Detail #357: A Syntactically Split Alignment

Most split alignment systems only have split alignment in the case marking, and not really in the underlying syntax.  The split often also either correlates to the TAM of the verb, or to certain grammatical persons. I don't recall seeing any other actual cause of an alignment split, and I would really be interested in hearing about any other triggers for it.

However, that's not what I am writing this post about. This post is about a split in syntactical alignment based on a semantic distinction among verbs. I further find this particular split rather likely to occur.

The particular thing I'd think would cause an intransitive subject to align with an object syntactically is existentialness. Verbs such as 'exist', 'be' (when used existentially), be found, be seen (maybe), be attested, etc all seem to lend themselves well to prefer an ergative syntax.

The most obvious phenomenon to investigate would be gaps, and we can easily imagine a language where

berries exist and I eat

would parse as

berries exist and I eat [them]

however, we can then expect that coordination over many VPs should also yield ergative patterns:

berries existed and I saw and ate and tasted sweetberries existed, and I saw and ate [them] and [they] tasted sweet

This should not be permitted when all verbs are just 'plain vanilla intransitives'

*the berries were tasty and I saw and ate and tasted sweet
*the berries were tasty and *I saw and *ate and *I tasted sweet

Of course, when multiple VPs are chained like this, the later the existential verb appears, the harder parsing correctly will be. For this reason, the language might either forbid existential verbs to be on the right hand of coordinations, especially after more than two or so,  or have some kind of 'de-existential' form that is semantically, but not syntactically existential. (This could be achieved by reusing some other thing from the language - maybe force quirky case existential subjects? maybe use usually non-finite verb forms? maybe have the usual existential verbs lack congruence, but forcing congruence on them turns them into 'regular' verbs?

Ćwarmin Geographical Terminology

Ćwarmin covers a relatively large area of plains, with some mountain ranges at the edges of the area, and a few hills and the occasional mountain dotting the plains. Two oceans also touch the plains. Due to the distances involved, the words for the ocean differ, as do the words for 'waterfall', which for obvious reasons only really occur at the mountaineous edges of the area.

Bodies of water:

sućma - lake
ləkir - swamp
wire - the ocean (northern word)
kaśku - the ocean (southern word)
telin - the coast (northern word)
sterim - the coast (southern word)

It is to be noted that the southern and northern words for the oceans are not 'names' – the same word would be used by a northerner (or a southerner) regardless if he's seeing the southern or northern ocean, or even some other ocean altogether.

sasra - river
sasruta xamku - waterfall (river-gen fall) (southeast)
kaluta xamku - waterfall (water-gen fall) (northeast)
sasruta korsa - waterfall (river-gen jump) (west)
ontas - ditch, small river

savar - travel upriver by barge
sivir - travel downriver by barge

insə - a place fit for wading

apśuta - rapids, from the verb apśun - to splatter, splash

kalak - the left side of the river as seen while looking downriver
perək - the right side of the river as seen while looking downriver

Terrain:

kaŋud - plains

nile - a 'depleted' area of pasture
micni - an area of pasture, regardless if still abundant or depleted
məcən - move towards areas suitable for pasture in winter
mocon - move towards areas suitable for pasture in summer
eŋmər - a large grassy area
leśśe - a small area with grassy vegetation, or a part of an eŋmər
rende - an area with bushy vegetation
leśen - to graze, to cause to graze, to lead to pasture

falsuc - desert
ŋormo - an area strewn with rocks
ŋoron - to pick rocks (for building with)
ŋor - stone

miker - a low, lengthy hill
miken - to travel along the crest of such a hill
mokan - to travel across multiple such hills
śorka - woods
śoran - to bewilder, to cause to be lost, to confuse
nunto - plains of permafrost
nunun - to be frozen
ərtər - cultivated land, from Bryatesle 'ırtız', acre.
ərtən - to cultivate

rogos - moor
bənel - marsh
sildil - quagmire
sildilin - to quake
rab - valley
kup - peak
kupon - to reach
walgor - mountain
walgrona - mountain range
fośtor - volcano
fośton - to explode (with anger), to erupt, to eject fire

parsu - glacier

kuruk - salt plain
kurmu - salt lake

egəd - road, naturally easily traversable path
egdin - to create a path

Among the verbs we find three interesting pairs here, savar/sivir and məcən/mocon and also miken/mokan. These are just three among many sets of lexemes that would seem suggestive of some kind of ablaut system. Similar hints exist in other ĆŊ languages as well.

Detail #356: Case, Gender and Copulas

In some languages, some complements of copulas can be in non-nominative cases, e.g. in Russian and Polish where they sometimes are in the instrumental case. A situation where such a thing could make sense in a language could be when there is some form of perceived gender disagreement between the complement and the subject, e.g. situations like 'she is a soldier', and this could make sense in a language even if the language lacks grammatical gender. However, I guess it would make most sense in a language with a grammatical gender system, whenever that gender system provides a mismatch.

A Number and Numeral-Related Thing in Sargaĺk and Dairwueh

There's no need for a language to have a 'perfect' analogy to the word 'both' (despite the fact that it exists in several subfamilies of European as well as in several Uralic languages, and these are only the ones I've been able to verify that they are not direct cognates or derivatives of the lexeme for 'two'). However, Sargaĺk manages to double that, by having two words with similar meanings but different morphosyntactical behaviours as well as slight differences in meaning.

In Sargaĺk, two is yor. 'Both' is either vrir or lyəs. Vrir takes a formally singular noun after it:

vrir	tame	-ta
both	man	pegative sg
both	men

There are some complications: vrir does not distinguish the absolutive and pegative. For nominative or pegative nouns, it is always itself unmarked, but has the pegative singular marker on the main noun. For all other cases, the noun is in the singular case, and vrir takes the singular oblique case congruence:

vri(r)	-ə	tame	-rne
both	singular masculine oblique	man	singular lative
both		men	to

Lyəs however, takes plural congruence with all cases and the main noun too is consistently plural. As subjects the verb for both of lyəs and vrir take plural congruence, except if vrir is used with certain words like 'hands', 'eyes', 'ears', 'nostrils', 'scissors' or 'the side of a boat'.

Both of these can also be used as pronouns, much like English 'both'. They can also be used for a dual reflexive construction which can be used with any subject numbering two, regardless of morphological number.

The semantic difference lies in the extent to which the two referents are seen as separate units ('lyəs') or a concerted group ('vrir')

In Dairwueh, a cognate of vrir exists, ŋrəz. This particle has a few uses that have developed out of an original meaning of 'both': in NPs it goes before any number to mark 'all N of', but without any explicit number present it signifies 'both'. In numeral complements it serves to mark the number as that of a group, rather than as a number of independent individuals.  This it also does with plural, indefinite determiners and pronouns, thus:

guni	ŋrəz	tirs
are.3pl	(both)	six
they are		six

They are (a group of) six.

v.s.

guni	tirs
be.3pl	six
they are	six

they number six, there are six of them, (but as individual things)

ŋrəz is the only 'numeral' in Dairwueh to take case. It roughly follows the plural paradigms:

masculine:

nom: ŋrəzo
acc: ŋrizna
dat: ŋrizit
gen: ŋriŋa / ŋridin
loc-instr: ŋriŋa / ŋrider

feminine:

nom: ŋriri
acc: ŋrizar
dat: ŋrizit
gen: ŋrizin
loc-instr: ŋrizar

neuter:

nom: ŋrəza
acc: ŋrəza
dat: ŋrizit
gen: ŋrizit
loc-instr: ŋriŋa

Conlanger Lore: Lists of Cases|Tenses|...

This isn't quite a piece of 'lore', but it's a common enough thing in conlang descriptions. I will also have to mention some notable, very thorough exceptions.

Conlangers, even fairly far into developing a language, sometimes are happy just to list the cases, tenses, etc, without ever really describing their use. This betrays, in my opinion, a very naive (or essentialist, or whatever) view of what such things - cases, tenses, aspects, etc - are. This post will focus on cases, because they illustrate the problem fairly well.

One point I like to drive home is that names like 'accusative' are but labels, and the accusative of one language does probably not behave like the accusative of another. (For a scholarly source, see this.) They are not the same case except by virtue of having the same name. Yes, the prototypical use may be the same, but the prototypical use may be but one of the many uses of a case, and might not even be the primary use in practice – see, for instance, the plural genitive in Russian.

The naivety that I accuse this of showcasing is simply the notion that labels for grammatical things are somehow rigid references: all datives are the same, all accusatives are the same, all past tenses are the same, etc. This is far from the case. The dative of German, and the dative of Icelandic, to pick two very closely related datives are distinct cases. Despite sharing a name and even a historical origin, they are not the same case; yes, they share some properties - including some of their most frequent uses, but they also have several differences. For one, they don't go with the same prepositions (and of course, what I am saying about cases also applies to prepositions - 'in' in different languages differ!). Secondarily, they appear as quirky case subjects or objects with different verbs. Thirdly, being a quirky case subject (or object) is not quite the same thing in Icelandic as it is in German.

Looking at other languages with a dative, we find even more of a divergence between them. We also find that things sometimes go by different names but would fit very well in that category - e.g. the Finnish allative. As a sort of mid-conclusion: names can be both one-to-many and many-to-one, i.e. many things can carry the same names yet be quite different things, and many similar things can have different names.

As for non-case things, even pretty obvious categories like grammatical number may present a similar trap: the singular vs plural distinction is not the same in all languages – a trivial example would be things in general. Some languages prefer generic nouns to be singular, some prefer them to be plural, some seem to accept both ways by different ways of delineating them (e.g. lexically determined vs. influenced by grammatical context vs. other things.)

Tenses, moods and aspects, obviously, can present even greater differences.

To get back on cases, I would like to point to some good descriptions of case systems or even just locative systems that I feel avoid falling in the trap of 'just being a list''. Examples include Salmoneus' description of the locative adpositions of his Rawang Ata. Yes, this isn't about a case system per se, but functionally equivalent so you better just tolerate my use of it as an example.

A good example of doing a rather no-frills case system right is Carsten Becker's Ayeri. Some of the interesting stuff there appears in the interaction of case, transitivity and pragmatic concerns.

And a very naturalistic, alt-historicky Slavic case system is presented by Martin Posthumus in his Novegradian.

Of course, I am vain enough to toot my own horn here: I think there's some merit to my descriptions of my conlangs' case systems as well. The Bryatesle case usage description is fairly in-depth, but even then somewhat incomplete (see I, II, III, IV, V, and VI).  Dairwueh has a short, but sweet description that attempts to analyze the cases in terms of abstract features. Ŋʒädär too has a nice description in the same style.

Sargalk and Cwarmin still have not gotten that treatment, but it'll happen soon enough.

ANADEWs: Complications in Nominal Marking with Numerals

In many languages around the world, numbers beyond 'one' are followed by plurals, because obviously, two, three, four etc are semantically plural. Likewise, in many languages, numbers beyond 'one' are followed by singulars, because a plural marking is superfluous. In some languages, two, and maybe other small numbers are followed by some form of paucal or dual or whatever.

However, some languages mess this up a bit, and I figure it might be of some interest to describe two examples.

1. Finnish
The Normal Noun
If the noun phrase is any other case than nominative or accusative, the noun is in the singular and its expected case, while the number likewise is marked for that case. With the nominative or accusative, the noun itself is in the partitive case (which also is the case when the number is in the partitive), and the number is in the nominative form (or rather, numbers have identical nominatives and accusatives, except for 'one').

The Abnormal Noun
Some nouns lack singular forms, and can thus not abide by the rules laid forth above. Instead, the number adjusts, and is marked for the plural. This even goes for the number 'one', giving us monstrosities like

'yksissä häissä' - 'one-plur-inessive wedding-plur-inessive' - at one wedding

but also

yhde-t bilee-t
one-plur party-plur
a party ("ones parties")

This is even more sick, as ordinals too get this treatment, giving us ugly monstrosities like

kolm-ans-i-ssa festare-i-ssa
three-ORD-plur-inessive festival-plural-inessive
at the third festival

Of course, in Finnish each element of the numeral (except 'toista', roughly "-teen" as in thirteen and such) is inflected for the case of the NP, and each element of a numeral is also inflected for ordinality, etc.

Further, the comitative case lacks formally singular forms, and thus whenever that is used, the numeral also needs to be plural - even if that plural is one.

2. Russian
Russian has a peculiarity going on, whose origin is the defunct dual form. The dual was identical for some nouns in the nominative to the genitive singular (but not for all nouns, e.g. feminines had a distinct dual). This has generalized so almost all nouns, when following the numbers two, three and four, take the genitive (when the numeral is in the nominative, mind you!). With other cases, the noun and the numeral are in the same case and in the plural number.

With accusatives, inanimates behave like in the nominative example above. Animates, however, take the plural genitive from two onwards.

Certain numbers - thousand, million, billion - are really nouns, and the "real noun" is in the genitive plural.

3. Hebrew
In Biblical Hebrew (maybe in modern too; I don't know and will not try to find it out today - no diss of modern Hebrew, but Biblical just is so much more cool) the numbers three to ten take the opposite gender's congruence marker. Thus, 'five lads' would be five-fem.sg lad-masc.plur

There is also a 'construct'-number, which signifies 'n of', but has no gender congruence. These construct numerals can also take possessive suffixes for 'two of us' and the like.

Finally, in modern Hebrew, there is still a dual, but this is used only with:

• nouns that naturally occur in pairs, even for genuinely plural numbers of the noun, and with some pluralia tantum (that also naturally occur in pair-like structures, I guess?)
• units of time

Detail #354: Complete Omission of some low Numerals

Consider a language in which the use of a singular pretty much implies exactly one, and never a 'generic' referent. In such a language, the number 'one' could be entirely omitted in favour of always using a noun instead, much like how Russian sometimes uses 'raz' ('a time, one time') instead of 'odin'.

Now, in such a language, one can imagine that mathematical notation would not develop very well, since the idea of a symbol for 'one' might be less obvious if there's no word for it.

If the language also has explicit duals, we could even consider dropping 'two' out of it as well.

EDIT: This post was renumbered due to previous omission of #354.