Shameless Self Promotion

I figured I could just as well promote my own music a bit.

Both pieces are in 11-tone equal temperament - a scale most of my readers probably never have heard anything in, since almost all western music is either in 12-tone equal temperament or in some "adaptive just intonation" of some vague sort. This adds quite a number of complications for the composer to grapple with, but I think I've rather managed to overcome some of these challenges with these two pieces – the first two instalments of a series of songs with somewhat similar ideas underlying them.

For the record, microtonal theorists kept saying that 11-tone equal temperament is very useless except for atonal music until at the very least a fair bit into the previous decade.

Detail #165: Mergers in Person * Gender

Imagine a language in which all persons are marked for gender. Further, there are at least three social genders, which I will call i, ii and iii.

I.i, I.ii, I.iii
you.i, you.ii, you.iii
3sg.i, 3sg.ii, 3sg.iii

Let's now imagine that there's some conflations, but that these conflations are not within one person, i.e. it's not the case that 2nd person conflates genders ii and iii or somesuch - it is the persons that are conflated instead:

1/2 sg.i	1 sg.ii	1/2/3 sg.iii
	2/3 sg.ii
3 sg.i

This means if you're gender iii, you use the same pronoun you'd use for a second or even third person of your gender. This would probably either be a gender considered very expendable or interchangeable, or one where the set of members is so small that it's no problem. For the other genders, the situation obtained by this complication might be less obvious, but could imaginably have interesting cultural reasons behind them.

Barxaw: Lexical Aspect and Verbs of Perception

In Barxáw, many verbs have an aspectual component to their meaning, for instance

xéλ - say, utter (perfective or rather maybe punctual)
najɛ̀  - talk, speak, orate, (imperfective, non-punctual) 

vár - to discover someone by hearing, to perk up from hearing something, to hear a very short sound
vìk - hear, listen

bìris - to decide in favour of something
t^harí - to think of something with an approving mindset

qursé - to be pricked by something, to experience a short flick of pain (that is generally not from heat)
sáré - to experience a pain that does not immediately pass (and is generally not from heat)

Now, with subordinated verbs where the main verb is one of perception, there's generally agreement between the lexical aspects of the main verb and the subordinate one, e.g.

nín qursé sa vár úŋ  - It stung, hearing you (enter or some other punctual thing like that)
*nin qursé sa vìk úŋ

nín sáré sa vìk úŋ - it pains/pained me to hear you (talk or sing or whatever)

 te úŋ vìk sa nín tiλì? - did you hear me sing?
*te ún vár sa nín tilì?

Situations where a non-punctual stimulus is perceived for just a punctual time-span requires some periphrasis:

nín tiλì, úŋ vár pex.
I sang, you discovered it/perked up from hearing it.

nín t^harí sa úŋ tiλì
I think with approval of your singing, "I like your singing"

Other new vocabulary in this post:

nín: I
úŋ: you
pex: a pronoun that refers to previous clauses. Also can mean something along the lines of "thus" or "such".
sa: a subordinating particle
tiλì: sing

Ćwarmin: Adjectival Attribute Syntax

Ćwarmin adjectives generally precede the noun they modify. However, some complications exist.

Possession is syntactically adjective-like, and genitives can appear anywhere in the string of adjectives. With genitives, any attributes of the genitive noun has to be in the genitive as well, and are more tightly coupled, syntactically to their noun:

gara migitite wicxitite lank[...]

red old-GEN.DEF house-GEN.DEF door[...CASE]
the/a red door of the old house

Proper nouns differ in that adjectives seldom are used with them unless the adjective is part of a proper noun phrase, i.e.

salcan Murustu
great tusk-NOM.DEF (the name of a town)

nerel salcuta Murstuta varsan
leeward great-GEN.DEF tusk-GEN.DEF harbour
salcuta Murstuta nerel varsan
great-GEN.DEF tusk-GEN.DEF leeward harbour
a/the leeward harbour of the Great Tusk.

However, adjectives can be compounded with proper names, and this is the common way of using adjectives with personal names and only slightly less common with names of places. It is not unusual for such a form to become the usual designation for a person, and official records may even prefer such designations over given names.

Salcaŋire ≃ Salcan Gire, "big Gire"
Misketadu ≃ Greedy Tadu
Farnapalb ≃ Farna Palb, "old Palb"

Compounds differ in intonation from phrases in that the initial stress of the second word is significantly weakened. 

Generally speaking, all attributes are marked as singular and indefinite. All adjectives can basically stand as nouns and vice versa, but statistically an adjective is very likely to stand as an adjective and vice versa, and this essentially helps identify whether a case-marked word is the head of a phrase or not. Intonation also helps out, with noun phrases tending to have a slight pitch raise on the stressed syllable of the head noun.

The order of adjectives in general tends to follow these orders, although the different orders seem to be somewhat independent of each other: personal characteristics > size > age > colour > material on one hand, value > shape > use-related adjectives on the other. A few other orders probably also exist. The order is not as fixed as in English, however. 

Subjects can be separated from their adjectives by the verb, simply putting an arbitrary number of its adjectives before the verb, and the subject directly after it. Likewise, the object can have an adjective shunted to the end of the clause. Such movements, the adjective often has a more central role in the clause - fronted adjectives say something about the state of the subject with regards to the verb phrase, adjectives that have been moved to clause-final position say something about the state of the object with regards to the verb phrase.

Challenge: Onwards with Detail #11

I went and edited a really old post recently, because I realized I had never visualized the idea in it well enough.

I realized while looking at it, that maybe a few quirks could make it even better. Right now it's fairly rigid - two or three words are resolved to their roles, and no way of changing how they are resolved. Every adverbial and other argument basically requires its own VP, so that's a bit unwieldy, but not all that impossible ultimately. The language probably needs a bunch of subordinating conjunctions for that kind of thing, but let's not get our hands in that deep just yet.

Let's consider the illustration of the words' "distributions" along the SOV circle:

life as a graphical artist is not in my future

Now, what operations would give us full control over this? How many possible elements do we have in the first place? An element is an n-tuplet of the form [S, O, V, d, z]. S, O and V ∈ {1, 0}, d ∈ {clockwise, counterclockwise}, z ∈ {s,v,o}.

Interpretation of this tuplet works like this: if S,V or O is 0, it is not close to that point along the circumference, if S,V or O is 1, it is close to the corresponding point along the circumference. d gives the direction along the perimeter as shown, and z (for 'zero') is the starting point, the most preferred element out of the elements.

Thus there are 2⁴*3 possible configurations for the words, a total of 48 possibilities. A three-word clause can be 48³ combinations of configurations. Meanwhile, three words can map onto subject, object and verb in six ways. We have a bit of a gap of orders of magnitude here. 48³ = 110592.

We have more than one hierarchy here: for each word, there's an internal hierarchy. [1, 1, 0, clockwise, s] prefers being Subject, over Verb, over Object. However, when resolving which out of a = [1,1,0,clockwise, s] and b = [1,0,0,clockwise,s] gets to be the subject, we find that b is more likely of the two to be the object, and a ends up our subject. 

Challenge: come up with a small set of markers that operate on this system in a way that does not reduce to directly marking role, but that operate on the tuplet-level of the representation, yet provides us a full way of getting any of the six possible SVO-assignments out of any of the 48³ possible configurations.

Detail #164: A Register in a Tonal Language

In a tonal language, have a specific register with some added complications to the phonology. This register deals with the music theory of the language, and the complications are entirely due to this.

Triads as well as common motifs all are thus given fairly straight. The underlying syllables, however, inform us a bit about the melodic/harmonic structure - so, i.e. "major triad played in first inversion simultaneously" comes out as na.a.a - with the three vowels having the tones of such a chord. An arpeggio is given with an exaggerated pause: naʔ:naʔ:naʔ:. Glissandos are expressed as actual glissandos. Closer details may be given: to specify which degree of the scale it starts and which it goes to, you use the tone's name's syllable and substitute the lexical tone it should have with a pitch (if possible, corresponding to the actual scale tone if a musical context is present), and a glissando upwards.

A lot of the musical terminology is basically more regular lexemes with tonal substitutions that illustrate the musical details. E.g. glissandos in general are the (monosyllabic) words for "jump" or "fall", but with a glissando (rising on jump, falling on fall) instead of the usual tone the words should have.

Chords. motifs, intervals and progressions in general may be given in slightly shortened forms, and with underlying words like "order", "sequence", etc. Such combinations enable theorists to innovate terminology rather efficiently, by reusing the word for "progression" with different tones to obtain different progressions, or the word for chord or motif or melody with added syllables or such to encode different chords, motifs or melodies.

Of course, "sequence" has its own real tone that is used in the rest of the language. These tones are carried over and remain - the "usual" tones of the language are pronounced in a way that is slightly off from the musical intervals, in a sufficiently audible way that a person who has mastered the register can distinguish "sequence" as "sequence" from "sequence" as "I-IV-V progression".

Detail #163: Interrogative Pronouns from Verbs

Imagine verbs with the rather specific meaning of 'to ask about the location of +OBJ', 'to ask about the manner of +OBJ', etc. Now, for whatever reason, these verbs are widely used in certain registers out of some deference-strategy where it's preferable to frame your question using a certain mood if asking your superior. This is then generalized, so that these verbs are widely used.

Now, instead of

where is the parrot(.nom)?

you get

may I [ask about the location of] the parrot(.acc)?

After a while, this turns into

[ask about the location of] the parrot.acc?

in whatever participle or infinitive or even finite form that the language happened to end up on. The interrogative word is reduced significantly, and we end up with a semantically bleached interrogative, which we'll know translate as 'where':

where (is) parrot.acc?

Suddenly, we've produced a split where intransitives where one side is an interrogative have accusative ~subjects. Now, in the transitive case we can envision different things having happened: the same strategy, and having this strategy somewhat ignore the roles of previous nouns, or we could have something like this:

may I [ask about the location of] his seeing the parrot.acc?

Possibly leading to an ergative pattern whenever interrogatives are present? However, we could also have the pattern differ between transitive and intransitive verbs, so that the transitivity of the verb forces the usual case marking onto the constituents, whereas intransitive verbs just get a tad odd.

 

 

Detail #162: Explicit Pronouns as Obviative

In a language with a rich congruence system, it might seem that third person pronouns would per se be a bit superfluous. However, what if explicit third person pronouns began having an obviative role, and emphasizing a proximate referent is done by other means - reduplicating the congruence morpheme or whatever?

Detail #161: Shapes

In Newmeyer's Possible and Probable Languages, Newmeyer lists a bunch of implicational universals. They're not listed as in any way necessarily absolute or anything. Now, we're not going to discuss the validity of universals or anything, but rather see what conlanging-ideas we could derive from them. I will now discuss the

"In every language in which the property concept of shape is expressed through adjectives, then those of color and size are also expressed through adjectives (Dixon 1977; TUA #141)." (Newmeyer, p. 5)

This opens the question of course as to what other ways of expressing shape exist. There's a few obvious alternatives: adpositional/case attributes, i.e. the property is expressed somewhat comparably to these English phrases, both somewhat 'off':

a man of rotundity, a face of angles : a rotund man, an angular face

 Now, such prepositional phrases could easily be used directly as complements of copulas:

the man is of rotundity

However, it could also be used in some other way:

the face has many angles

However, let's get a bit wilder, shall we? How about shape being encoded as a set of optional derivational suffixes similar to, say, diminutives in how they're applied.

man-rotund-[number]-[case]
face-angular-[number]-[case]
stick-bent-[number]-[case]

For the predicative sense, let's derive verbs in some special ways, or use some kind of dummy noun with the suffix. All suffixes may not have verbal versions.


Bibliography:
Newmeyer, Frederick J., Possible and Probable Languages - A Generative Perspective on Linguistic Typology, Oxford University Press, 2005

Dairwueh: Personal Pronouns

I notice I have not designed the Dairwueh pronoun system yet, which was a bit of a surprise. So, time to get around to it, I guess.

Personal pronouns:

	I sg	II sg	IIIsg masc	IIIsg fem	IIIsg neut	I pl	II pl	IIIpl anim	IIIpl inanim
nom	ver	xot	ker	si	je, te	int	xin	sera	ja
acc	vena	xota	kn/kna	s/su	je, te	ina	xita	setar	ja
dat	vevit	xovit	kar	sir	in	init	xivit	serit	jivit
gen	vedin	xodin	kenat	sura	iŋa	indin	xidin	serin	jivit
loc-instr	veder	xoder	keŋa	suvat	iŋa	inder	xider	serar	(j)iŋa

 Like in your average Indo-European language, possession is not generally expressed by genitive forms of pronouns.

Possession, instead, is formed by these pronouns - which are inflected for case and number congruence with the possessed noun:

	I	II	IIImasc	IIIfem	IIIneut
singular	vev-	xov-	kev-	suv-	jo
plural	ade-	bev-	arv-	sarv-	(siŋ)

The plural forms may seem rather far removed from the personal pronouns. They are in fact closer related to personal past tense suffixes, which are closely related to former ergative forms that have been lost. For the plural nouns, the ergative forms had a separate oblique stem. The neuter forms are not declined for congruence with their heads.

As for the genitive, there is one peculiarity of some interest. Normally in Dairwueh, a definite subject of a transitive verb is in the genitive, whereas all other subjects are nominative. Pronouns are obviously always definite; however, a split has occurred in the third person.

Both "ker/si" and as "kenat/sura" can stand as subjects of transitive as well as intransitive verbs. The difference is that kenat/sura refers to a more prominent referent, and ker/si to a less prominent one. However, in all other positions, they generally refer to the more prominent referent in general. A dropped subject is generally interpreted as referring to the more prominent referent as well.

With possession, kenat/sura are parsed as reflexive possession, i.e. as referring back to the subject. Kev-/suv- are parsed as referring to some non-subject possessor.

Dairwueh: Prepositions and the Preprepositional

Dairwueh has ten very commonly used prepositions, and about a handful of less common ones. They are:

ma		at, in, on
i		to, at
di(n)		as, than, like, of
lo		from, of
əre		for, by
gir		along, through
priv		under, inside
aub		with, for, by, in response to
ves		around, across, along, near to, before
nist		against, in opposition to, in response to, after
yil		about, like, made from, consisting of, during

I've previously given a very short sketch of the Dairwueh Preprepositional construction. This particular detail of the language requires some more elucidation, as it is quite a common structure.

Structure: a noun phrase in the genitive precedes a prepositional phrase.

[[NP₁] [PP [NP₂]]]

...	vərxi	-ra	ma	buner	-ŋa
...	coat	-FEM.SG.GEN	on	shoulder	-MASC.SG.GEN
with a	coat		on	(his/her) shoulder

Prepositions in this construction differ slightly from regular prepositions - whereas in other positions, prepositions come in one form, and parsing for whether it codes for location or direction is based on the semantics of the verb rather than the case of the noun or any inflection on the preposition. However, the preprepositional construction has two forms for most prepositions.

A few prepositions can appear in intransitive forms in the preprepositional construction; these then have a special form that appears for this. Generally, the preposition then ends in -s.

vərxira mas – "with a coat on"
kamrin dis – "(more) mushrooms than (you'd believe)"
ustira yis – "in the meantime, there was peace"
kavat nis – "and thus (appeared) the squirrel"
sparagedin pris – "and so, the farmers have been dealt with (basically 'buried' or covered)"

There's also a dynamic form, -(ə)bi. (Since the number of prepositions is fairly limited, we can just as well give the full set: mabi, ibi, dəbi, lobi, ərəbi, girvi, prəbi, avəbi, vesvi, nisvi, yilvi.) Since the verb is less closely tied to the preprepositional construction than to prepositional phrases in general, the semantics of the verb phrase do not inform us of whether the preposition is one encoding direction or location; this is in fact a retention from a previous stage of the language that got lost in the regular prepositional phrase due to its superfluousness with regards to the way the verb started coding for the same things.

vərxira mabi bunerŋa – a coat onto his/her shoulders
vaŋra  dəbi kumana – the ship, like a stone
ustira yilvi Erhana Makisna – peace, appearing in king Maki's days
sordara nisvi sarkarir – rain replaced by drought

Example number two could fit into a sentence like:

as parune vaŋra dəbi kumana - it sank, the ship like a stone.

The last example is from an invocation in times of drought:

viskar purkeb sordara nisvi sarkarir – you have closed the heavens, rain has become drought

Most of these preprepositional constructions have rather idiomatic uses, which can be hard to acquire. A systematic description is difficult to present. Sometimes, the preprepositional noun seems to be an almost-subject of some transition or observation, sometimes it seems a mere afterthought. Oftentimes, the dynamic forms are used with dynamic, telic verbs - thus implying that the change expressed by the preprepositional phrase was the result of the verb. However, sometimes such changes are linked to non-telic, static verbs. Sometimes, they may be the main content of a sentence - a throwaway semantically vague verb without subject or anything, and a preprepositional coding for most of the information.

In some registers, nominal verbs tend to appear - both as part of the prepositional phrase, and as the preprepositional noun. The syntax and semantics of these constructions get even more complicated.

Oftentimes, there's no 'obvious' non-clunky way of integrating the preprepositional nouns as arguments of the main verb - they're in complicated relations to the nouns in the verb phrase or to the verb phrase itself, and thus the "separation" into main clause and preprepositional tail helps indicate to the listener that some (culturally deduceable) complicated relation holds.

Detail #160: A Quirk in Plurality Congruence for Objects

Consider a language with some kind of number congruence for objects on the verb. Let's add a few quirks:

The indirect object is considered an object too for this purpose.

This would mean that 'I gave him a book' has plural congruence, since the number of objects is greater than one.

Let us assume there's no pro-drop in the language. Now, plural reciprocals and reflexive verbs simply obtain reflexivity by plural congruence:

they saw.PLUROBJ = they saw each other or themselves

Detail #159: "Indexes that are Off"

Lots of things and grammatical categories form rather reasonably indexed sequences, and for some of these, there's also a sort of 'obvious' starting point, i.e.

1: default – big

2: comparative – bigger

3: superlative – biggest

We can compare some other things where the indexing doesn't necessarily seem very much like the natural numbers, but rather like the integers:

-1: yesterday

0: today

1: tomorrow

Similarly, tenses could be considered as mapping to negative and positive numbers. A system that combines aspects and tenses could reasonably be seen as mapping to some subset of the rationals (-1.5, -1, 0, 0.5, 1, 1.5 or whatever you fancy, where '.5' is perfect or whatever. Coming up with a reasonably interpretation for that series is left as an exercise for the reader!).

Now, this isn't really a way of categorizing or describing or classifying or anything along those lines - saying that 'comparative is two' or 'today is zero' does not really inform us about anything at all.

What would be interesting, though, is derived forms that reuse morphology from one category, say the natural numbers onto something where the number metaphor suggests a mix of negative and positive numbers. Say

1 → -1: "day": yesterday

2 → 0: "day-er": today

3 → 1: "day-est": tomorrow

I have not come up with any more interesting examples than this – that'll basically end up a challenge for the reader. I imagine such twists could be interesting though.

Barxaw: Diminutives and Augmentatives

Barxaw has both diminutives and augmentatives. They are formed with the particles tíi (diminutive) and ŋò (augmentative). They have certain syntactical quirks.

1. They can replace counters.

2. Unlike adjectives, they are not used as predicates.

3. When replacing counters, they're also doubled just before the noun.

4. 

The social role of diminutives is to reduce potential tension, by - as it were - diminishing the importance of the things talked about. Of course, they also serve to mark diminutive sizes. A certain amount of lexicalized information is necessary.

Tíi sùy genuinely denotes a 'small meal', whereas ŋò sùy denotes that the meal is of some importance.

Tíi də̀m signifies any matter about which there's no reason to be overly concerned, ŋò də̀m signifies a physically major conflagration. Tíi quyén signifies a specific breed of small dogs, quyén signifies horses. Ŋò quyén usually signifies breeding-quality horses, but is attested in some dialects to signify a nearly monstrous mythical being. 

Any serious dictionary of Barxaw would include separate entries for all such separate meanings. For most nouns, however, assuming big or important for ŋò, and small or harmless for tíi should work out fairly well. 

Sociolinguistically, diminutives are used in similar ways in all classes - with women also using them slightly more often than men in all classes - but they are avoided in communication over class boundaries, except when lexicalized - although even then, synonyms are preferred if possible. Augmentatives have no such restriction, and are often used even in communication between state officials.

Barxaw: Counters

The syntax of counters has been described previously, but no catalogue has been presented this far. These are only the most frequent ones. The original meaning - and commonly used meaning when the counter is used as a noun is given to its right. Things for which it serves as a counter is given with indentation. Quite a few of these do not fit into the "heads of cattle"-style construction in any literal way, but should rather be understood as, say "a congregation of N (people) ..." and the like. Notice that this is but a preliminary sample.

evé : congregation.

a counter used for groups of humans; however, not used for groups primarily distinguished by familial ties - for familial groups, use éŋmu

màg : stem

trees and large plants

pin : straw

small plants

sún : bushel, bunch

long thin things (mainly of agricultural origins, but also some types of fish)

tipá : bag

flours, dust, dirt, gravel

qiní :  small bag

salt, spices, jewellery

dim : a small bottle

potions, strong drinks, fragrances, various important ingredients in Barxaw cooking

sìdu : spoon

bits of foodstuff, spices, strong sauces

bán : disk

text and therefore also letters, orders, agreements, wood for building purposes, stone for building purposes, metal for building purposes

qìx : stick

tiny tools, such as cutlery, and various such things, pens, some small edible things.

mùm : bucket

liquids, various small tubers, fruits, leaves for culinary uses

fím : palm

a unit for breads that are bshaped by pressing with the palm before baking

busí : fabric

used for clothes, bags, and textile-based parts of housing

 méx : herd

most tame animals

wùr : snouts of carnivores

wild carnivores

xáp : "ears"

cats, dogs

qàm : mast (obsolete as a noun)

larger sea-faring vessels originally, but generalized to all vessels - even land-based, and now also temporary dwellings. Sometimes metaphorically for graves.

A sample of a few less common ones:
mim : row

structural parts in architectural contexts.

pál : glass of strong drink

used as a counter for certain types of agreements

Detail #158: An Origin for a Causative Construction

Let's imagine a language that has developed out of a formerly inverse alignment language or an obviative-proximative language into something more "average". Now, we also know that "and" tends to have a temporal ordering effect when coordinating VPs, i.e. an utterance like "he stopped and got out of the car", we parse this as his stopping first and getting out of the car later, even for verb-pairs that aren't necessarily as clearly temporarily ordered. We also like to think there's a kind of causal link unless we have good extralinguistic knowledge to imply there shouldn't be one.

Anyways, now imagine that the direct marker or proximative marker is generalized and a case system starts distinguishing subjects and objects. However, the obviative congruence marker (or the inverse marker) survives in one context and one construction - originally, following a conjunction, it had a causative meaning, i.e.

Subject Verb₁-DIRECT Object-ACC and Verb₂-INVERSE

Subject Verb₁ed the Object and made it Verb₂

Over time, the requirement for a Verb₁ to be present weakened:

Subject and Verb₂-INVERSE Object

Subject made Object Verb₂

Of course, now 'INVERSE' (or obviative) no longer signifies INVERSE (or ...), but rather 'causative'.  

"And" as a grammaticalized thing is known from Hebrew, but there it has somewhat different use. This could of course be very fun if, in addition, and and with were conflated in the language. 

Detail 158: Even More Inverse Alignment and Ditransitives

So, the situation for which I came up with the least elegant solutions for a wider inverse alignment was the ditransitive situation. This is a result of the factorial increase in rearrangements, obviously.

one argument: one possible arrangement

two arguments: two possible arrangements

three arguments: six possible arrangements

Having to have markings for six possible arrangements does feel mighty wasteful, even if you construct the markings from a smaller set of morphemes that combine to create the six possible syntactic role assignments.

A suggestion gleaned from IRC (thanks my favourite Swiss-Sqiptar person h13) is to incorporate the theme into the verb, but this of course only works if incorporation is a thing in the relevant language. 

Now, one thing we may notice is that IO and Subject often may be of comparable levels of agency, animacy, etc. Of course, there's no guarantee that this would hold, but ... let's assume it does hold in this language.

One thing we could do is have the inverse marking on the verb only affect the relative order of two nouns (usually the two that are the least similar in agency/animacy/etc). The third noun, however, takes an explicit case marker. The verb then simply gets to assign the two remaining roles - and maybe it'd be nice with one extra marker for the verb for the situation when it's IO vs. Subj that are being inversed, or some other arbitrarily picked pair.

Thus we get, assuming man > bear >> fish. Notice that I've set the definiteness arbitrarily, but of course these different phrasings could possibly affect definiteness, i.e. maybe the case-marker tends to be drawn to definite nouns. Notice however that definiteness in this language is not quite the same as in English - unmarked is just that: unmarked, can be either definite or indefinite, marked is explicitly definite. :

man kill-DIRECT fish: a man kills a fish

man kill-INVERSE fish: a fish kills a man

no case markers:

man kill-DIRECT bear fish: a man kills a fish for a bear

*man kill-INVERSE bear fish

case marker present: 

man kill-DIRECT bear-DAT fish: a man kills a fish for the bear 

man-ERG kill-DIRECT bear fish: the man kills a fish for a bear

man kill-DIRECT bear fish-ACC: a man kills the fish for a bear 

man kill-DAT.INVERSE bear fish-ACC: a bear kills the fish for a man

man-DAT kill-DIRECT bear fish: a bear kills a fish for the man

man kill-DIRECT bear-ERG fish: the bear kills a fish for a man 

man kill-INVERSE bear-DAT fish: a fish kills a man for the bear

man-ACC kill-INVERSE bear fish: a fish kills the man for a bear

man kill-INVERSE bear fish-ERG: the fish kills a man for a bear 

man-DAT kill-INVERSE bear fish: a fish kills a bear for the man

man kill-INVERSE bear-ACC fish: a fish kills the bear for a man

man kill-DIRECT bear fish-ERG: the fish kills a bear for the man 

man-ACC kill-DIRECT bear fish: a bear kills the man for a fish

man kill-INVERSE bear-ERG fish: the bear kills a man for a fish

man kill-INVERSE bear fish-DAT: a bear kills a man for the fish 

etc. (The remainder, of course, being the man killing the bear for the fish, wherein man-ERG, fish-DAT and bear-ACC occur. You can probably figure the verb markings out by now).

Personally, I prefer a system that omits at least one of the possibilities above - I wouldn't mind dropping the accusative or the dative altogether. Maybe have a system whereby accusative and dative are distinguished solely by discourse pragmatics or by some weirdo marking on the verb.

Maybe we could go even further and have a separate bunch of cases with somewhat restricted use:

two arguments high in the animacy hierarchy, one argument low: IO can take -DAT, SUBJ can take -ERG
one argument high in the animacy hierarchy, two arguments low: IO can take -DAT2, DO can take -ACC
roughly equidistant distribution: SUBJ can take -NOM, DO can take -ACC2

Anyways, I shouldn't be trying to fully exhaust the possibilities of weird things to do with inverse alignment - I just hope to show some examples of directions one could pursue within the world of inverse markings. So for now, I think the inverse marking spree I've been on suffices for a while.

Of course, to make it even more interesting, the available number of distinctions might not be the same for all genders or all numbers - plurals and low-animacy nouns might not distiguish ACC and ACC2, for instance.

A Link (a webcomic)

Jumping a bit late on the bandwagon (since badconlangingideas got there first), but ... just in case I have readers who don't also read BCI (like such readers exist, right?), I might as well recommend this webcomic:

Grammaticality and other Judgments

A pretty clever webcomic about linguistics. Should appeal to linguistics nerds.

Detail 156: Inverse Alignment and Splits

It seems to me that the most natural split-alignment where an inverse system is part of it would be along these lines:

present tense, non-perfect(ive), ... : inverse

past tense, perfect(ive), ... : ergative

The appearance of the ergative part following the same lines as usual in split-ergative languages. However, we could maybe do something else. How about a scenario as follows:

A nom-acc or erg-abs language that has a reflexive possessor along the lines of Swedish "sin", i.e. 'his/her/its own' where his/her/it refers to the subject (or the absolutive, if we want hardcore syntactice ergativity in the ergative subsystem). We form a bunch of adverbs along the lines of "after doing VERB" and so on as "after his_refl VERBING y", whereas intransitives don't take subjects: "after VERBING", and with the other subject we get "after his_nonreflexive VERBING y". We go on to grammaticalize these types of adverbs into predicatives. We lose the reflexive pronouns elsewhere, after first reanalyzing them as part of past tense agreement morphology. We further reinterpret the two morphemes as referring to whichever noun is higher on the animacy hierarchy (or other relevant hierarchies) rather than to previous subject/previous (dominant?) non-subject.

Now we have a nom-acc or erg-abs language with a good start at getting an inverse system rolling for past(/...) verbs. One thing that would help rid these past tense verbs from a more clearly nominative-accusative pattern is if non-finite verbs take nominative rather than accusative objects. This does happen in some languages for whatever reason so no weirdness there - except we of course need some rather specific circumstances for that to vanish. For an ergative-absolutive language, the problem with accusative objects steering the thing back onto an erg-abs pattern is not applicable. 

However, since ergatives and genitives often are morphologically identical, it could initially seem unlikely for something like this to appear, as the non-finite past tense verbs would be parsed as having ergative subjects anyway - but there's the clever thing, the reflexive/nonreflexive distinction for possession might not really apply for pronouns-as-subjects but very well for pronouns-as-possessors. Or alternatively, they may be to distinguish whether it's the previous ergative or absolutive noun that is referred to in a new sentence (or subclause), and this - I would figure - increases the likelihood slightly for something like this to come about. 

Detail 157: A Register Thing

Imagine a language with a register for certain religious holy days. On these days, lots of distinctions are conflated: 'dog', 'lamb' and 'cow' generalize to cover all four-legged furry animals; lizard generalizes to cover four-legged and fewer-legged small animals (including also small mammals - which are included under dog/lamb/cow as well). All tools for woodwork are 'hammers' or 'saws' - depending whether they can be used to merge or split things.

The majority of words that are not "prototypical" for a wider class are left out of use on these days.

This also happens with verbs; however, multiple different conflations may overlap partially, and sometimes, two verbs are used together to 'cheat' the taboo - i.e. "walk stand" for walk, since "walk" now includes dancing, crawling, trodding, trampling, moving about by cart, riding horses, etc; "stand" includes standing, sitting, lying, etc. Likewise, nouns sometimes are combined - dog lizard for mouse, for instance.

This idea was posted out of order with regards to idea 156 because it occurred significantly later (I had already written most of idea 156, but this took way less time to write).

Finally, adverbs too are made less specific - tomorrow now signifies any future time and is the only word used for that, yesterday likewise in relation to the past.