Re: Vardelm's Scratchpad (NP: Devani OVS order)

[Vardelm]

So..... maybe probably scrap that whole last post?

I want to use body part based adpositions for Yokai, and writing the above posts made me realize I need to do a bit more planning on how those will work & fit in with the system above, or something similar to it.


The main decision point right now: do I want to use a locative case in combination with body parts, or just have them use the associative/genitive/attributive case?


For the below examples, I'll use my "adjectival" forms from the concord tables above as the locative form. This is something I'm actually considering if I do use a locative case.


With a locative case:

shengromai madubiure jenangshiol chiunoni aidaijina zhunkhaine shengro-ma-i castle-C14-ASC madu-biu-re belly-C11-LOC jenang-shi-ol queen-C02-POS chiun-on-i loyal-C05-ASC aidaijina guards.PLR.C05.DIR zhunkha-ine feast-VRB The queen's loyal guards in the castle feasted.

Using the associative case (genitive or general attributive):

shengromai madubi jenangshiol chiunoni aidaijina zhunkhaine shengro-ma-i castle-C14-ASC madu-b-i belly-C11-ASC jenang-shi-ol queen-C02-POS chiun-on-i loyal-C05-ASC aidaijina guards.PLR.C05.DIR zhunkha-ine feast-VRB The queen's loyal guards in the castle feasted.

The locative version is pretty straight-forward since it's just "at belly", meaning "in". The associative version works more like Zapotec, although there I'm not sure there's any kind of case used. Here, the noun class also plays a part in that - for a locative meaning - "belly" does not concord with the head noun "guard". Attributive uses DO show concordance, and in this case will also probably move their position.

Attributive for comparison:

jenangshiol chiunoni shengroni madubi aidaijina zhunkhaine jenang-shi-ol queen-C02-POS chiun-on-i loyal-C05-ASC shengro-n-i castle-C05-ASC madu-on-i belly-C05-ASC aidaijina guards.PLR.C05.DIR zhunkha-ine feast-VRB The queen's loyal guards in the castle feasted.

Yeah, that a SUPER arbitrary example. I guess the guards protect ... the castle's belly? Or maybe it's just that they guard parts that are the castle's interior. However, the example shows the point about the marking at least.

[keenir]

jenangshiol chiunoni shengroni maduoni aidaijina zhunkhaine

jenang-shi-ol

queen-C02-POS

chiun-on-i

loyal-C05-ASC

shengro-n-i

castle-C05-ASC

madu-on-i

belly-C05-ASC

aidaijina

guards.PLR.C05.DIR

zhunkha-ine

feast-VRB

The queen's loyal castle-belly guards in the castle feasted.

Yeah, that a SUPER arbitrary example. I guess the guards protect ... the castle's belly? Or maybe it's just that they guard parts that are the castle's interior. However, the example shows the point about the marking at least.

*nods*

Would the "belly" portion move to another part of the statement, if, say, we wanted to focus on the location of the feast (inside the castle), rather than where the guards were posted prior to the feast? (maybe "in the castle-belly" instead?)

thank you. keep up the great work.

[Vardelm]

Would the "belly" portion move to another part of the statement, if, say, we wanted to focus on the location of the feast (inside the castle), rather than where the guards were posted prior to the feast? (maybe "in the castle-belly" instead?)

So far, I have no idea! I have a bit about focus in the other languages here (notably Dwarven), so I'll probably get to that at some point. Still hammering out basic stuff.

thank you. keep up the great work.

Thank YOU, and I'll see what I can do!

[Vardelm]

And the winner is...... LOCATIVE CASE!

The syntax feels just a little easier (for me) to manage if I have a locative case vs. parsing the context to make sure it works. Some of the languages I'm using as inspiration also have locative cases, too. Finally, I like having another form that will be regularly used that has a different suffix ending. That breaks things up and is a bit more aesthetically pleasing to me. To that end I'm going ahead & using the adjectival form.

I think I'll be using "associative" instead of "genitive" as well. Thanks, Bradrn!

So, this example is what I would currently consider "correct":

shengromai madubiure jenangshiol chiunoni aidaijina zhunkhaine shengro-ma-i castle-C14-ASC madu-biu-re belly-C11-LOC jenang-shi-ol queen-C02-POS chiun-on-i loyal-C05-ASC aidaijina guards.PLR.C05.DIR zhunkha-ine feast-VRB The queen's loyal guards in the castle feasted.

Also, the associative will be used for the participle instead of the prior adjectival/similative/semblative. To return to & update these examples:

tonglaosha dzulasha. tonglao-sha tiger-C04 dzula-sha slept-C04 The tiger slept.

tonglaosha dzulashai indusha. tonglao-sha tiger-C04 dzula-sha-i sleeping-C04-ASC indu-sha is-C04.VRB The tiger is sleeping.

dzulashai tonglaosha... dzula-sha-i sleeping-C04-ASC tonglao-sha tiger-C04 The sleeping tiger...

[keenir]

And the winner is...... LOCATIVE CASE!

excellent!

So, this example is what I would currently consider "correct":

shengromai madubiure jenangshiol chiunoni aidaijina zhunkhaine shengro-ma-i castle-C14-ASC madu-biu-re belly-C11-LOC jenang-shi-ol queen-C02-POS chiun-on-i loyal-C05-ASC aidaijina guards.PLR.C05.DIR zhunkha-ine feast-VRB The queen's loyal guards in the castle feasted.

splendid.

great work on this and the other demonstrations of this conlang

ps: I assume that "C14 and such, are the cases from the previous page...or am i incorrect? thank you

[Vardelm]

ps: I assume that "C14 and such, are the cases from the previous page...or am i incorrect? thank you

Yep. I'm using the convention C## since I haven't seen much in the way of noun classes in glossing, and I tend to like sticking with 3-character labels as to me it makes the gloss look more organized. C14 would thus be "noun class #14 (large objects, structures)".

[Vijay]

(And if this bothers you I advise you particularly to stay away from the literature on SVCs, which is mostly an endless sequence of, ‘Here’s what an SVC is!’ ‘No, this is how you define an SVC!’ ‘Here’s six types of SVC!’ ‘No, here’s four types of SVC!’ ‘English has no SVCs!’ ‘Oh yes it does!’ ‘SVCs aren’t real, I’ll call them “multiverb constructions” instead!’. I may be scarred for life by now…)

Yeah, I've had a hint of that earlier in this thread!

All this while I was drafting my lesson on SVCs in Malayalam

[Vardelm]

1st cut on pronouns. I'll just leave this here without comment.


Yokai Personal Pronouns

			Singular	Plural
1P	Familiar (Humble)	Masculine	akia	akisan
		Feminine	ashu	ashusan
	Honorific	Masculine	akanim	wakasanim
		Feminine	ashunim	washusanim
2P	Familiar	Masculine	kai
		Feminine	shao
		Animate		xaduo
	Honorific	Masculine	kasong
		Feminine	shusong
		Animate		xadusong
3P	Proximate	Masculine	ichen
		Feminine	ulim
		Animate		ansang
		Inanimate	tshazo	tshazang
	Obviative	Masculine	ichai
		Feminine	ulai
		Animate		ansai
		Inanimate	daizo	daizang

[Vardelm]

Yokai Verb Voice

As mentioned in this post, Yokai is a direct/inverse language. The obviative pronouns in the prior post might have given that away as well. So, a little more on Yokai's verb voices.

Voice	Affix	Notes
Active	∅	default, unmarked voice intransitive uses intransitive verb endings seen in noun concords
Direct	g(a)	transitive subject is higher on hierarchy than object
Inverse	+an	transitive object is higher on hierarchy than subject
Passive	+amn(i)	intransitive typical passive voice subject is patient uses intransitive verb endings seen in noun concords

Active Voice:

tonglaosha kiosha tonglao-sha tiger-C04 kio-sha see-C04 The tiger sees/saw (something).

Direct Voice:

tonglaosha yolinsha kiogasha tonglao-sha tiger-C04 yolin-sha gazelle-C04 kio-ga-sha see-DIR-C04 The tiger saw the gazelle.

Notes:

• Verbal concord is the same as intransitive concords when both subject & object are of the same noun class (I still need to work on concords for other combinations).
• Since both actors are of the same noun class, word order determines which is subject & object.

Inverse Voice:

tonglaosha yolinsha kansha tonglao-sha tiger-C04 yolin-sha gazelle-C04 k-an-sha see-INV-C04 The gazelle saw the tiger.

Passive Voice:

yolinsha kamnisha yolin-sha gazelle-C04 k-amni-sha see-PAS-C04 The gazelle was seen (by something).

So...... what's the animacy hierarchy in Yokai? Well, like relationships, it's complicated.

Stay tuned.

[Vardelm]

FINALLY!!!

Since my last post - almost 6 months ago! - I finished up my masters project for grad school, in mid-July. That took quite a bit of mental energy. Since then, I've been trying to work on affixes for Yokai moods & person agreement with very little success/progress. That is, until this past week. It took a while, but I managed to get back into the creative headspace I had when posting in the rest of this thread. I'm pretty excited that I finally have enough of Yokai's many interacting parts in place that I can present more of a complete picture of the language.


Yokai Speech Levels

So far, I've shown 2 major features of Yokai: noun classes/concords and direct/inverse alignment. The 3rd leg of the stool plays with both of those, which is speech levels. These are multiple levels of politeness distinctions that are very similar to Korean. The levels differ through the following:

• politeness/formality affix on the verb
• simple vs honorific 1st & 2nd person pronoun use
• use vs non-use of 1st & 2nd person pronouns
• hierarchy of 1st & 2nd person pronouns (direct/inverse voices)
• verb ending particles
• noun honorifics
• observation of taboo speech

Broadly speaking, Yokai has 4 major speech levels, defined by the verb's politeness/formality affix. The affix also changes depending on the verb's mood. The table below shows the 3rd person affix for each mood.

Speech Level	Declaritive	Inferential	Subjunctive	Interrogative	Usage
Ceremonial	-ji-	-xu-	-hai-	-xol(u)-	The ceremonial level is used in highly formal settings, such as official functions performed by royalty or religious observances by priests. It includes a high level of politeness in addition to formality.
Formal	-de-	-seo-	-wa-	-siol(u)-	The formal is used where a degree of formality (surprise!) and politeness is needed, but not as much as the ceremonial. This includes interacting with elders, superiors, low/mid level officials, work colleagues in a formal setting, business transactions, & strangers.
Casual	-	-ru-	-yo-	-l(u)-	The casual appears in informal setting where the participants are well known to each other and there is no particular assertion of deference or superiority in social position.
Assertive	-kho-	-kho-	-zai-	-bi-	The assertive is used in casual or formal settings where the speaker projects social superiority over the addressee. This can be to lower social ranks or to children/students when giving them instruction, or even to lower social ranks and criminals when scolding/lecturing them.

Simple, right? But why would we want simple?

For each of those main speech levels, there are a number of sub-levels. Depending on the scholar studying the Yokai language, these might just be variations on the 4 main levels, or they might count as individual speech levels in their own right. They are defined by all the other factors besides the verb's mood/formality affix.

Speech Level	1st P Pronoun	2nd P Pronoun	Subject Pronoun	Taboos	Usage
Ceremonial
Courtly	Hon	Hon	1st	Yes	by royalty, high officials to each other with same/similar rank
Stately	Hon	Fam	1st	No	by royalty, high official making pronouncements taboos set aside for official/ceremonial functions
Reverential	- (not used)	Hon	2nd	Yes	to royalty, high officials, etc. shows extreme deference
Formal
Deferential	Fam	Hon	2nd	Yes	to low/mid officials, superiors, elders
Proper	Hon	Hon	2nd	Yes	to strangers to colleagues in formal setting to customers
Casual
Polite	Fam	Hon	1st	Taboo addressee Public	to colleagues, acquaintances
Simple	Fam	Fam	1st	Public	friends, relatives similar of age by adults to children when comforting, having fun
Intimate	Fam	Fam	1st	No	to close friends & relatives
Assertive
Instructive	Hon	Fam	1st	Yes	to lower social rank by teacher to students by adults to children when instructing
Pejorative	Hon	- (not used)	1st	Yes	to criminals to lower social ranks when scolding/berating

EDIT: adding a bit more description.

• For 1st & 2nd pronouns, Fam is "familiar" and Hon is "honorific". The term "familiar" might be backwards from languages that makes the distinction. Familiar implies "I am not claiming any status", and therefor is "humble". The 1st person honorific does "claim status".

• Subject pronoun determines whether 1st or 2nd pronoun is the subject for a verb with the direct voice in a transitive sentence.

• For Taboos, Taboo addressee means taboo speech is used only if the direct addressee is or includes someone in the taboo group, while Public means just that: only when in public.

That should provide a sketch of the speech levels. I still need to work on defining the usages more, but the basics are in. Also, I haven't even started on the verb & noun particles/honorifics, and I have just some basic thoughts on taboos so far. Some of the sub-levels may expand or disappear as I work on this, but I feel like this is a good starting point.

Next up will be affixes for concords in all speech levels, persons, and moods, which is what I've been trying to work on for the last 5+ months. Until I have all of that posted, I won't bother with examples sentences.

[Vardelm]

Yokai Verb Concords

This post will be mostly tables of morphology will little comment. I intend to go back and provide example sentences with commentary. There's so much here that I feel it's useful to just see all of the agreement suffixes together. Before getting to the affix tables, one aspect of the noun classes needs to be explained.



Noun Macro-Classes

I've mentioned the "noun macro-classes" before. These are classes that share similar plural forms in that they have common prefixes. They also share related semantic space. They are arranged according to the 5 traditional Chinese elements: water, wood, fire, earth, and metal. I describe these here just because some of the below tables refer to the macro-classes by name.


Noun Macro-Classes

Macro-Class	Included Classes
Water	01 - low-caste women, aquatic wildlife 06 - fruit, aquatic plants, oceans, lakes, rivers 11 - hollow objects, containers,flexible objects 16 - events, time 21 - liquids, food
Wood	02 - women, birds 07 - deciduous plants & forests 12 - small objects 17 - states, qualities 22 - plant matter
Fire	03 - leaders, elders, reptiles 08 - flowers, seeds, leaves, deserts, volcanoes, sky, heavens 13 - insubstantial objects, man-made, random-shaped objects 18 - ideas, concepts, emotions 23 - coals, cinders, lava, smoke, flesh/meat
Earth	04 - men, furred animals, crawling insects 09 - grasses, mushrooms, vegetables, mountains, hills, plains, caves 14 - large objects, structures 19 - actions 24 - rock, stone, sand, dirt
Metal	01 - low-caste men, shelled, flying insects 06 - pines, cactus, pinecones, nuts, glaciers, tundra, grave 11 - tools, hard objects 16 - death, darkness 21 - metal, shell, crystal, ice

---

Verb Concords, Politeness, & Mood

In addition to the below tables, intransitive 3rd person verbs agree with their subject's noun class. The endings are already listed in the Noun Classes chart.


Mood & politeness
- 3rd person
- 1st & 2nd person intransitive

		Declaritive	Inferential	Subjunctive	Interrogative
	3P	-ji-	-xu-	-hai-	-xol(u)-
Ceremonial	2P	-jian	-xuen	-hun	-xolun
	1P	-junkayo	-xunkayo	-hun	-xolun
	3P	-de-	-seo-	-wa-	-siol(u)-
Formal	2P	-dan	-san	-won	-siolun
	1P	-donyo	-siunyo	-won	-siolun
	3P	-	-ru-	-yo-	-l(u)-
Casual	2P	-ming	-rong	-yang	-long
	1P	-ng	-ng	-yang	-long
	3P	-kho-	-kho-	-zai-	-bi-
Assertive	2P	-khang	-khang	-zeng	-be
	1P	-khang	-khang	-zeng	-be

Notes:

• 3rd person mood is indicated by a separate suffix. The same mostly appears with 1st & 2nd person, but have fused to some degree.

• It would be nice to be able to merge cells for this table (rowspan & colspan). Since that's not available, you'll note that a number of cells have the same entry. Notably, 1st & 2nd person are the same in the subjunctive & interrogative.

Mood & politeness
- Transitive 1st vs 2nd person

	Declaritive	Inferential	Subjunctive	Interrogative
Ceremonial	-jinyo	-xunyo	-hun	-xolun
Formal	-denyo	-seonyo	-won	-siolun
Casual	-nyo	-runyo	-yang	-long
Casual	-khang	-khang	-zeng	-be

Notes:

• Endings are used on transitive verbs where the subject & object are 1st & 2nd person, regardless of which is which.

Transitive 3rd person vs 1st & 2nd person

3P Noun Class	vs 1st Person	vs 2nd Person
Water animate	-yon	-ano
Wood animate	-yoshe	-anshe
Fire animate	-yua	-ayua
Earth animate	-yosha	-nsha
Metal animate	-yan	-ane
Water inanimate	-yoza	-anza
Wood inanimate	-ya	-anya
Fire inanimate	-yonka	-anka
Earth inanimate	-yoja	-anja
Metal inanimate	-yuna	-ana

Notes:

• Endings are used on transitive verbs where there is a 3rd person referrent and either a 1st or 2nd person.

• Animate almost always refers to noun classes 1-5 (of theh appropriate macro-class). Inanimate is for the other classes.

Transitive 3rd Person Animate vs 3rd Person Animate or Inanimate

Macro Classes	Metal	Earth	Fire	Wood	Water
Water	-nan	-shon	-nga	-nshe	-no
Wood	-nshe	-sha	-shega	-she
Fire	-gen	-gasha	-(g)a
Earth	-shen	-sha
Metal	-(i)na

Notes:

• Endings are used on transitive verbs where at least one of the subject or object is animate.

• It's worth pointing out that when the subject & object are from the same noun class, the ending used is identical to the intransitive. The verb is still marked with the direct or indirect voice affix.

Transitive 3rd Person Inanimate vs 3rd Person Inanimate

Macro Classes	Metal	Earth	Fire	Wood	Water
Water	-nza	-jaza	-zanka	-zai	-za
Wood	-inai	-eja	-enka	-e
Fire	-nken	-nkaja	-nka
Earth	-jan	-ja
Metal	-(i)na

Notes:

• Endings are used on transitive verbs where both subject & object are inanimate.

• I considered (and am still considering) having a table with animate vs inanimate, separate from animate vs animate. There's already a lot of agreement endings, so having just 1 animate seemed like a good cutoff. There may have been such a set of endings in the past, and this would show that the system is already starting to collapse a bit. Future languages could reduce to using just the macro-classes or stop marking for both subject & object. Plenty of good options.

[Rounin Ryuuji]

What would that look like in practice?

[Vardelm]

What would that look like in practice?

I assume you mean all of this, correct?

That's basically the next few posts. I don't have a set plan yet, but I'll probably start with just showing the basics of various transitive sentences in the declarative mood. Maybe then move on to the hierarchy for direct/inverse, and then to situations that would involve different views of politeness, etc. I foresee a bunch more work on how the noun classes really work in the society, so I'd guess this will change. I don't mind saying I don't have all of it worked out yet. This IS a scratchpad, after all.

[Rounin Ryuuji]

What would that look like in practice?

I assume you mean all of this, correct?

That's basically the next few posts. I don't have a set plan yet, but I'll probably start with just showing the basics of various transitive sentences in the declarative mood. Maybe then move on to the hierarchy for direct/inverse, and then to situations that would involve different views of politeness, etc. I foresee a bunch more work on how the noun classes really work in the society, so I'd guess this will change. I don't mind saying I don't have all of it worked out yet. This IS a scratchpad, after all.

Yes, but more specifically all the verby stuff. I would've liked to see lots of examples, but I suppose that's why this is a scratchpad.

[Vardelm]

Yes, but more specifically all the verby stuff. I would've liked to see lots of examples, but I suppose that's why this is a scratchpad.

Gocha. Yes, I plan on lots of examples, same as the rest of the thread. I just needed to get these 2 posts in to show the high level view. I think without that, examples for this lang wouldn't tell the whole story. There's too many interconnecting parts.

[Rounin Ryuuji]

Yes, but more specifically all the verby stuff. I would've liked to see lots of examples, but I suppose that's why this is a scratchpad.

Gocha. Yes, I plan on lots of examples, same as the rest of the thread. I just needed to get these 2 posts in to show the high level view. I think without that, examples for this lang wouldn't tell the whole story. There's too many interconnecting parts.

That's the idea I get. I'm having some trouble piecing it together mentally (though this is probably owing to gaps in my knowledge more than anything else).

[keenir]

Welcome back.


(always good to see new Yokai material - better to see new material with the return of a conlanger)

[Vardelm]

Welcome back.

(always good to see new Yokai material - better to see new material with the return of a conlanger)

Thanks! I've been on the board, reading almost daily, & posting a little, but it's good to finally get back to posting in this thread.

[Vardelm]

Yokai Verb Examples

This post will provide a bunch of examples to show the verb inflections mentioned in the posts above.


Intransitives

I figured it's worth showing some more intransitives as a recap, and because I have 1st & 2nd person endings now (not to mention pronouns).

These examples are all in the "casual" speech level since that is the least marked (in the declaritive).

akia jihang akia 1PS jiha-ng ate-1P I ate.

akisan jihang akisan 1PP jiha-ng ate-1P We ate.

kai leungdaming kai 2PS leungda-ming fall.unconsicous-2P You passed out.

kasong leungdaming kasong 2PS.HON leungda-ming fall.unconsicous-2P You passed out.

tsumakhomdai jeotade tsuma-khom-dai pine-pod-C10 jeota-de fall-C10 The pinecones have fallen.

Notes:

• 1st & 2nd agreement verb concords are found in the last post: Verb Concords

• 3rd person subjects use verb concords based on their noun class, from this post: Concords

• Verbs do not agree with subject/object pronouns for politeness. The speech levels (politeness, formality) are based on who you are speaking to, not whether an honorific pronoun is used for subject or object.

• Verbs do not inflect for number.

Speech Levels

I won't go through details of every speech level. This will just show some of the verb forms used in the speech levels to provide an idea of how the affixes are used mechanically. These examples will still be in the declarative mood.

akanim chumsuedonyo akanim 1PS.HON chumsue-donyo refuse-1P.DEC.FRM I refused.

kasong goshengdan kasong 2PS.HON gosheng-dan succeed-2P.DEC.FRM You succeeded.

yoxanxio biadodexa yoxan-xio cloud-C08 biado-de-xa arrive-FRM-C08 The clouds arrived.

Notes:

• Each example is in the declarative mood & formal speech level.

• Intransitive 1st & 2nd person endings have formality fused with their endings

• 3rd person has the affix -de-, followed by the intransitive concord for noun class.

Moods

daijioni dzularune daiji-oni guard-C05 dzula-ru-ne sleep-INF-C05 The guard slept (I assume).

Notes:

• Moods are fused with formality/politeness, so they work identically. 1P & 2P have fused endings, while 3P has a separate affix.

• I intend for the "inferential" to be a non-direct evidential like Turkish has. Beyond that, I'm not 100% sure how it will function, and I have even less ideas on the subjunctive & interrogative.

Voices

rosangsha lujimunxun paomagadeshon rosang-sha merchant-C04 lujimunxun C06.PLR-bananas-C06 paoma-ga-de-shon sell-DIR-FRM-TR.C04.C06 The merchant sold bananas.

rosangsha zhuduoni thiekogadeshen rosang-sha merchant-C04 zhuduo-ni thief-C05 thieko-ga-de-shen strike-DIR-FRM-TR.C04.C05 The merchant struck the thief.

rosangsha zhuduoni thiekandeshen rosang-sha merchant-C04 zhuduo-ni thief-C05 thiek-an-de-shen strike-INV-FRM-TR.C04.C05 The thief struck the merchant. It was the merchant the thief struck.

jimunxun paomamnideni jimunxun C06.PLR-bananas-C06 paoma-mni-de-xun sell-PAS-FRM-C06 The bananas were sold.

Notes:

• All of the voice examples use the formal -de- marker, mostly to show where mood/polities affixes show up relative to voice affixes.

• I don't have the animacy hierarchy worked out yet. It will take typical considerations like sentience & person, but will also add relative social status, noun class, & possibly taboos. The 3rd example shows the inverse marker -an-, although I don't know if a merchant or thief would be higher (probably the merchant). Word order will play some role as well in disambiguation as well as focus, so provided the translation "it was the merchant..." to show this a bit.

• The passive is intransitive so it uses intransitive concords. I think an oblique argument for an agent will be rarely used since the inverse voice exists. If an oblique is used, it would probably indicate a lower level of agency/causation.

---

One thing I realized is I need to decide on what verb endings to use for 3rd person pronouns. I might just use the noun class the participant the pronoun would refer to, but not sure.

I think that's it for now.

[Vardelm]

It's been a while, so time to revisit Dwarven a little bit!


Dwarven: Sandhi Sketch

The plan for Dwarven (or maybe its descendant languages) is to make frequent use of consonant mutations. This will primarily be at the start of words like the Celtic languages, but there will also be mutations at or towards the end of words, (probably) more in the style of the Uralic languages. I have a start on the sound changes for the Celtic-inspired mutations, but not the Uralic ones. As such, I'm posting just to force myself to write & think about it in a different manner than I have been. Chances are high that I'll be posting modified versions of this as my thoughts develop.

For now, this is more of a sandhi system than true consonant mutation. All of the sound changes are basically predictable based on phonological environment. Later, I may switch so that the environments were present in a previous language and removed. Or, I may just let the sandhi system stand in this language and develop it into various directions in the descendants. I actually kind of like that possibility.

Lenition
Lenition will be the main driver of consonant sandhi in Dwarven. It results purely from a preceding word in a phrase ending in a vowel. The resulting lenition is primarily adds voicing, but can also yield spirantization, debuccalization, or elision.

Normal	Lenited
p	b
b	v
pj	bj
bj	vj
t	d
d	ð
tj	dj
dj	ðj
k	g
g	ch
kj	kj
gj	chj
q	h
qj	j
f	v
v	∅ (elided)
fj	vj
vj	j
ð	l
s	h
ðj	lj
sh	h
ch	r
chj	j
h	∅ (elided)
m	v
mj	vj
n	ð
nj	ðj
ñ	ch
ñj	chj
l	l (no change)
lj	lj (no change)
j	j (no change)
r	r (no change)

Devoicing Assimilation
Where a preceding word ends in an unvoiced consonant with the same manner & placement, the initial consonant of the following word will also be devoiced.

Normal	Devoiced
-p b-	-p p-
-t d-	-t t-
-k g-	-k k-
-f v-	-f f-

Future Devoicing Assimilation
At least one branch of Dwarven will go on to have even more devoicing. Essentially, a preceding word ending in any unvoiced consonant will devoice the initial consonant of the following.

Normal	Devoiced
-p b-	-p p-
-p d-	-p t-
-p g-	-p k-
-t b-	-t p-
-t d-	-t t-
etc	etc

Future Devoicing & Hardening
With the devoicing in place, the same environment (preceding unvoiced consonant) will lead to hardening of fricatives, nasals, & continuants. The devoicing & hardening will form 1 set of mutations that are - vaguely - the "opposite" of lenition.

1 issue is with <s sh>. I'd like to have them harden to something, but not sure what. With that said, when I work out other sound changes happening to the descendant langs, there may or probably will be affricates <ts dz> by that point, so it could end up being a non-issue.

Normal	Devoiced/ Hardened
p	p (no change)
b	p
pj	pj (no change)
bj	pj
t	t (no change)
d	t
tj	tj (no change)
dj	tj
k	k (no change)
g	k
kj	kj (no change)
gj	kj
q	q (no change)
qj	qj (no change)
f	p
v	f
fj	pj
vj	bj
ð	t
s	???
ðj	tj
sh	???
ch	k
chj	kj
h	q
m	bj
mj	bj
n	d
nj	dj
ñ	g
ñj	gj
l	ð
lj	ðj
j	ch
r	h

Consonant Mutations / Grades
With both lenition and devoicing/hardening in place, that should create a situation where there are some bi-level mutations and some tri-level. The plan is - for at least one language - to remove as much of the other inflections and phonological environments such that the consonant mutations are the only thing that indicate the inflection in question. At this point, it's a consonant gradation system.

Normal	Lenited	Devoiced/ Hardened
p	b	p (no change)
b	v	p
pj	bj	pj (no change)
bj	vj	pj
t	d	t (no change)
d	ð	t
tj	dj	tj (no change)
dj	ðj	tj
k	g	k (no change)
g	ch	k
kj	kj	kj (no change)
gj	chj	kj
q	h	q (no change)
qj	j	qj (no change)
f	v	p
v	∅ (elided)	f
fj	vj	pj
vj	j	bj
ð	l	t
s	h	???
ðj	lj	tj
sh	h	???
ch	r	k
chj	j	kj
h	∅ (elided)	q
m	v	bj
mj	vj	bj
n	ð	d
nj	ðj	dj
ñ	ch	g
ñj	chj	gj
l	l (no change)	ð
lj	lj (no change)	ðj
j	j (no change)	ch
r	r (no change)	h

If you have been around the ZBB since the 2008-2010 timeframe, you may see where this is headed. Eventually, I want to have an improved and diachronically created version of my Tibetan Dwarvish nooblang, which was never finished.

I may create some example sentences, but probably not tonight.