Monday, March 27, 2017

Addendum (#333B): Some Details

One thing to note with a rule that duplicated the tense-carrying morpheme for creating habituals is of course that different types of verbs may differ as to whether they carry tense at all. One could easily imagine tense-carrying infinitives (after all, participles are a kind of infinitive) – but it's also conceivable that not even certain types of tenses would be classed as tenses in a language - their distribution and behaviour should really determine whether tenses A and B actually belong to the same class of 'things marked on the verb'.

Sunday, March 26, 2017

Detail #333: A Way of Forming Habituals and some Morphological Quirks to that Idea

An idea that probably has occurred to many is using reduplication for marking habituals. However, what about reduplicating the tense marker instead? Now, a different option appears: don't just reduplicate it, apply it twice instead.

What is the difference? Reduplication takes phonological matter and copies it, with some possible morphophonological rules applied. Applying it twice applies all morphological rules that are relevant, and then morphophonological rules. Thus, if there's two verb conjugations, and the first person singular past verb in the first conjugation looks like a  second conjugation stem, the first person singular past habitual would consist of a first conjugation suffix followed by a second conjugation suffix.

Different persons may behave differently with regards to that, due to the first inflected form appearance possibly deciding which conjugation the second suffix takes. Of course, one could have more conjugations that interplay in complicated ways as well. Other morphophonology could of course also apply.

One could of course consider similar things for plurals - duplicate case suffixes or gender suffixes, and do so by rules that make them vary a bit at times. However, this creates a fun situation with regards to the nominative, a case that oftentimes is not marked by any explicit morpheme. Maybe the reduplication then defaults to reduplicating the root or some syllable of it - or uses a different case marker, e.g. the accusative, thus conflating the two in the plural (not an unusual thing in the languages of the world).

Wednesday, March 22, 2017

Verbs of Perception in Sargaĺk

In Sargaĺk, there is a way of forming new verbs of perception that is fairly interesting. All of the suffixes mentioned here create citation form verbs when applied to roots.

The derivative suffix -at'on goes on either the noun stem for an organ of perception (which either is the body part that felt a touch, or the eyes, or the face in the face in case of olfactory and gustatory perception), or the ears, or on an onomatopoeic string representing a sound. This leads to some lexemes which are basically forbidden by Sargaĺk phonotactics appearing anyway: s:::::::at'on ('to hear a light wind'), k'rktat'on (which in the causative means 'to eat nuts'), pŕtpŕtpat'on (which in the causative means 'to fart'), y::::::::at'on (to hear a wind howling in something), auwo:::::at'on (to hear wolves howling), pst'at'on: to hear water sloshing

An NP or an adjective you've visually perceived can also have -at'on on it, or even a number - normally indicating that you've counted them by eye. The way of perception can also be mentioned even if the verb has incorporated some adjective or noun or onomatopoeia, then as an absolutive (dative) argument. A thing that has been perceived as being something will be indicated as an absolutive (accusative) object, however, the thing can also be incorporated with the adjective left as an absolutive object complement.

-at'p'a indicates the sound one makes when perceiving a thing, quality or sensation, or a quality one senses, or even a noun one becomes.

Detail #332: Articles for Names Intertwining with Moods

In some languages, even names take articles. Fairly un-exotic examples of these include some dialects of German and Scandinavian. However, we can go on and do some fairly interesting things here.

We can obviously distinguish a variety of social status and age by such articles, but another thing we could include in them would be mood! We can of course also consider case, if we go for a German-style system with case distinguished in articles. However, subtypes of the vocative could have different markings - maybe also distinguishing a different number of subtypes in different social statuses, and maybe also distinguishing both the status of speaker and recipient. Consider, imperatives. A particle could easily change to indicate that the verb is to be parsed as a command, or indeed as a confirmation of a command. An article could also indicate that the verb is meant as a question to the listener. One could also imagine various optatives and jussives and the like having special markings with names (but no special markings on other NPs, where auxiliaries, independent particles and verb forms serve the whole heavy lifting duties.)

However, one could also have the question-articles appear on non-vocatives whenever the addressee also is a participant in the verb phrase, so some non-vocative cases also may need to have question-address-forms. A system where some cases are conflated may appear with those forms.

 The historical origin of these forms may be rich in different types of lexemes: verbs ('hear', 'do', '(I) obey', etc...), nouns ('boss', 'word (to)'), adjectives ('kind', 'right', 'worthy', etc), adverbs ('immediately'), etc. The dividing line between these and cognates in the language is that stress patterns for these have been different, leading to significantly different sound changes over time.

Wednesday, March 15, 2017

Ćwarmin: The Polar Question and Its Other Uses

The polar question in Ćwarmin is formed partially by syntactical means, partially by morphological ones. In addition, intonation also plays a role, with rising intonation towards the end of the phrase.

The main morphological indication is a suffix, for human nouns -k[ə|o]r, for inanimates -t[ə|o]r. This is an exceptional morpheme in that it agrees with the animacy of the relevant noun – when marking a verb, it may distinguish whether the question pertains (more) to an animate or an inanimate argument of that verb. It usually is located on the verb, but can also be moved to a fronted NP. 

Here are different questions regarding whether Kier steals apples:
Kier-kər lend-itiś mesl-i
does Kier steal apples?

Kier lend-ijiś mesl-i-kər
does Kier steal apples?

lend-ijiś Kier mesl-i-tər
Kier lend-ijiś mesl-i-tər
Apples is what Kier steals, aren't they?

lend-(ijiś)-tər Kier mesl-i
does Kier steal apples?

-tər and -kər both sometimes surpress accusative marking, with inanimates even plural marking is sometimes omitted in the core cases before -tər. This is a place where some differential object marking actually occurs in Ćwarmin, with indefinites generally omitting the accusative or plural marking before the interrogative marker.

However, the same structure with -kər/-tər appears in 'regardless of'-style meanings, and in 'whether'-style meanings. In 'regardless of', the tone is falling throughout, whereas in 'whether' the tone has a light premature rise before falling.
bec a Alas staŋn-u-tor kinl-əc
you (c) Alas answer_yes-Q ask-2sg ?
you ask whether Alas answers yes?
'a' is an optional complementizer that introduces subordinate clauses. 
Parsing -tor/-tər in this way is only permitted with verbs of knowledge and perception.

The 'regardless of'-meaning usually goes after the verb, but can go elsewhere - clause-initial position or even just before the verb.
wərs-ic sewk-ər kurćap(-utćo) au-tor.
walrus-meat eat-1sg salty(-obj.compl.) is-Q
I eat walrus-meat regardless if it's salty
Autor sometimes is rendered as 'ator', aukor as 'akor'.

Friday, March 3, 2017

A Bit on the Ethics and Aesthethics of Prescriptivist Thought

This is a response to a thing written maybe two years ago in a facebook group; I hope whoever wrote it has forgotten about it and does not feel harassed or pointed out. This post has been a very long time in the works, and while I could come up with better examples, more in-depth explanations etc, I think I will just call it a day for this one by now. Indented text are quotes, all the rest is my own writing.Finally, I did ask for permission to repost this and comment to it, but I don't even remember who wrote the original text. No malice intended.
I think I've finally managed to put my finger on what exactly bothers me about language evolution, particularly of my native language. It's not the fact that languages change, period, that gets on my nerves; I accept without grievance the fact that the present is ephemeral and all things must change - and I would be proud of contributing to a shared achievement of our species that continued to grow in complexity, nuance, or efficiency.
This is a nice onset. Of course, the context in which it was posted puts some limitations on it: facebook comments and statuses and 'posts' don't really provide a venue for any depth. So, I'll point out a few things I find missing this far: metrics for complexity, nuance and efficiency.

Complexity itself is not an obvious concept here: complexity for complexity's sake is often among the most wasteful things imaginable, and therefore at odds with another desideratum: efficiency. What is complexity supposed to mean? Let us imagine a rule that says that words that begin in clusters cannot be preceded by the preposition 'for', but need to be preceded by the preposition 'otaque'. This would increase complexity without making the language any more expressive. Complexity is anything that increases the amount of data needed to describe the language's workings.

Further, we can come up with quite different metrics for the other thing I mentioned, viz. efficiency: the most important are probably precision in expression, data compression, noise tolerance, effort for the brain etc

Turns out these are somewhat incompatible too: more compression / shorter exponents for information leads to less noise tolerance. Greater precision leads to greater effort for the brain.
It's the fact that it doesn't do that, at all. The principle of least effort predicts that almost invariably, the people who will end up being vindicated by evolution are the majority, the people who put the least pride into how they speak and typically carry the least amount of linguistic knowledge. It's demagoguery, but made inevitable not through governmental decree but by the core tenets of human social behavior.
Given the quality of governmental decrees throughout history, I think we should be happy it's not for the government to decide on what your language is supposed to be like. Let us give an example of an inconsistent ruling given by a language academy, viz. the Swedish Academy. Much like some flavours of prescriptivist English, the Swedish Academy frowned on 'better than me'-style constructions. The justification was that 'it is a shortened form of 'better than I am', and thus calls for the nominative rather than the accusative form. However, the Swedish Academy permitted the use of reflexive possessive pronouns in that position, e.g. 'better than her own sister'. Here's the kicker: you can't under any circumstance have a noun phrase with her own as its possessor be a subject in Swedish, so clearly the Academy were either unaware of their own failure to be consistent or they didn't care that their explanation was mistaken.

Yet that particular alleged grammatical error was often used by teachers to harass kids in school, or by people who felt they spoke "better" than others to illustrate that they indeed were better. In many things, pride often precludes clear thinking. And language is one of the fields in which pride both precludes clear thinking and becomes a whip with which to punish those you don't approve of. (And often, those are of social classes who just don't have the time to invest to learn this things, nor would get any actual tangible benefits from investing it.)

The linguistic systems that have been formalized as Sanskrit, Latin, Greek, Finnish, Russian, etc, are in part of course inventions of a few scholars - but the bulk of them are the results of undirected linguistic evolution! Not a bad day's work done by random drift in a speaker community. Most of what those scholars did was just analyze what random drift had come up with.

These people whose linguistic knowledge is being put down by the poster to whom I am responding, are also among those who need language as a tool in their daily life, and they are representatives of the kind of hardware that needs to be able to interact with it - their brains contain the heuristics that sample the string of phonemes and do crazy good reconstructions of the underlying sentence that only sometimes get it wrong – and they adjust to whether they're often misunderstood or they often are met with the reaction 'oh you misheard' by adding redundant information that helps in recovering the message; and thanks to the magic of evolutionary design methods, this takes their brains' limitations with regards to memory and speed, their own limitations with regards to time they have at their disposal to invest in learning additional vocabulary and additional grammatical quirks to satisfy upper middle class wankers' linguistic aesthethics, et.c. into account. Luckily, most won't even desire to suck up to upper middle class wankers, thus prioritizing the other desiderata higher than the satisfaction of those upper middle class wankers.
I guess that's why I've always been, to my own occasional chagrin, more into engineered languages than ones designed to resemble natlangs, which - while I continue to have a lot of fun with - I ultimately only enjoy to the extent that they let us break human convention, not imitate it.
Anyway, rant over, I understand I'm being a pretentious ass, yada yada yada, we all get to have one thing we're like that over, mine's language. Just thought I'd share.

Yes, you are being a pretentious ass, but I also think you're being an ignorant ass, which in my view of things is worse.

Thursday, March 2, 2017

Inraj Sargaĺk Generation-Specific Terminology

Family terminology in Inraj Sargaĺk is one of the parts that most certainly are a holdover from the substrate. In part, a strong reason to suspect that this system is conservative is its unusual traits.

First, comparing the most immediate family terminology, we find that the system of distinguishing siblings, uncles and aunts by age is not well-established in Inraj Sargaĺk. Thus simi signifies all male siblings, and tame all female siblings.

The Inraj family terminology with regards to offspring, aunts, uncles, grandparents is specific to generations in a cyclical manner. Thus, a given person was born during the time considered to be the time of generation 1. He is considered the ospa of his father, who belongs to generation 3. His son will not be his ospa, however, but his ərok.

The graph below has 'female siblings' to the left, male siblings to the right, and parent/descendant in the middle - with females to the left, males to the right. Thus, aunts' and uncles' side of the family are not distinguished. The graph is cyclical, i.e. going downward past "mile / ərok", you get "adan / mota" again, etc.

Thus, a person of the ərok generation will have a mota for grandfather, a motbor for great uncle, an ospa for father and an ospor for uncle, an adkas for great aunt, a diskes for aunt and dise for mother. A mota will have an ərok for dad, etc. In the unusual case where greater spans of generations have survived, the prefixes mar-/mer- and sul-/sil- signify 'old' or 'young' to distinguish the two, e.g. maradkas : 'an adkas of the older generation when two adkas generations coexist', a sildise is the younger person that could be termed dise.
The most immediate family terms - sister, brother, father, mother, son, daughter - are often the same as in regular Sargaĺk, but in religious contexts even those are replaced by the terms here. For even slightly more distant relatives - uncles, aunts, grandparents, grandchildren - these terms are the usual terms.

Wednesday, March 1, 2017

Sargaĺk Possessives

Possessive formation in Sargaĺk utilizes two cases - the pegative as well as the absolutive. There are certain syntactical restrictions on both.

First, the default possessive case when the possessum is oblique is the pegative. Thus
However, if there are adjectives pertaining to the possessum to the left of the possessor, the pegative is blocked:
There are certain reasons why an adjective would be to the left of the possessor:
  •  Whenever the possessor rather signifies the kind of possessor, than an individuated possessor, or even the kind of  the possessum (e.g. fish guts) rather than an individuated possessum, the possessor and possessum are syntactically closer together than the possessum and its adjectives are.
  • Whenever the adjective distinguishes an individuated possessum among a potential multitude of possessum's owned by the same possessor (e.g. 'John's red hat (as opposed to his blue hat)')
Whenever the possessum is in one of the core cases, viz. pegative or absolutive, the same rule with regards to adjectives surrounding a possessor holds. Another rule that holds for core case NPs is that the case of the possessor is partially influenced by the transitivity of the verb, and the possessor may be dislocated from its possessum. The basic rule for possessors is: the more transitive the verb is, the more likely for a possessum to be pegative, and the higher up the hierarchy ditransitive subject > indirect object > transitive subject > direct object > intransitive subject, the more likely the possessor is to be in the pegative case.

It is not uncommon for a possessor of core cases to become a "pretend-subject"; this pretend-subject doesn't trigger any ditransitive marking or anything such on the verb, though. The pretend-subject can be marked for pegative even if the situation isn't "pretend-ditransitive". Since subjects usually go sentence-initially, this means the possessor can be offset from its possessum. It is generally speaking not possible to decide which noun is the possessum from any syntactical or morphological cues – contextual knowledge and a sort of noun hierarchy are relevant parsing cues.