Lkal sik

A forum for all topics related to constructed languages
Post Reply
User avatar
Posts: 711
Joined: 07 Aug 2013 00:26
Location: displaced from Helsinki

Lkal sik

Post by gach » 03 Apr 2016 13:47

Here's a side project language, called Lkal sik, on which I've been writing some notes throughout the winter and spring. You can think of it as a way to get something presentable while at the same time developing ideas for Kišta behind the scene. Lkal sik is one of those conlangs that draws inspiration from the Tibetosphere. I'm exploring a bit more rarely used evidential strategies with it by including an egophoric evidential to the language and am also using quite a bit of consonantal prefixes in a way that gives a flavour of Old Tibetan to the aesthetic. On the other hand, I'm considering making these prefixes originally classifiers on nouns so they might actually more resemble the Austroasiatic sesquisyllabic prefixes. On verbs the prefixes can be either directionals or vaguely classificatory fused nominal adjuncts.

Lkal sik is spoken in the same world as Kišta and is the language of an influential state called Skal located some distance upriver from where Kišta and its related languages are spoken. There is trade between the two areas resulting in loan words and new technologies adopted by the Kišta speakers, but also a possibility for political tension since states like this tend to expand their influence over time. I'll be describing here the classical norm of the language which was spoken and got standardised maybe 700 to 900 years before the present time in the world. During this time the language was a regional administrative language in a border province of a larger empire. Later the empire disintegrated and after getting its act together the province of Skal formed its own independent state with Lkal sik as its main language. I should eventually also describe the later stages of the spoken language, but this has to wait until I work out satisfactory sound changes from the classical language. Not surprisingly, especially the modern written Lkal sik rarely follows exactly the grammar or vocabulary of the old classical language and it has always been more or less influenced by the contemporary spoken language. The normative classical language continues to exist as a grammatical ideal.

Both the name of the language and the state are based on the root kal with different prefixes. The prefix l- in the language's name is found in many nouns related to sound or the mouth while the prefix s- in the name of the state is an areal prefix. The second element sik in the language's name simply stands for "speech" or "language" and is derived from the verb si, "speak".


The consonant and vowel phonemes of Classical Lkal sik and the orthography I'll use for them are

/p t k q/ <p t k q>
/b d dʑ g/ <b d dź g>
/f s χ/ <f s h>
/v r ʑ ʁ/ <v r ź ř>
/m n/ <m n>
/l j w/ <l j/i w/u>

/a e i o u (ə)/ <a e i o u (e)>
/ae ei ao ou/ <ae ei ao ou>

The glides /j w/ are written as j w at the start of the word or after vowels and as i u when they come after consonants. The off-glides on the low diphthongs /ae ao/ are something around [e̝ o̝] or [i̞ u̞]. They are lower than the off-glides on the high diphthongs but higher than /e o/ or the base vowels of /ei ou/. A big reason for deciding this was simply that when these diphthongs are preceded by glides, writing iae and uao instead of iai and uau looks much more pleasing. The vowel [ə] never appears on the main syllable and its phonemic status is up for debate. The word initial voiceless stops tend all to be aspirated.

The typical allowed word shapes are ((Cə)C(C))V(C(s)), ((Cə)C(C))VCV(s,n), and ((Cə)C(C))VCVCV(s,n). That's to say that a word may begin with a two consonant cluster and on top of that have a prefixed syllable of the form C[ə]-. Here the vowel [ə] is epenthetic and fully predictable and so I won't be writing it down. The allowed consonant clusters at the start of the main syllable are

/sp st sk sq/
/br dr gr/
/bʑ gʑ/
/Cj Cw/ (except for double glide clusters)

A prefix may end up forming an allowed cluster with the initial consonant of the main syllable. In these cases a true cluster is formed and no epenthetic [ə] is interted after the prefix consonant.

Words that have only one full syllable may end with the consonants /p t k s χ m n l/ or the clusters /ps ts ks ns/. In these places the nasal /n/ has the allophone [ŋ]. On longer words all consonants may appear between two vowels but no clusters are allowed and at the end of the word /s n/ are the only allowed consonants. The vocalism of the suffixal syllables following the main syllable is also reduced. Only /i u [ə]/ may appear there. I'll write the suffixal [ə] as e but it's probably better described as an allophone of /a/.
ImageKištaLkal sikSeic

User avatar
Posts: 711
Joined: 07 Aug 2013 00:26
Location: displaced from Helsinki

Re: Lkal sik

Post by gach » 03 Apr 2016 14:30

Tense and mood

There are two tenses in classical Lkal sik, an unmarked non-past and a past marked by the suffixes -is, -s, -i, -je. The basic form of the tense suffix is -is, which appears as the word final form after the main syllable vowels a (forming the diphthong ae) and e and all stem final consonants,

qel rne-is
SG3 come-PST
"he/she come"

qel buh-is
SG3 sleep-PST
"he/she slept"

The vowel of the past tense marker is merged with the high front vowel or glide and elided after the rest of the vowels, leading to the allomorph -s,

qel si-s
SG3 speak-PST
"he/she spoke"

If the the past tense marker isn't the last element in the word, it drops the sibilant and becomes typically becomes -i,

dźun rne-i-n
SG1 come-PST-EGO
"I come"

dźun buh-i-n
SG1 sleep-PST-EGO
"I slept"

Needless to say, there are cases leading the total elimination of the past tense suffix. In these cases there is a variant -je that's occasionally attested on some verbs,

dźun si-n ~ dźun si-je-n
SG1 speak-EGO(PST) ~ speak-PST-EGO
"I spoke"

There isn't a whole lot to say on the subject of morphological mood. The imperative is formed with a simple suffixless stem verb.

The evidential system

Classical Lkal sik has a three way split between an egophoric (or personal experience) evidential -n, a direct evidence evidential , and an indirect evidence evidential -ke. The divide between the direct and indirect evidentials is quite straight forward. The direct evidential is used to indicate direct sensory evidence for the statement while the indirect evidential signals the lack of this. Thus, the direct evidential gets a lot of use when talking about visual evidence and the indirect evidential when reporting inference or hearsay. There's some fluidity between these two forms. If you are reporting some information based on a sound you have heard, you can use either the direct or the indirect evidential. The direct evidential is likewise some times used when talking about information contained in well known and trusted written sources. In both of these cases the choice of the direct evidential indicates a greater personal trust in the validity of the statement.

The egophoric evidential is used when the speaker has personal involvement in the action. This typically means that the speaker is either the subject or recipient in the clause or an affected object. Control matter that much that unconscious actions by a first person subject don't trigger egophoric marking but conscious unintentional actions do. Non-realised action (future or hypothetical events) also never get egophoric marking. As a defining characteristic for the egophoric, it's used in an anticipatory way in questions. Thus a typical pattern for indicative sentences is that the first person gets egophoric marking and the other persons don't,

Dźun sne-i-n.
SG1 leave-PST-EGO
"I left."

Ho sne-is.
SG2 leave-PST
"You left."

Qel sne-is.
SG3 leave-PST
"He left."

while in questions it's the second person that gets the egophoric marking and the other persons don't,

No dźun sne-is?
Q SG1 leave-PST
"Did I leave?"

No ho sne-i-n?
Q SG2 leave-PST-EGO
"Did you leave?"

No qel sne-is?
Q SG3 leave-PST
"Did he leave?"

Polar questions can be formed either by the interrogative particle no or simply with rising intonation. In the written language second person questions may thus be implied merely by the use of the egophoric marking. On the other hand, rhetoric questions, where no personal involvement by the second person is expected, are regularly formed with the direct evidential in all persons.

The egophoric evidential also gets used with certain third person subjects without any first person involvement. This happens with complement clauses where the third person is reporting their own personal experience and the same subject carries over from the main clause. Thus the following sentence has egophoric marking in the complement clause since the information contained within it stems from the third person's own experience

Qel sak-is qel Teni be rne-n.
SG3 say-PST SG3 Tenni ABL come-EGO
"He said that he is from Tenni."

While in the next sentence there is a break in the subject between the two clauses and no personal involvement about the complement clause can be assigned to either the speaker or the third person originating this information.

Qel sak-is ho Teni be rne-ke.
SG3 say-PST SG2 Tenni ABL come-INDIR
"He said that you are from Tenni."

In the direct evidential it's further possible to show that the speaker expects the audience to already be familiar with the information. In the earliest records of the language, this was done with an independent particle da that could also be used together with the egoporic. The particle quickly merged together with the verb into -de and became restricted to be used with only the direct evidential,

Qel Teni be rne-de.
SG3 Tenni ABL come-SHEARED
"He is from Tenni (as you surely know)."
ImageKištaLkal sikSeic

User avatar
Posts: 711
Joined: 07 Aug 2013 00:26
Location: displaced from Helsinki

Re: Lkal sik

Post by gach » 03 Apr 2016 17:57

Pronominal and nominal morphology

The personal pronouns of classical Lkal sik are dźun "I", ho "you.SG", qel "he/she" in the singular and nei "we.EXCL", ma "we.INCL", hei "you.PL", qele "they" in the plural.

Demonstratives include fa (proximal) and sie (distal) in singular and fao (proximal) and siou (distal) in plural. These are used both for deixis and anaphora but only the singular forms are used attributively,

sie nřu
DIST horse
"that horse"

sie nřu-gu
DIST horse-PL
"those horses"

The suffix -(g)u marks plurality but is only available for animate nouns. It's also not required for plural reference and never appears with plural quantifiers. The labial off-glide of the plural deictics is related to the -gu plural of nouns. The -e suffix on the PL3 pronoun qele could be so as well but is more likely related to a historical -i plural seen on the PL1.EXL and PL2 pronouns nei and hei. In the spoken language the -gu plural became defunct quite early on and got replaced by a new plural marker originating from kip, "many". The new plural is available for inanimates as well and in later times becomes the most common plural marker in the written language as well.

There are three morphological cases in classical Lkal sik: nominative , genitive -(i)s, and accusative -s. The genitive and accusative endings differ only by the vowel i in the genitive ending, and this gets often elided. The genitive ending has the form -is only after consonant final stems, where it contrasts either with an epenthetic e or no vowel at all in the accusative, and on stems where a final unstressed e is a part of the stem and gets replaces by the suffix vowel. In all other cases genitive and accusative are identical. As a result, the genitive and accusative endings also fall together on all animate plural nouns. Some noun paradigms for bźen, "woman", kame, "child", and nřu, "horse", are

Code: Select all

        SG      PL          SG      PL          SG     PL
NOM     bźen    bźen-u      kame    kam-u       nřu    nřu-gu
ACC     bzén-s  bźen-u-s    kame-s  kam-u-s     nřu-s  nřu-gu-s
GEN     bźen-is bźen-u-s    kam-is  kam-u-s     nřu-s  nřu-gu-s
Personal pronouns have some irregularities in their paradigms,

Code: Select all

      NOM             ACC              GEN
      SG     PL       SG     PL        SG     PL
1     dźun   nei      dźuns  neis      dźuns  neis
12     -     ma        -     mas        -     maes
2     ho     hei      hos    heis      hos    heis
3     qel    qele     qes    qeles     qeis   qelis
ImageKištaLkal sikSeic

User avatar
Posts: 711
Joined: 07 Aug 2013 00:26
Location: displaced from Helsinki

Re: Lkal sik

Post by gach » 20 Jan 2019 21:34

Due to work, I haven't found too much time or motivation for some time to be productive on the conlanging front (what you get for ending up in charge of three first author papers at the same time). I did, however, do a little more reading on egoforicity over the Christmas break, which made me reconsider the epistemic marking of Lkal sik. What better reason to dig up one of these old threads from the mothballs.

Epistemic marking reconsidered

The main structure of the Lkal sik epistemic marking system remains unchanged in that it's a unified paradigm of egoforic and non-egoforic evidential forms. The difference is that I decided to explore how a complex egoforic system with two separate egoforic forms would work. Such systems are found in the real world in some languages from the Tibetosphere (such as Bunan) and the Barbacoan languages of Ecuador and Columbia. When this separation into two separate egoforic forms occurs, one of them indicates privileged access to knowledge out of being an actively operating agent while the other one indicates privileged knowledge through being otherwise affected or even being linked to the action. Naturally there's quite a lot of room to play with the scope of especially the latter experiencer egoforic.

The four epistemic forms of Lkal sik, agent egoforic -n (EGO1), experiencer egofiric -me + ablaut (EGO2), sensory evidential (SENS), and indirect evidential -sik (INDIR), are tied together with the tenses. For some reason, the complex egoforic systems are only seen in the past tense forms in real world languages. I don't know why this would necessarily be so, but at the same time it feels like a nice distribution for the features. So instead of doing more research on the subject, I decided to go with it with the internal explanation that the Lkal sik EGO2 form is derived from an older perfect participle and has thus already comes with past tense reference by default. Here are a couple of example paradigms:

Code: Select all

            "eat"        "rain"       "fall"       "come"

PRS EGO1     brane-n       -          (ete-n)       dźo-n
    SENS     bran         lih          et           dźo
    INDIR    bran-sik     lih-sik      et-sik       dźo-sik

PST EGO1     bran-i-n      -          (et-i-n)      dźo-je-n
    EGO2     bru-me       liuh-me      out-me       dźou-me
    SENS     bran-is      lih-is       et-is        dźo-jes
    INDIR    bran-i-sik   lih-i-sik    et-i-sik     dźo-je-sik

FUT          bran         lih          et           dźo
I give a "future" tense here in addition to the present and past tenses even though it's not an independent tense of its own. Future reference is done using the present tense with the exception that all epistemic categories are neutralised in favour of an unmarked form. The EGO2 stem includes a rounding ablaut that's derived from an infix in the proto-language. I haven't even touched the details of this, though, so it will probable still go though some changes.

In the present tense the system is quite straight forward. In declarative sentences the EGO1 form is used when the speaker is an agent in control of the action. Hence, it's only applicable for first person reference. In the remaining cases the sensory and indirect evidentials are used depending on if the speaker has witnessed the action first hand or not.

Dźun brane-n.
SG1 eat-EGO1
"I'm eating."

Ho/qel bran.
SG2/SG3 eat
"You are eating." / "He/she is eating."

Dfam-u bran-sik.
guest-PL eat-INDIR
"The guests seem to be eating."

The one complexity is that if the speaker is not in control of the action, the EGO1 form is not applicable and has to be changed for the sensory evidential.

Dźun/ho/qel et.
SG1/SG2/SG3 fall
"I am falling." / "You are falling." / "He/she is falling."

In the past tense we also have the EGO2 form or the experiencer egoforic. In declarative speech this corresponds to first person subjects not in control of the action, first person direct objects of all types, as well as the speaker simply being physically affected by the action.

Dźun out-me.
SG1 fall-EGO2
"I fell."

Ho/qel et-is.
SG2/SG3 fall-PST
"You fell." / "He/she fell."

Nřu dźun-s sout-me.
horse SG1-ACC drop-EGO2
"The horse dropped me."

Hgem liuh-me.
morning rain-EGO2
"It rained in the morning." (It soaked me.)

It is possible to extend the scope of the EGO2 form to cover first person recipients and possessors or even less direct connections to the speaker but I've decided not to include these under the Lkal sik EGO2 marking.

The hallmark feature of egoforicity is that the egoforic form related to the first person in declarative speech is related to the second person in questions. What this comes down to in Lkal sik is that in questions you always anticipate the epistemic status of the answer from the addressee's perspective. Hence, typically questions follow the pattern of marking second person subjects with EGO1 and otherwise using the sensory evidential,

Da ho bran-i-n?
Q SG2 eat-PST-EGO1
"Did you eat?"

Da qel bran-is?
Q SG3 eat-PST
"Did he/she eat?"

but often it's also possible to make more fine grained distinctions.

Da lih-is hgem?
Q rain-PST morning
"Did it rain in the morning?" (Did you witness it?)

Da liuh-me hgem?
Q rain-EGO2 morning
"Did it rain in the morning?" (You've got yourself wet.)

Da lih-i-sik hgem?
Q rain-PST-INDIR morning
"Did it rain in the morning?" (Is there any indication of it?)

A similar shift of stance happens in reported speech, also prototypical for egoforic systems. Here egoforic marking is used from the perspective of the speaker of the reported clause. As a result, it often doubles as switch reference marking between the matrix clause and the reported clause.

Qel sak-is qel bran-i-n.
SG3 say-PST SG3 eat-PST-EGO1
"He/she said that he/she has eaten."

Qel sak-is dfam-u bran-is.
SG3 say-PST guest-PL eat-PST
"He/she said that the guests have eaten."
Last edited by gach on 21 Jan 2019 02:41, edited 1 time in total.
ImageKištaLkal sikSeic

User avatar
Posts: 1768
Joined: 11 Feb 2015 11:23

Re: Lkal sik

Post by gestaltist » 20 Jan 2019 22:29

I'm glad you wrote this post. I've never seen this thread before and I find it quite interesting. My current project (Īsmay) was originally going to have egoforicity but I wasn't familiar with the term and I couldn't put my finger on how exactly I want it to work or what I want, exactly, so I dropped it. I wish I'd read this a month ago...

I like the additional egoforic "past" marker you've added. Regarding glossing, though, EGO1 and EGO2 sounds a bit confusing. Why not use something like EGO.A (agentive) and EGO.E (experiential) instead? It would be easier to parse, I feel.

User avatar
Posts: 711
Joined: 07 Aug 2013 00:26
Location: displaced from Helsinki

Re: Lkal sik

Post by gach » 20 Jan 2019 22:58

Glad that you like it. Egoforicity is one of those things which takes time before you start to understand it intuitively, especially in relation with classical evidentiality. Hence also the differences between my old and new posts. If you still wish to do reading for a future project using egoforicity, this category is probably the best place to start,
gestaltist wrote:
20 Jan 2019 22:29
I like the additional egoforic "past" marker you've added. Regarding glossing, though, EGO1 and EGO2 sounds a bit confusing. Why not use something like EGO.A (agentive) and EGO.E (experiential) instead? It would be easier to parse, I feel.
Those certainly are clearer glosses and I was considering something similar when I was figuring out my terminology. In the end I went with these simple non-informative glosses that I'd seen used in the literature. Luckily it's a pretty simple task to change glossing conventions.
ImageKištaLkal sikSeic

User avatar
Posts: 711
Joined: 07 Aug 2013 00:26
Location: displaced from Helsinki

Re: Lkal sik

Post by gach » 23 Jan 2019 01:14

Another thing I've been thinking of is that despite paying a lot of attention to dealing with information structure in conlanging, I've never actually started a topic prominent language. This is an obvious gap that needs to be filled and I've chosen to use Lkal sik for the purpose.

As you may notice, there's nothing pointing to topic prominence in the previous examples. I intend this to remain the case in the older stages of the classical language. Instead, the topic prominence will be a feature that evolves over time and reaches its full extent in the later, or more modern, stages of the language.

Initially at least some fronted topicalised arguments were marked with the postposition ta (meaning undecided, probably some generic locative). This quickly agglutinated in the spoken language, loosing the final vowel beyond the second syllable. At this point you essentially had a case like topic form parallel to the three original cases: nominative, accusative, and genitive. The written language retained a more conservative grammar, however, and took some time to accept the morphologised topic marking. The topic forms are used in the written language of the later classical period but a lot more sparingly than in the spoken language of the era. Most notably, subject topic is not marked.

Over time the usage of the classical language broke down and eventually people started to write the spoken language of their own time with its fully developed topic prominent grammar. I assume that the evolution of written Lkal sik will then be dominated by a classical norm with only limited topic marking and a later developing modern standard that matches the topic marking of the spoken language. I'm interested, though, by what will happen during the intermediary period where people preferred to use the classical standard but increasingly slipped in colloquial topic marking.

Now, I'm not aware how much documented evidence there is for how topic prominent grammar develops and evolves over time. If anyone knows any relevant literature on the subject, I'd be interested to know. This would be useful for figuring out what will the earliest stages of the topic marking grammar most likely look like and what sort of grammatical errors would we expect from sloppy writers during different stages of the language's evolution.

Lastly, here are some example topic forms from Classical Lkal sik: nřu (horse) ~ nřu-te (horse-TOP), kame (child) ~ kame-t (child-TOP), diom (man) ~ diom-et (man-TOP), bźen (woman) ~ bźen-et; and the corresponding forms in the modern langauge: qu ~ qùt, kàm ~ káma, dźom ~ dźóme, źe ~ źéne. In the orthography for the modern language <q> stands for /χ/ derived from /q ʁ/, the vowels all have their IPA values, and the diacritics denote word level tones: acute for high or high rising, grave for falling, and unmarked for low.
ImageKištaLkal sikSeic

Post Reply