Wednesday, September 30, 2015

Detail #212: Things with Participles

Make case inflection of participles affect the marker that forms the participle so much that in certain cases, different types of participles are conflated. E.g. "the robbing man" and "the robbed man" are distinct when these men also are the subjects or objects of some other verb, but when they're a comitative, the grammar doesn't give a shit whether they're robbed or robbing. This would be interesting with a slightly greater number of pariticple forms than English has, so something like recipient-participles and the product sets of tenses and voices, and not just a weird conflated thing like what English has (conflating present with active and passive with past).

Diminutives in Sargaĺk

Sargaĺk, much like its southern neighbours, has diminutives. It forms them in its own ways, however. Three common ways, of varying productiveness, are these:

Initial clipping: for many bisyllabic or longer nouns, the first syllable can be dropped or replaced by the vowel in it:
barxas : xas, axas sheep, lamb
setirmun : tirmun coat, vest
resvat: svat rope, short piece of rope
Derivational suffixes:
-pe- (m), -gi- (f), -sni- (m), -sa- (f), -se- (m), -si- (f)
Stress movement: for many words, stress movement to the last syllable can be an indicator of diminutiveness.
kádil, kadíl : tree, little tree
vípek, vipék : sharp edge, sharp point

Tuesday, September 29, 2015

Neuro-Linguistic Programming is a Pseudo-Science

Here's a paper. Pseudo-sciences rot the brain of their adherents, so please spread the knowledge that Neuro-Linguistic Programming is bullshit.

Monday, September 28, 2015

Animacy and Case in Sargaĺk

The Sargaĺk gender system has a further subdivision – there is an animate vs. inanimate distinction as well. This does not appear very clearly in the case morphology, but in some constructions they are treated in different manners. This distinction cuts through both of the genders, in effect giving us a 2*2 gender system. A relatively small number of nouns are inanimate feminines, the majority of inanimates being masculine.

1. Causatives
An animate causer argument of a causative is in the pegative or nominative case (depending on how transitive the resulting verb phrase is). However, inanimate causers take a postposition - ips - which requires the accusative case.

(A further detail: there can at most be one constituent in the pegative case in a VP, so a structure like "A made B give C D" comes out with A in the pegative, B in the nominative, C in the nominative and D in the accusative. However, if A is inanimate, B is pegative.)

Coordination between subjects requires both to be marked the same way. Usually, an animate subject can "promote" an inanimate subject to take the same case marking, but some speakers seem to favour the other approach.

2. Demoted Subjects of Passives
The agent of a passive verb can be represented using a comitative-instrumental for animates, but takes acc + ips for inanimates.

3.Subjectless Verbs
There is a set of verbs that do not, normally, take (syntactical) subjects at all. These include nagan, slumber, imbur, be temporarily settled somewhere, urdrys, to grow, izgər, to breathe, uvis, to whistle (or 'there's noise from the wind'), mondyr, to listen to, anmir, (of ice on the sea), to melt. For all of these, an inanimate subject is acc + ips, an animate subject is comitative.

Sunday, September 27, 2015

A New Approach to Copulas?

Some languages have an explicit copula. Some have some extent of zero-copulas. Some languages permit putting TAM-markers and person congruence on the complement, i.e.
car new-3sg.neuter → (the) car is new
man town-loc-3sg.pres → the man is in town
However, an approach I've not seen much of - except maybe borderlinely so in English - is marking the subject for TAM and person, and have the complement unmarked for such things.

If this is attested (beyond clearly reduced copulas as in English), I'd be interested in knowing of it.

Friday, September 25, 2015

Reduplication in Sargaĺk

Sargaĺk has a few types of reduplication.

  • Non-morphological Reduplication
    These are words that are made to reappear with the same marking in a sentence, with the same referent (if they're nouns), or the same verb (possibly with slightly different markings). With verbs, it restricts the meaning to very literal. In the present tense, it also implies habitual, whereas in the past tense it is very much perfect and intense.

    For nouns, it marks the importance of the noun (or adjective or adverb) with regards to the discourse.
  • Full reduplication in situ
    Intensifies verbs, With nouns it can be restrictive: only x, the only x. For some indefinite pronouns, this changes the meaning (any → some, (roughly speaking); one → each, what → is there even such a thing )
  • Single-syllable triplication
    This has a rather specific use: to imply that someone is going on about something too much. Any verb, noun or adjective can be used this way. The syllable before the stressed syllable is tripled.
    Inik mramramrasuta
    stop(imperative) (your) whine-acc.

    (you) with your cousins!
    savt kaveveveĺtva nisissu
    she speak-ak-ak-3sg.fem dog-loc
    she is speaking of dogs all the time

Thursday, September 24, 2015

Detail #211: Numbers with Rhythm and Metre

It seems to me to be pretty likely that some languages might have a very strong idea of numbers forming a very rhythmic and potentially rhyming structure. (Note: rhyme need not be the main repeating thing - it might be alliteration or any kind of phonological similarity combined with some type of rhythmic structure, really). Basically, you'd expect this to happen in some culture where counting is almost always done out loud, and where numbers seldom are used in any more abstract sense.

However, sound changes wear down the rhythmic structure, and undo the rhymes – however, due to the sprachbund having that thing for such structures, the number systems tend to be reshaped so as to replace the old, lost structure with a new structure. This means numbers change faster than other lexemes, as they are reshaped to create some kind of regular, appealing rhythmic flow and rhyme where it has been lost - and not necessarily with the same flow and rhyme as previously; new rhymes, new rhythmic patterns keep emerging.

Wednesday, September 23, 2015

A Question of Attestation: Extracting Evidentiality

Does any language with grammaticalized evidentiality permit extracting the evidentiality-marker with regards to mood and aspect?

I.e. the normal conditional
she killed the monster (I saw it myself) → if she killed the monster (no evidentiality?)
However, with a type of auxiliary, you would have
if she AUX (I saw it myself) killed the monster → if I had seen her kill the monster
For a conlang, we could here permit counterfactual statements not to have any evidentiality marker on the embedded verb, whereas factual statements with counterfactual evidentials will have evidentiality on both.

Sargaĺk: Taboo Lexemes

A number of lexemes in Sargaĺk have taboos over them. Unlike most similar taboos, however, the main type of lexeme to be subject to such taboos are verbs. These include verbs like
  • give birth
  • die
  • bury
  • bless
  • curse
  • breast-feed
  • care for (the sick)
  • hunt
  • trap
  • rowing, sailing and navigating
  • any verb relating to rites of transition
  • marry
Curiously enough, verbs related to sexual activities are not taboos. The taboos tend not to be absolute - using the "main" verbs for these things when talking about giving birth is considered acceptable for a while after the birth has been carried through. Likewise, from death to burial, speaking about the dead person dying is permissible. Similarly, the burial can be mentioned using that verb for about a forthnight after  the actual death. Blessing and cursing are only mentioned by those verbs when blessing or cursing. The "actual" verb for Breast-feeding is only mentioned by women, and only while having infants to nurse. Caring for the sick is only mentioned while not in the presence of someone who is sick, but also only when someone is sick. Hunting and trapping are only talked about using the "real" lexeme once the hunt or trapping has succeeded. Rowing, sailing and navigating are only talked about when participating in such navigatory activities.

The idea seems to be that talking about these activities with their "real" designation will attract the attention of malevolent spirits – they'll cause deaths so you will have to bury someone, they'll put madness in the minds of the bride or the groom, they'll kill whoever is approaching a rite of transition, they'll bring storms or sickness, they'll make the mother's milk run dry, they'll turn curses into blessings or vice versa, and they'll seal the womb so the baby remains there forever, and they'll scary off the prey. However, these spirits are slow and inefficient, whereas good spirits who are awakened by the actions will be strengthened by hearing the right words.

The words are replaced by more generic verbs:
bury: cover, with a reduced noun meaning 'ground' somewhere in the NP
give birth: cause to go forth
bless, curse: say, with a reduced noun meaning 'strength' somewhere in the NP
care for: to carry drink to
hunt, trap: to tame, with a reduced noun meaning 'food'
rowing: to pull wood
sailing: to catch wind
breast-feed: to
Over time, the reduced nouns tend to get assimilated into the verb, and as that form over time gets more common, it too can become taboo, causing a need for a new, non-taboo construction. The hamster wheel of taboo words keeps rolling.

Tuesday, September 22, 2015

Detail #210: Comparatives, Superlatives and Numbers

Use comparatives and superlatives to mark accuracy. Use negative comparatives to mark approximate, but less than or equal. Superlatives mark "more than" in general, and comparatives mark slightly more than or equal.

Stack comparatives and negative comparatives for "about N".

Onwards with #141: Other Ways of Combining Inverse Number with Regular Number

I previously posted about combining inverse number and regular numbers

One could imaginably have nouns keep marking number in a regular way, and have articles that appear if the unexpected number is being marked, so, e.g.

e men

e cow
 If the language is like English with regards to nominal morphological irregularities, we could also have cases along the lines of
e sheep
and cases like
e moose
I have noticed there's someone using google translate to read this blog in Spanish recently. That's not a big problem, but Google translate marks the latter nouns in the "e sheep-sheep", "moose-e moose" tuplets with a plural marker, and this makes the point invisible in google-translated Spanish.
Thus, in general, the nouns' number marking would be somewhat redundant, i.e. both the inverse and the regular marking appears. However,  for some nouns, only the inverse marker is of any relevance with regards to number, and for those both ways occur. In the given examples, sheep are expected to appear in flocks, and thus the plural is more common. Moose are more likely to appear one by one (with the exception of the cow and her calves for a part of the year ­ – but these groups too appear as individual groups, so you maybe a separate word for these groups?)

Monday, September 21, 2015

Introducing Sargaĺk

Sargaĺk appeared to me as an idea a few days ago, and it's already generating a fair share of notions with me. So, here's what I know this far about Sargaĺk.

In the conworld in which it's set, it is related - albeit distantly - to Dairwueh and Bryatesle. It is possible the notion of "relation" breaks down at the time-depths involved, and that the proto-languages of both just were parts of some very tight sprachbunds. Similarities do occur in phonology, lexicon and grammar, although seldom in ways that would lend themselves to any very regular reconstruction.

Sargaĺk is spoken off the northern coast of the continent on which Dairwueh, Bryatesle, Ćwarmin and so on all are spoken.

Besides the odd alignment (pegative-accusative?), it does use the pegative case for a few more uses that we will get into soon. First, the case system:
  • nominative
    The nominative marks most subjects, as well as the indirect object of ditransitive verbs and non-pronominal direct objects.
  • accusative
    The accusative marks direct objects, as well as some objects of adpositions. It also marks at what time something occurs. It is only distinct from the nominative for pronouns and a handful of nouns, including most nouns that designate times.
  • pegative-genitive
    The pegative marks subjects of ditransitive verbs. It also marks possessors in noun phrases. In a number of fossilized expressions it also marks origins and qualities, and there are traces of its origin as an ablative.  It also marks subjects of verbs of transition, i.e. "X becomes Y" is "X.peg Y grow/go/shrink/..." and "X makes Y be Z" is "X Y.peg Z.acc makes"

    Due to sound changes, not all nouns have unique pegatives, having been conflated with the nominative in some forms (and for the pronoun sib, distant third person masculine, the pegative and accusative are conflated). If such a noun is supposed to be in the pegative and is in an unusual position with regards to word order, it often gets an adverbial close by. Which particular adverbial depends on the verb.
  • comitative-instrumental
    The comitative marks instruments and accompanying people. Both of the objects of verbs meaning things like "replace, change, switch" are also in the instrumental case.  (Holds, sort of, for one verb, referring to barter trade?)
  • locative cases
    • locative
      The locative marks, as the name implies, location. It also marks by whose opinion, until which time, out of what material, for what reason.
    • ablative
      Origin, from which time, by whose permission, by whose perception.
    • lative
      The lative marks to where. It also marks at what intervals or during which recurring time (i.e. 'during the days' or 'each day') something occurs.
Beyond this, there is a comitative that only appears with humans, and is only used for extended family members of the speaker's family.

The morphological markers are as follows:

mascfemplur mascplur fem
fam. com-mime-mimis-mimes-mimes

So much for nominal morphological tables.

Word order is SOV, and we have the usual order-related things that tend to go along with that.

The verb is more complex than in most nearby Ćwarminian languages: several aspects, markers for ditransitivity (and on some verbs for transitivity), gender congruence (to some extent), many moods and a fair share of stackable derivative morphology.

The adjective is fairly simple, but it does have - unlike almost all of my conlangs - a fairly close equivalent to the comparative and superlative. These morphemes also can be used on verbs as well as adpositions.

There are also a number of nouns that appear in slightly reduced form as particles, and help distinguish various meanings of a verb, e.g. drown, disappear, be covered, be hidden, be lost, flee, be buried all use the same verb, but with water, sky, hand, mind, people (or carnivorous animals), ground as reduced nouns somewhere in the clause. A huge number of nouns have such forms, but they are not entirely regularly formed.

This is but an early draft of the Sargaĺk language's features.

A Dissertation with a Fair Share on Subject and Object Case in Finnish and Estonian

Here. Definitely worth a read. Its main topic is second language learning, but the particular case usages are the main phenomenon used to measure and reason about it.

Saturday, September 19, 2015

Detail #209: Case-like Affixes for Finite Verbs

Let's consider a situation along these lines: a language has TAM markers that serve to subordinate a verb in a sort of adverbial way: the verb now marks, for instance, reason, temporal things, and so on, e.g.:
eat-1sg-duration sit-1sg
I sit while eating
More complicated notions can easily be expressed:
fear-1sg juggle-2sg-reason knives
I am afraid because you juggle knives
In combination with aspects, we can turn duration into "ever since", "until", and so on. 

The language does not have a comparable set of things for nouns, but can use auxiliary verbs to some extent:
harvest being-3sg-reason have-1sg no extra time
because it is harvest/because of harvest, I have no extra time
The language might also have some more normal adpositions or such for similar uses with regards to nouns. However, and here's the fun bit: the language has no nouns for time units, holidays, seasons or moments: it has verbs.

These verbs generally are in the third person without any subject, but can at times take normal subjects. For instance,
day-2sg-duration drink-2sg enough?
do you drink enough during your day?
how christmas-2sg-past-perfect?
how was your christmas?
Certain verbs conflate time-spans with things we do in those time-spans: night and sleep are conflated, but can be disambiguated with an adverb. Likewise, wake up and morning are conflated. (The imagined ethnicity place morning much earlier than we do, due to agricultural needs.)

Weeks (or their equivalent), months (or their equivalent), years, minutes, hours, etc all have their own verbs. We also get a kind of "case rection" kind of analogue with regards to verbs with embedded verb phrases - 'to think that X verbs', for instance, takes the embedded verb in the consequent form:
moment-3sg-duration thought-1sg go-2sg-future-consequence
for a moment, I thought you were going mad
Similarities to case thus abound, yet these markers only appear on verbs. 

Friday, September 18, 2015

The Verb Alignment in Sargaĺk

Among the isolates that are geographically surrounded by Ćwarmin-Ŋʒädär we find Sargalk. It is spoken on a number of islands off the northern continental coast. The Sargalk subsist on fishing and hunting large marine mammals. It shows several typological similarities to the ĆŊ languages as well as to two other isolates in the region. However, lexically it has a number of tantalizing similarities to the Bryatesle-Dairwueh family.

One particular detail that attracts attention, however, is the "indirect ergative" structure of its alignment. Transitive as well as intransitive sentences follow the familiar nominative structure:
nen ood-as
I  sleep-1sg
I sleep

te ood-ar
you sleep-2sg

nen saĺp'a kin-es
I   meat eat-1sg
I eat meat

angas nev-ex sruk'-ju
walrus me-acc notice-3sg.perf.
the walrus spotted me
 However, the situation goes weird once ditransitives are introduced:
ne-tta te saĺp'a gʒup-s-an
I-"erg" you meat bring-1sg-[ergative marker]
I bring you meat
The "ergative case" is also used as a genitive case. In verb phrases with no object, but a recipient, the recipient is either marked with a postposition ('ete') or a dummy object is introduced ('ir', also the root of the inanimate indefinite pronoun). Two types of divergent dialects occur: on the two centralmost islands, it is permitted to omit either the subject or object (without change in marking), mark the verb with the ergative marker, and thus show that the nominative argument is the recipient (whereas the object and subject are marked with the accusative and ~ergative). The other type permits marking the recipient with either nominative or accusative. It seems speakers of the latter dialects prefer taking the case which the referent of the noun is less likely to be the other possible interpretation of - i.e. a noun that could likely be the object is marked with nominative, a noun that could likely be the subject is marked with the accusative.

If the ergative congruence marker on the verb is omitted, the person suffix should agree with the recipient. 

Further, a handful of verbs require the ergative for its subject even if they're intransitive. These include "build", dreĺiʒe and "survive", mak'ugu.

The voices interact unexpectedly with this: the recipient or the direct object can both be promoted to subjects of a passive transitive verb. The subject, the object and the indirect object can also all be demoted to adpositional phrases: k'ik for the subject, ete for the indirect object, and ʒva for the direct object. However, there is an antipassive of sorts that promotes the subject to the nominative case, and demotes the recipient to an oblique argument (again, marked with ete). 

Syntactically, the subject is a subject regardless of whether the verb agrees with it or the recipient with regards to gaps of this type. Thus, the following two examples have the same subjects for all verbs:
she-erg gave-3sg-ERG me a kiss and told-1sg a secret
she-erg gave-3sg-ERG me a kiss and told-3sg-ERG a secret
she-erg gave-3sg-ERG me a kiss and told-1sg-ERG a secret
means that I told a secret, and presumably to her.

As can be inferred from the examples, the ergative marking often can be used to imply the presence of an implicit recipient.

Thursday, September 17, 2015

A Conreligious Detail: Competition for the Priesthood as Source of Scripture

Consider a religion where a number of families have some kind of legitimate claim to the position of high priest. Being the high-priest family brings perks beside status: it gives some right of taxation of the population.

Consider if such a religion is rather conservative with regards to its ritual: even the slightest change could be seen as invalidating a priest's legitimacy, and therefore also forcing the current family of the high priest to abdicate from their position.

Pointing out these changes requires good memory, though, and if there's daily rituals (with different details for almost every day), an attempt at calling someone out for changed rituals requires a great deal of knowledge as well as the resources to keep a person posted at the temple scrutinizing the activities of the high priest and his attendant clergy (from his tribe, obviously). So, only competing families that currently are fairly well of could post someone there, and even have the records that codify what has previously been observed.

Originally, different families may have their own codifications. Disputes as to what are the accurate codifications may arise. Some kind of arbitration process between different codifications arises, and slowly leads to a consensus version.

The scripture has little to no impact on the life of the regular believer, but for a handful of families, it is an important political tool. Not all these families can keep observers constantly, so attempts every few years might occur - a family might have a particularly observant (and literate) son, and figure that this year he's not needed on the fields, this might be a chance to the priesthood. Might be that some certain particular rituals of importance are more likely than others to go wrong, so when the high priest is getting old and more likely to misspeak something due to frailty, representatives of several families show up at all the days of difficult rituals. Of course, the priestly family also has its own reader there, following along so as to prevent false accusations of infidelity to the text.

Over time, language changes but the texts do not, and during several years it might be possible that only a few token readers have appeared on a few days. Thus, there is a possibility that sacrificial rituals which are carried out by learned rote no longer fit with how people understand the codified text - the term for some body part of an animal has changed, and now the sacrifice is carried out in a way that disagrees with the language. Terms for movements change, and suddenly a certain invocation done with a certain movement is invalid.

Detail #207: Even More Inverse-Alignment Things

Let's have a variety of ditransitive verbs: dative-centered, subject-centered and split ones. These signify the following situations:
subject centered: subject, indirect, object
dative centered: subject, indirect, object
split: subject, indirect, object
That is, there is a hierarchy - subject, indirect object, object. In this hierarchy, the leftmost element of the subset we pick is always the "dominant" one; we have all the possible subsets with two members. Dominance in this case means that this is what the most animate argument is parsed as when the verb has the direct marker. The other bolded argument is what it's parsed as with the inverse marker.

The third object has a specific position: it is the preverbal argument (or it could be the sentence-final one, or it could be the post-verbal one or whatever). Thus, the language permits any of
A and B are "bolded" arguments, C is the non-bolded argument. The verb has a marker for whether A or B is the subject (or (indirect) object).

We could go on a bit: maybe the indirect object and the direct object share a marker when they're the unbolded argument, i.e. the same preposition is used for both of them. An intransitive subject, however, also uses the same preposition, whereas a transitive subject of a dative-centered verb has a specific preposition.

Wednesday, September 16, 2015

Detail #206: Some Mucking about with Lexemes and Grammatical Number

Dyadic kinship terms are a somewhat interesting lexical detail for which we could create some interesting things.

Consider, for instance, nouns such as cousins. We have a few situations:
they are cousins (reflexive)
they are my cousins (non-reflexive - they need not be cousins of each other)
the two of them are cousins (reflexive)
all of them are cousins (reflexive)
all of them are cousins (partially reflexive - cousins of someone else among them) - essentially, "all of these form sets of cousins"
Now, we could go on with this; some nouns may only have one reasonable interpretation for the ~reflexive plural interpretation: 
father and son →plural father and sons (a single family unit)

However, we might also want to permit 
father and son →plural fathers and sons (multiple family units of two or more)
and maybe we want to distinguish this without really bringing in anything like "collective" plurals or anything (at least as a regular thing in the language). So, what if we form one of the plurals irregularly (i.e. unexpected conjugations, suppletion, small stem changes, unusual morphemes?) - but with a twist. For some nouns, the regular form marks the singular set with multiple members, the irregular form marks the multiple sets with pairwise or more members; for other nouns, the opposite occurs.

Tuesday, September 15, 2015

Does This Exist?

Does any language have different family terms apply before and after transitional rites? The closest thing I can think of exclusively relate to kin-by-marriage, i.e. you become a son-in-law when you marry someone, and in some circles I bet you get some kind of specific terms for almost-N-in-laws.

However, it'd seem obvious that, say, the word for son or mother or grandfather or sibling or even cousin might change due to a ritual. It could also be possible, of course, that these changes are not symmetrical - i.e. a change in one term might not imply a change in another, and there may be acts for different parts of a relationship that change different terms.

The only even remotely similar thing I know attestation of is that in some tribes in America, there are rites that turn the word cousin into wife.

A Play by David Crystal

David Crystal, a fairly prolific linguist who has written a great number of books on topics such as language death, historical linguistics, clinical linguistics, has also written - back in the 1990s - a play. The play deals with the topic of language death. It is obviously a dramatization of the topic, and deals with a rather traumatic form of language death (that of a tribe being devastated by disease) - but still, it manages, I suspect, to give a good hint at the experience of being the last speaker of a language rather well.

For people with an interest in linguistics, however, parts of it does read a bit like an info-dump. However, for most people I guess the info-dump might be genuine news. The bits that are impressive generally deal with the character Shalema, the last speaker of his language.

The manuscript is available for interested readers here.

Friday, September 11, 2015

Detail #205: Relative Ordinals

Have a set of ordinals that simply mean "n later in the sequence" and "n earlier in the sequence".

Detail #204: Forming Numbers Additively and Subtractively

Let us imagine a language that forms tens up to 50 "additively" in some sense - twenty, thirty, forty and fifty derive from the numbers up to six. Sixty, seventy, eighty and ninety, however, derive from the number of tens missing: ninety is one short, eighty is two short, etc. We could derive these maybe from the ordinals, so "one short of first five" = 95.

However, let's add a complication: small negative numbers were accepted quite early as a way of expressing debts. These exceptionally enough start out as a direct offset of the first hundred numbers into the negatives. Thus -5 is "one short five". However, from the first -100 downwards, numbers are formed by putting some word amounting to "minus" before the positive number phrase.

Thus down to -100, each hundred has the same "structure" to the individual numbers, but from -100 downwards, an inverse structure obtains.

Thursday, September 10, 2015

Detail #203: Restricting Imperatives to Certain Verbs

Imagine a language where only kinetic verbs with outcomes that physically affect the subject or object have imperative forms, whereas all other verbs conflate the command-form with some irrealis form, say subjunctive or somesuch.

Ćwarmin: Demonstrative Pronouns and a Suffix

Ćwarmin demonstrative pronouns are among the few words to have case congruence on them. The situation is even more complicated, since the congruence operates slightly differently depending on whether the noun is animate or inanimate - with inanimates, the demonstrative only marks for case. Definiteness and number are omitted. For animates, the demonstrative usually is either marked with the definite or the specific case marker, and number is also marked (although the paucal number may be substituted by the plural).
arna / ərnə = this
olba / elbə = that
Some morphological abbreviations do happen, however:
arna → artu / artak (definite/specific singular nominative)
arna → araś/arok (plural definite/specific)
More generally, the -na or -ba syllable is generally dropped, and if the suffix has more than one syllable, the first vowel is reduced.

Ćwarmin genitives cannot stand as independent noun phrases. However, with the suffix -(a|e)rn(a|e) or -(o|e)lb(a|ə) a genitive is turned into an independent ~noun. The genitive suffix is also slightly reduced:
-itite+ erne → -itern-
Unlike the full pronouns, the n is not lost when further inflecting this noun for case. Some syllable reduction does happen with regards to the first syllable of bisyllabic case suffixes, however. Although these originate with a demonstrative pronoun, all three definitenesses can be marked on this, correlating to meanings like
a so-and-so of so-and-so's
this particular so-and-so of so-and-so's
the so-and-so of so-and-so's

Wednesday, September 9, 2015

A Musing

For years now, I've wanted to come up with a morphological category for nouns that isn't case, but is somewhat like it. In some sense, a case marking that is decoupled from theta roles as well as from notions like subject and object - although I am not quite sure what remains to go for once those notions have been removed from the playing field.

I go from wanting to include things like animacy, volition, whether the outcome is beneficial or not to entire rejecting those as part of it. Other times I want some more exotic things: whether a noun is normally part of the action, legitimacy of claims or statuses, social status, mechanics of what happens in the event the verb phrase refers to, imagined emotions of the involved, importance visavis the discourse, etc. 

Never have I come up with anything I've been happy with - in part I guess the reason is that what I am looking for is just something vaguely other yet vaguely similar, and I want something fresh and unattested.

Onwards with Ideas for a Musical Language

Music is not just pitches ordered in time. There are a number of other things: duration, rhythmic patterns, tempo, tonality, harmony, counterpoint, timbre, intensity, consonance, dissonance, resolution - and for all of these, sustained values, quick shifts and slow transitions are possible.

I previously described some properties of the diatonic scale with regards to "relatively identifiable" subsets. However, if we go beyond monophonic music (i.e. beyond music where one note is played at a time, no more than that) some of those problems can be resolved without all too many problems.

One easy solution would be to keep a single drone constantly throbbing along:

C - - -
c d c e
The c d c e melody snippet is now distinguishable from f g f a or g a g b on account of the constant C drone. The others would be audibly different, e.g.:
C - - -
f g f a
Music with this kind of structure does exist, most famously perhaps in the tradition of the Scottish bagpipe, but also elsewhere - in traditional musics all around the world. Since we have an established tone that frames the entire thing, we always know what tone we're hearing (or at least, after a while of practising, we are likely to be able to identify the intervals in all but the noisiest of circumstances, for at least those timbres we've been practising with)

A rather central thing to western music (and conspicuously absent in all traditional non-western music) are chords and their progressions. These have some fairly interesting properties as well, which might simplify the 'identification' issue; since each traditional chord contains at least three notes from the scale, and these are quite evenly distributed:

Chord tones of the C major chord as per the cycle of fifths.

Chord tones of the C major chord as per the major scale.
There's only two relationships between major chords within a major key that are ambiguous: F to C is identical to C to G; mistaking G to F for either G to C or C to F is unlikely with some practice. Beyond these, we get minor chords:
A minor
C D E F G A B (c d e ...)
A minor
Now, the same holds true with regards to minor-to-minor transitions as what held true with regards to major-to-major transitions. The minor-major (and vice versa) transitions contains some more ambiguities, though:
Fmaj Cmaj Gmaj Dmin Amin Emin (Bdim*)
Cmaj Dmin Emin Fmaj Gmaj Amin (Bdim)
Dmin → Fmaj = Amin → Cmaj = Emin → Gmaj. Beyond this, however, we also have:
Dmin → Cmaj = Amin → Gmaj, as well as Emin → Cmaj = Amin → Fmaj.

It does seem we may have some chance of identifying them correctly even in isolation, but there's a significant probability that people without perfect pitch will misidentify these in isolation.

Chords have rather nice properties in general: they do sound like their member tones together form a "supertone" of some kind. They have functions in keys. They can easily be split into melodic snippets that also hint at the "supertone" just by playing them in sequence instead of simultaneously.

We could take that and run with it, though! Let's first introduce another chord concept: inversions. Basically, the C major chord is a C major chord regardless of which pitch is lowest and highest. In the following set, pitches are ordered so that low-to-high corresponds to left-to-right:
    E1     C2 E2 G2 C3
        G1 C2            C3 E3
C1    G1       E2
All these lines are C major chords. However, we can describe the different forms using the term 'inversion'. We have three inversions of triads: root position, first inversion and second inversion. Root inversion is, say, CEG, first inversion is EGc, second inversion is Gce. Modern theorists generally don't distinguish the function of these, although some baroque theorists did. However, and this is a funny thing: trained musicians are less likely to hear inversions as different chords than amateurs are. So apparently there's a thing there we might be able to use!

Let us use a different notation now, where r, f, s stand for root, first and second. A letter normally codes for "major chord", and a letter followed by an m stands for a minor chord. We could now maybe start doing some kind of Semitic-inspired thing. Most words are three chords, except we permit a few two-chord words where the chords may be any two chords that form the right relation.

We could add two other things: order of arpeggiation. An arpeggio is any way of playing a chord in sequence rather than as a simultaneous block. Since we only deal with three tones for now, we can pick six different ways - one starting tone out of three, one next tone out of two, and there's one remainder. 3!, in other words. We simply number the tones from top to bottom, and write the order as 123, 132, 213, 231, 312, 321.

So, we have 18 possible arpeggios per root - three different inversions, six different orders. 6 (or 7) * {r,f,s} * {123, ...}.

 So, about 100-120 different building blocks. If a word consists of at least three chords, normally, we get about a million different 'words' (we do end up with less than that, due to some reasons, but still we're in the hundreds of thousands). That should be sufficient, although our words are somewhat longish: about nine tones each. Of course, it's possible that background noise and such might cover a tone or so, and you'll end up being misaligned among those chords.

Thus, we can have words like:
If we restricted ourselves a bit, we could of course limit the onset to patterns where the first note is the lowest of the set: CEG, DAF, etc. We could restrict the types of inversion in the middle bit, and we could have some other restriction on the final bit. And now we could start doing something a bit "triconsonantal root"-inspired; maybe root voicings in the first element mark first person, first inversions mark second person, and second inversions mark third person; if it's ascending all the way it's plural, if the middle tone is higher than the last tone, it's singular. We might add that some "not quite chords" at the onset encode non-verbs, maybe any sequence of seconds at the onset encodes 'nouns'. For whatever reason, though, we prefer to retain the same starting note as the verb usually would have if we've derived the noun from a verb; [CEG][DAF][B'EG] inflected into a verbal form would either be [CDE][DAF][B'EG] or [CB'A'][DAF][B'EG]. We might derive participles or whatever by some other not-quite-chord: maybe fifths along this line: [CGC][DAF][B'EG]

Another thing we could use to encode different things, of course, is rhythm, as well as other things such as staccato vs. legato or somesuch.

We could also introduce some sound symbolism kinds of things: a very fast and high minor second motif like cc#c could encode "small", a slow similar thing over an octave could encode large. A segment of the chromatic scale could encode "all", whereas a pentatonic scale could encode "some". A trill could indicate 'many', glissandos '-ish'.

Imagination is all that limits us here.

Monday, September 7, 2015

Musical Conlangs: Some Observations and Suggestions (Part I)

Many musical conlangs fail to take into account the different modality of music contra language. The most common "template" for musical languages is one where tone straightforwardly replaces phoneme. So instead of /p t k b d g ... i e a u o ... / you get /A B C D E F G/ or some other set of pitches - possibly over more than one octave, possibly with length distinctions and such thrown in for good measure.

This, to me, seems like rather too simple a way of using music for language. It sells music short. This is the first post in a series that will look a bit into the structure of music in order to suggest more interesting

Most humans do not have perfect pitch. Thus, in isolation, the sequence A C sounds much the same as any out of the sequences E G, B D or D F. Lots of similar groups of two notes, where the two notes cannot be distinguished by just hearing the two notes themselves exist. Let us restrict ourselves to the diatonic scale for now.

In fact, we can identify some subsets within the diatonic scale that are not "unique", i.e. there will be other subsets with the same structure to them.

CDE=modulo pitch GAB
In fact, we find several similar structures where any melodic segment that only moves in that subset can be transposed with its structure intact into some other subset. We take a look at the structure of the thing:
F (5) - C (0) - G (7) - D (2) - A (9) - E (4) - B (11)
This is the cycle of fifths, a notion that might be familiar to musicians. The number gives an index - we simply index the tones by the chromatic scale. Simply put, the tones of the C major scale can be obtained by starting at F and ascending by 7 semitones repeatedly until you reach B. Now, this can be used to identify subsets with identical structures within them!

Any two structures that can be transposed one to the other will cover the same length of the cycle. Consider CDE and GAB:
Basically, we can form a metric for this! If we strip the melody of its melodic structure and only represent it as a set of tones, and then take the range from the leftmost to the rightmost tone as our metric, we have away of identifying whether a melodic snippet could have occurred elsewhere in the same key. The "width" of a melody in this metric is the difference of the leftmost and the rightmost tone present in it as per the following:
0 1  2  3 4 5  6
Thus a melody which jumps around G, D, A, E and B goes from 6-2 = 4, and same goes for one that moves around F, C, G, D and A – 4-0 = 4.

And it turns out the only "unique" span is that of F to B (however, looking at the chromatic side of things, it turns out B to F is exactly the same distance, and thus B to F is not distinguishable from F to B ... in isolation, that is.)

However, whenever F and B and a third note are involved, the resulting set can only happen on "one side" of the F-B endpoint. What can we do with this knowledge? Well, we could use 'structures' as lexemes instead of notes; thus, say, "F C D C" is a surface form of "0 1 3 1", where we simply add this to whichever starting point we use: if we were starting from G, we'd get "G D E D", if we were starting from D, we'd get "D A B A".

We could also try to do something that would ensure that a certain pitch is firmly etched as the "root" in the minds of the listener. Maybe we generate the vocabulary so that, say, forty percent of the phonemes in the dictionary are C, fifteen percent G, and each other tone something like 11%. In such a thing, it should be possible to orient oneself tonally fairly solidly.

Other possibilities include having some kind of cadenzas that introduce a statement (or appear somewhat regularly), grammatical prefixes or suffixes that occur in a limited subset (and therefore are easy to identify with regards to the scale structure). Finally, we could go for a different scale altogether, one where no structural similarities exist between any subsets. This makes for a very limited scale - if we permit similar structures between two-note subsets but not between three-note or larger subsets, we get more possible scales to chose from. Here's an example:
CDb Eb F#G [big gap] c

C-Eb and Eb-F# are identical, as are C-Db and F#-G, as are C-F#, F#-C, but other than that no other subset's structure recurs in another subset of this scale. However, we might want to stick to the diatonic scale due to its great melodic properties.

Of course, there's more to music than melody as well, and we'll get on that in the next installment.

Sunday, September 6, 2015

Detail #203: Quirky-Case Subjects in Isolating Languages

Imagine an isolating language that typologically has evolved a lot like English, but perhaps slightly further. However, it retains three or four cases for its personal pronouns. However, now we'll take a look at its properties before becoming isolating and after.

mar-ka ǵudu-rek wubin-t
man-acc wisdom-gen/dat possess_as_intrinsic_quality-3sg
the man has wisdom/ is wise
wáli-rek lar-rek du’ik-t
stone-gen/dat moss-gen/dat overflow with/exude-3sg
stones exude moss (a mistaken belief about how moss on stones grows)
Much like in Old Norse, however, what we translate as subjects here are not really  subjects. They differ syntactically speaking, and as far as congruence goes as well. The verbs in the example sentences have 3sg markers, but so would they even if the "subject" were any other person or number - 3sg doubles as a neutral present marker. This lack of real subjecthood for non-nominative subjects of these verbs means that:
  • no reflexive pronouns can refer back to these subjects
  • no gapping over coordination is permitted, i.e. "(the stone) remains there and [____] exudes moss to this day" is not permissible, nor "the stone exudes moss and [_____] is quite magnificent really"
  • the verb cannot be used as an imperative
  • no extraction (i.e. no constructions along the lines of "stones seem to exude moss", but rather you are forced to construct it as "it seems that stones exude moss"
  • unlike subjects, they are not obligatory; however, unlike subjects, they cannot be implicit

These cases have been lost except for the pronouns, as stated, and so we now get:

mar ubĩ xudu
man have_as_intrinsic_quality wisdom

merge ubĩ pura
me-acc have_as_intrinsic_quality toughness

weyi dyik lar
stone exude moss
merek dyik tur
I-gen exude sweat

What is retained, however, is that these verbs do not, syntactically speaking have subjects. The (partial) list of exceptional properties carries over - the subjects of these verbs are, for a while at least, exceptional. Over time, regular subject properties bleed over to them due to analogy with other subjects (and maybe the case marking on the pronominal subjects is dropped).

Wednesday, September 2, 2015

Detail #202: A Different Approach to Copulas

Consider a normal NP:
three red houses
the little man
those rocks in the valley
the two parts
a new idea
Consider also clauses such as 
the red houses are three / there are three of the red houses
the three houses are red
the man is small
those rocks are in the valley
the rocks in the valley are those
there are (the) two parts / the parts are two in numberthe idea is new
There's obviously a relation between these; however, doing something like conflating the two feels rather boring. As does just adding a verb to either end of the latter.

Adding a verb after it, with no other marking that distinguishes the complement from the NP obviously makes it unclear what particular adjective or determiner or whatever that is relevant?

So, we add a marker that tells us which particular type of thing the verb pertains to:
the three red houses QUANTIFIER_are: the read houses are three in number
the three red houses ADJ_are: the three houses are red
the three red houses DET_are: those are the three red houses
the three red houses NOUN_are: the three red (things) are houses
We could of course have some categories: several noun-markers (mass noun, identity wrt noun, type wrt noun), some classes of adjectives (by some semantic distinction). When several things of the same type are stacked, some ambiguity of course ensues. We could reduce the likelihood of such ambiguity by having, say, a few classes of adjectives (semantically delineated, I figure). Many adjectives might have synonyms in several classes, where the semantic difference also plays in, i.e. the synonyms differ in subtle ways by virtue of being in different classes. Agreement for the right adjective class then reduces the likelihood of ambiguity (since it's less likely that an adjective of the same type is present.) Stacked quantifiers might not be all that common, nor stacked determiners and such, but let's not assume they are entirely absent. The ambiguity, though, probably is not all that problematic.

Tuesday, September 1, 2015

Detail #201: Participles, Again

Let's derive verbs whose meaning aligns with "[to act upon something using a] tool". Let's derive those not by just zero-deriving them into verbs or by using instrumentals as verbs or something; let's instead take a slightly circuitous route: taking the instrumental case, attaching an active participle suffix, and then combining this with auxiliaries to form finite forms.

This would maybe work best in a language where the instrumental is not much longer than the regular form, say:
körs - körön (hammer, by hammer)
The participle should also probably be quite succinct:
nükörön ("hammering")
As an adjective, this would work fine: nükörön tarstas would be "hammering hangover", despite the fact that the hammer involved is nothing but metaphorical. However
bi nükörön ser  = I am hammering
bi nükörön sentri  = I was about to be hammering
A real passive is impossible for this construction, since this derivation strategy simply does not permit the passive participle prefix; instead, the object remains an object:
bir nükörön se = I am being hammered, "me is hammering"
However, the accusative argument is, from a syntactical point of view treated as a subject, e.g.
tor nükörön se ta stumdu mehede
him hammered is and still drink
he is hammered and still he drinks
(Assuming "hammered" has the same meanings as in English in this language as well)