations? There are quite many variants within each stem form group (the [p,t,k]-groups etc.) and it doesn't seem obvious at all how for example the distinction between hpp-hpp-hpp and hpp-hpp-pp came to be.

A related question is how predictable or unpredictable is the alternation type of a given word if you know its shape in a given inflected form?

[Travis B.]

The consonant alternations originated in terms of fortition and lenition based on, generally, the medium stem, but sometimes the long stem or the short stem (hence why there are occasional short stems that contain /xtː/ <htt> or long stems that contain /ð/ <d>), and came to be more based in analogy and lexical diffusion rather than any regular phonological process (and then there are words that resisted that analogy, e.g. loans and neologisms, or which only applied it partially, e.g. having alternations like /tː/ <tt> - /t/ <t> - /t/ <t>).

As for clusters like /xtː/ <htt>, the originated from overlong consonants /tːː/ > [ʰtː] > /xtː/.

As for how predictable the overall alternation is if you know its form in a given inflected form, that would generally give you one of the different stems, for which generally you can predict what the stems longer than it would be, and sometimes the stems shorter (e.g. knowing that the medium stem has /t/ <d> would generally let you know that the short stem will regularly have /ð/ <d>, except in exceptions like mentioned above, but knowing that the medium stem has /β/ <b> would generally let you know that the short stem will regularly have either ∅ or /w/ <w>, which could only be determined more specifically by associating the stem with a particular paradigm.)

[gach]

As for how predictable the overall alternation is if you know its form in a given inflected form, that would generally give you one of the different stems, for which generally you can predict what the stems longer than it would be, and sometimes the stems shorter (e.g. knowing that the medium stem has /t/ <d> would generally let you know that the short stem will regularly have /ð/ <d>, except in exceptions like mentioned above, but knowing that the medium stem has /β/ <b> would generally let you know that the short stem will regularly have either ∅ or /w/ <w>, which could only be determined more specifically by associating the stem with a particular paradigm.)

How well are the paradigms then predictable from the shapes of the words? So that if you encounter a word that has /r:/ in the strong stem, can you guess whether it shows /r: ~ r:/, /r: ~ r/ or /r ~ r/ in the medium and weak stems or is your best bet going to be to memorise the paradigm? At least some irregularity should be unavoidable with a complex system like this and you could potentially go bonkers with different dialects showing very different patterns of levelling as well as irregular patterns caused by inter dialectal loaning.

It's nice to see different people's takes on a gradation system. I need to flesh out mine for posting here so that you guys can rip it apart.

[Travis B.]

How well are the paradigms then predictable from the shapes of the words? So that if you encounter a word that has /r:/ in the strong stem, can you guess whether it shows /r: ~ r:/, /r: ~ r/ or /r ~ r/ in the medium and weak stems or is your best bet going to be to memorise the paradigm? At least some irregularity should be unavoidable with a complex system like this and you could potentially go bonkers with different dialects showing very different patterns of levelling as well as irregular patterns caused by inter dialectal loaning.

It's nice to see different people's takes on a gradation system. I need to flesh out mine for posting here so that you guys can rip it apart.

In the case of /rː/ <rr> in the long stem, there really is no good way to predict necessarily what the medium and short stems could have, as they could form any one of /rː/ <rr> and /rː/ <rr>, /rː/ <rr> and /r/ <r>, or /r/ <r> and /r/ <r>, but one can guess the probability of which it will have, because as the majority of stems are originally based off the medium stem, the most likely patterns here are /rː/ <rr> - /rː/ <rr> - /r/ <r> and /rː/ <rr> - /r/ <r> - /r/ <r>, with the form with /rː/ <rr> in the short stem being less common.

However, for a good part one is going to have to memorize the gradation patterns, especially considering that on one hand there will be words that follow no regular pattern at all (e.g. the sound changes that formed them were only applied partially), and multiple gradation patterns coexisting for the same word (e.g. through dialectal borrowing).

[Travis B.]

Note that through the analogy process that produced consonant gradation, when the original stem became something other than the medium stem, consonants such as /β ð ɣ x/ <b d g h> gravitated towards the short stem, consonants such as /pː tː kː tːs tːɕ sː ɕː mː nː ɲː rː lː ʎː/ <pp tt kk cc ćć ss śś mm nn ńń rr ll ĺĺ> gravitated weakly towards the long stem, and consonants such as /xpː xtː xkː xtːs xtːɕ βmː ðnː ɣɲː ðlː ɣʎː/ <hpp htt hkk hcc hćć bmm dnn gńń dll gĺĺ> gravitated strongly towards the long stem. Hence it is relatively rare that consonants of the first set would end up in long stems or consonants of the second set and especially the third set would end up in short stems.

[Travis B.]

.

[Vardelm]

I'd like to see some sample nouns declined for each case. It would be a little easier to understand the gradations by looking at them in action.

Also, huzzah for Uralic/Samoyedic inspired conlangs!

[Travis B.]

Now that I made something Uralic-inspired I now feel obligated to add a fuckload of locative cases...