Laashu - The Five Mutations

A forum for all topics related to constructed languages
Post Reply
Porphyrogenitos
sinic
sinic
Posts: 313
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Laashu - The Five Mutations

Post by Porphyrogenitos »

This is an idea I have been puzzling over for a week or two, and I am finally ready to present an initial sketch of it.

In Bloomfield's Language (1933), he presents a fragmentary sketch of Chinese, stating that Chinese can be described largely in terms of just three constructions. I decided I wanted to create a conlang in which all grammatical structures could be described in just those terms - in which those were the only structures available in the language. Although, I ended up adding one more. They are:
  • Subject - predicate
  • Action - goal
  • Construct - modifier
  • Topic - comment
I was pulled in several directions by this. I was also influenced by Lojban and especially Toaq, and was leaning towards making it an engelang with unambiguous syntax. Unambiguous syntax ended up being too hard to implement, so I allowed the language to have a more naturalistic syntax where multiple interpretations and garden paths are possible. But I kept the engelang concept of self-segregating morophology - i.e. unambiguous morpheme boundaries.

I decided the constructions would be head-marked - counting the subject, action, construct, and topic as the head of each respective construction. The marking was initially a particle placed after the end of the phrase, but then I decided it would fuse with the preceding word, in the manner of a clitic.

Another thing I wanted to do was to have monosyllabic roots, like Old Chinese or Toaq. I really really liked the concision of monosyllabic roots. And I wanted them to stay monosyllabic even when inflected with a phrase clitic - but how to do it? I seriously considered suffixes, which weren't too bad, but raised some mild issues with self-segregating morphology. The most elegant and aesthetically pleasing solution, which would give the greatest freedom to root shape within the monosyllabic template, would be to implement the clitics tonally - like the grammatical tones in Toaq. One problem: I prefer to be able to pronounce my conlangs, and while I can sometimes dispense with that, I really wanted to be able to pronounce this one. And I'm not so good with tones. So, I ended up deciding on something else.

The grammatical "clitics" - by all appearances, an ablauting root vowel - consist of the rhyme of each monosyllabic root. There are fifteen rhymes in this language, but only three lexically contrastive categories of rhymes. That is, each root has five forms - the unmarked or citation form, which has an important role of its own in the grammar, and the subject, action, construct, and topic forms.

Early on, before I had even decided on the phonology of the language, I decided that the name of the language would be of the form change-CONST five, i.e. 'five changes', referring, naturally, to the five forms of each word. I have not decided on many lexical items yet, just enough to provide some basic examples, but I've decided on the relevant roots for the name of the language: laa shu, 'the five changes', or perhaps more artistically, or more linguistically, 'the five alternations', 'the five mutations', or even 'the five desinences'. Technically, this is a metonymy - referring to the language's most characteristic feature, instead of the language itself. A fuller name might be dle laa shu, 'the language of the five mutations', lit. 'the tongue of the five mutations'. In the language's roman orthography I intend to write each syllable/root separated by a space, but in English usage Laashu would be fine. The doubled <a> has orthographic import in the native orthography, to be explained below, but to an English speaker it could just as well be "Lashu", though for the sake of consistency Laashu would be better.

Next, I will explain the language's phonology, including the shape of roots and the system of alternations.
Last edited by Porphyrogenitos on 13 Jan 2021 05:41, edited 1 time in total.
Porphyrogenitos
sinic
sinic
Posts: 313
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Laashu - The Five Mutations

Post by Porphyrogenitos »

Phonology

Laashu has a maximum syllable structure of CCCVN, or more specifically, sCRVŋ, where R is any of /w j r l/ and V may be a diphthong. However, it is more helpful to think of it in terms of onsets and rhymes. Different consonants may combine in different ways to produce different onsets; not all combinations are possible. Rhymes basically behave as units. The consonants are:

/m n ŋ/ m n ng
/p t k/ p t k
/b d g/ b d g
/ts tʃ/ ts ch
/dz dʒ/ dz j
/f s ʃ x/ f s sh kh
/v z ʒ/ v z zh
/l r/ l r
/w j/ w y

An empty onset may be orthographically represented as h. (This dates from when I was intending to have roots be onset-obligatory, with an epenthetic [h] always inserted. But now I think that the epenthetic [h] will only occur intervocalially, i.e. word-medially, and may also be substituted by a glottal stop.) Note that /ŋ/ only occurs in the coda.

The vowels are:

/i u/ i u
/e ə o/ e a o
/ɛ ɔ/ ae ao
/a/ aa

/iə̯ uə̯/ ia ua
/ai̯ au̯/ ai au

/ɛ ɔ/ may vary between [ɛ ɔ] ~ [æ ɒ].

This accounts for 12 of the 15 possible rhymes. The remainder are formed with /ŋ/:

/iŋ aŋ uŋ/ ing ang ung

These may just as well be pronounced as nasal vowels.

I may post the whole inventory of possible onsets sometime later. I will note that the voicing distinction in stops and affricates is neutralized after /s/, and that onset /w/ may not precede a back vowel and onset /j/ may not precede a front vowel.
Last edited by Porphyrogenitos on 13 Jan 2021 05:42, edited 2 times in total.
Porphyrogenitos
sinic
sinic
Posts: 313
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Laashu - The Five Mutations

Post by Porphyrogenitos »

Morphophonology

The syllable rhyme encodes lexical and grammatical information in Laashu - more grammatical than lexical, in fact. I'm not all that pleased that more lexical contrasts aren't possible, but this was basically my compromise system that I came up with after abandoning tone and suffixes - although these still sort of are suffixes?

The only vowels that appear in the unmarked (base, uninflected) form of lexical items are /i ə u/. Each one inflects for grammatical categories in different ways.

I-stems
Unmarked: -i
Subject: -ae
Action: -ing
Construct: -e
Topic: -ia

U-stems
Unmarked: -u
Subject: -ao
Action: -ung
Construct: -o
Topic: -ua

A-stems
Unmarked: -a
Subject: -ai
Action: -ang
Construct: -aa
Topic: -au

I had actually thought about perhaps providing one derivational/classificatory operation built into the monosyllable, e.g. allowing for a masculine-feminine distinction or something, but that didn't end up happening.

As it stands, this is the entirety of bound morphology in Laashu - though perhaps various syntactic constructions may be morphologized.

Stem shapes

All native Laashu stems/morphemes are strictly monosyllabic. Proper names and other terms may be loaned in as multisyllabic items, subject to the restriction that only the basic three vowels /i ə u/ may appear in the uninflected word: English America amirika

However, there is a very strong preference to nativize these loans by reducing them to a monosyllable: amirikami. These shortened proper names are usually situated in context, so homophony is not necessarily an issue: ko mi 'America', lit. 'country of America', 'America-land'. With geographic names, two proper names may be often be used in conjunction, much like the English construction "Columbus, Ohio" or "Dublin, Ireland": paa fra 'Paris, France', lit. 'Paris of France'. (note that the preceding two examples are construct-modifier constructions)

The syllable and root structure of Laashu does raise questions about homophony, however. At the current count, there are only 219 possible (lexically-constrastive) syllables in the language (i.e. not counting rhyme distinctions that are only used for grammatical inflection). I don't recall the actual numbers on this, but I think it's often bandied about that the typical language has around 3000 commonly-used roots. (Or was it 800? 2000?) In any case, this is clearly far less than that.

I am somewhat influenced by the minimalism of toki pona, but I don't want to go quite as far as toki pona in lexical impoverishment. There are just 120 words in toki pona, so at least we can do better than that. According to a Quora answer, the 6000 most commonly used Chinese characters account for 372 different syllables, ignoring tone. Obviously it isn't fair to ignore tone, but my understanding is that Mandarin can be understood well enough in context, usually, even if someone completely butchers the tones. (Probably the same with English vowels.)

In any case, some homophony will be necessary. If each syllable is given 2 meanings, that's 438 roots, which isn't completely horrible. Compounding will probably play a big role in a lot of relatively basic vocabulary anyways.

I may also try to expand the number of possible onsets (though I have a countervailing impulse to simplify the consonant and onset inventory) or even the number of rhymes. But I won't mess around with that more for the time being.

Word classes

In Bloomfield's sketch of Chinese that inspired this, he said Chinese only had two primary word classes: content words and particles, with content words being used freely in the different constructions as subjects, predicates, modifiers, etc.

Well, as it happens, there is really only one top-level word class in Lashu - I was initially going to have particles, either they got fused into the words as inflections, or I eliminated them as unnecessary. But, you could say - considering the four (or five?) inflected forms as morphemes - that there are two classes of morphemes in Laashu, one free and lexical and one bound and grammatical.

In general, Laashu lexical items are free to fill any slot in any construction. Of course, there are likely to be distinctions within this one big mono-word-class, even if they only emerge at the statistical level: ma 'mother' is surely less likely to fill the action slot than tlu 'hit'. What specific structural consequences this will have remains to be seen.

Although of course I have many things already planned out ahead, I don't intend this as an encyclopedic portrait of a fully-formed natlang set in a fictional word - at least at this point, I have no plans to set Laashu in a fictional world. It is more of an experimental language. What this means is that I would be happy for some issues and features in the language to be fleshed out and decided by usage - the approach taken by Lojbanists and tokiponists, i.e. to consult usage (including and perhaps especially the creator's usage) when figuring out the best/right way to do something.

There is, of course, no body of Laashu usage yet - there is hardly even a lexicon. Basically, I will be setting down some ground rules, but this will be an evolving project. But I suppose that's true of anyone's conlang?

More on how the "four (or five??) constructions" actually work tomorrow...
User avatar
Omzinesý
runic
runic
Posts: 2932
Joined: 27 Aug 2010 08:17
Location: nowhere [naʊhɪɚ]

Re: Laashu - The Five Mutations

Post by Omzinesý »

Porphyrogenitos wrote: 13 Jan 2021 05:01
In Bloomfield's Language (1933),
Is this a paper in journal "Language" or is Bloomfield's Language a conlang or is it a term I shoul know?

Sorry this is a stupid question, but I am interested to check how those constructions are defined.
User avatar
WeepingElf
sinic
sinic
Posts: 222
Joined: 23 Feb 2016 18:42
Location: Braunschweig, Germany
Contact:

Re: Laashu - The Five Mutations

Post by WeepingElf »

AFAIK, Language is the title of classic linguistics textbook by Leonard Bloomfield, and I guess that is what Porphyrogenitos meant.
... brought to you by the Weeping Elf
User avatar
Omzinesý
runic
runic
Posts: 2932
Joined: 27 Aug 2010 08:17
Location: nowhere [naʊhɪɚ]

Re: Laashu - The Five Mutations

Post by Omzinesý »

WeepingElf wrote: 13 Jan 2021 16:16 AFAIK, Language is the title of classic linguistics textbook by Leonard Bloomfield, and I guess that is what Porphyrogenitos meant.
OK, I see.
In the category "term I should know".
Porphyrogenitos
sinic
sinic
Posts: 313
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Laashu - The Five Mutations

Post by Porphyrogenitos »

Omzinesý wrote: 13 Jan 2021 12:14
Porphyrogenitos wrote: 13 Jan 2021 05:01
In Bloomfield's Language (1933),
Is this a paper in journal "Language" or is Bloomfield's Language a conlang or is it a term I shoul know?

Sorry this is a stupid question, but I am interested to check how those constructions are defined.
WeepingElf wrote: 13 Jan 2021 16:16 AFAIK, Language is the title of classic linguistics textbook by Leonard Bloomfield, and I guess that is what Porphyrogenitos meant.
Yes, thank you. Here is a pdf of it; it's a great book, very relevant to conlangers. I think conlangers in particular really tend to discuss language in Bloomfield's structuralist idiom - I mean, linguistics as we know it wouldn't exist without the structuralism of Sapir, Whorf, Bloomfield, etc, but the later innovations of generativism and functionalism don't seem to have touched mainstream "naturalistic" conlanging so much.

I'll provide screenshots of the passage that inspired this conlang. Bloomfield discusses Chinese in the context of word classes. He explains that "parts of speech" and basic construction types are language-specific, and that Indo-European is somewhat unusual for having many very clearly differentiated word classes:

Image

Image

So, the three constructions Bloomfield gives for Chinese are:
  • subject-predicate
  • attribute-head
  • action-goal or relation-axis
In Laashu, I flipped the order of attribute-head (and renamed "head" to "construct", after the Semitic construct state) and added a fourth construction, topic-comment. Also, in Laashu all constructions (specifically, their heads) receive a marker, but it is phonologically fused to the preceding word.

In a moment I'll get to how the four/five constructions actually work in Laashu.
Porphyrogenitos
sinic
sinic
Posts: 313
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Laashu - The Five Mutations

Post by Porphyrogenitos »

Syntax: Basic constructions

I have vacillated between saying there are four and five constructions in Laashu. The subject-predicate, action-goal, construct-modifier, and topic-comment constructions are all explicitly marked on their respective heads. The absence of one of these markers can be just that - an absence, which serves as the citation form of a word, which would be used in a list or as a vocative. But two unmarked words next to each other - parataxis - also serves as significant role of its own, which I will describe further below.

I will be using the following abbreviations:
  • subject - S
  • action - A
  • construct - C
  • topic - T
Subject - predicate

The subject-predicate construction basically corresponds to an English subject+verb phrase, including copulae joining NPs to adjectives, other NPs, participles, etc.

Here are some basic examples:

mai kri
mother-S old
'mother is old'

mai hyu
mother-S sleep
'mother is sleeping'

snyai pa
man-S father
'a/the man is a father'

grae sta
house-S big
'a/the house is big'

The citation forms of 'mother', 'man', and 'house' are ma, snya, and gri, respectively. They take the subject form to mark them as the head of a subject-predicate construction: You know that anything that follows them must be a predicate.

Constructions themselves may partake in other constructions, either as heads or as dependents. The Laashu equivalent of a transitive verb involves an action-goal construction forming the predicate in a subject-predicate construction:

snyai tlung sku
man-S hit-A pet
'a/the man hits a/the pet/dog'

The inflection of 'man' signals that what follows it is a predicate, and the inflection of 'hit' signals that what follows it is a goal. The goal, sku 'pet, dog, domestic quadruped' has no dependent - and so, not being the head of any construction, is unmarked.

An action-goal construction can also serve as the head of a subject-predicate construction:

tlung skao mla
hit-A pet-S bad
'hitting pets is bad', 'to hit a dog is bad'

This clearly demonstrates the clitic character of the inflections - they affect the last word in each phrase. 'hit' is the head of an action-goal phrase, so it takes the action inflection; the phrase 'hit pet' as a whole is the subject of a subject-predicate phrase, so only the last word, 'pet', takes the subject marking. The predicate of the subject-predicate phrase, 'bad', takes no marking.

Now, wait - couldn't the above sentence also be interpreted with 'pet-S bad' as a subject-predicate construction on its own, serving as the goal of an action-goal construction headed by 'hit'? Yes, it could. That is where the natlang character of Laashu comes in - it is not syntactically unambiguous.

However, much of the time, context should be enough to disambiguate between rival parsings. In this case, the aforementioned alternative parsing has an incoherent or infelicitous interpretation - the action is 'hit', and the goal is 'the/a dog is bad', so...I'm not even sure what that would mean? Something like '[he, someone] hits [the fact that] a dog is bad' - it can't be '[someone] hits a bad dog', 'bad dog' would have to be a construct-modifier construction. So in this case, at least, we're saved from ambiguity by the nonsensical nature of an alternative reading.

I think I will pause for a moment here before moving on to the action-goal construction, about which there is more to say.
Porphyrogenitos
sinic
sinic
Posts: 313
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Laashu - The Five Mutations

Post by Porphyrogenitos »

Syntax: Basic constructions - continued

Action-goal

The action-goal construction is quite multifunctional, and is used to express things that would be expressed in English with prepositional phrases and verb phrases, including transitive verbs and their objects, as well as motion verbs.

Action-goal constructions that express an action and a patient do not have to occur within a subject-predicate construction, as in the examples in the previous post; there may be no agent or the agent may be left implicit:

tlung snya
hit-A man
'[someone] hit a/the man', '[he] hit a/the man', 'a/the man was hit'

Though, as also seen in the previous post, in the appropriate context an action-goal construction may also be interpreted as a "nominal" or "nominalized form" - kind of meaningless to say, since there are no "nouns", per se, in the language - in this case, the phrase tlung snya could be interpreted as 'to hit a man' or 'that [someone, he] hit a man', or perhaps 'hitting men' (note that there is no number distinction!)

Okay, now for motion. Action-goal constructions basically express everything motion-related in Laashu, taking on the role of both motion verbs and prepositions. An example:

sking gri
go.into-A house
'into a/the house', '[someone/he] went into the house'

On its own, this could be interpreted with an implicit agent, like with tlung snya. As the predicate of a subject-predicate construction, it basically behaves like the English verb enter:

snyai sking gri
man-S go.into-A house
'a/the man entered the house'

If an action-goal construction serves as the modifier of another phrase, it takes on a prepositional or adverbial sense:

mai hyo king gri
mother-S sleep-C be.in-A house
'mother is sleeping in the house'

Less idiomatically, this might be rendered 'Mother is sleeping, being in the house'.

Action-goal constructions may themselves be modified in an adverbial manner. This is the main way to express manner of motion:

snyai sking gre gla
man-S go.in-A house-C run
'a/the man entered the house running', 'a/the man ran into the house'

The 'goal' in an action-goal construction may also be a phrase/clause:

mai bing hyo king gri
mother-S love-A sleep-C be.in-A house
'mother loves to sleep in the house'

So, the term 'action-goal' is kind of misleading, since "love" isn't really an "action" and "sleeping in the house" isn't really a "goal" here. But I'm not sure what a better label would be.

I also might think of a different way to express "love" and other attitudes. The above seems too straightforwardly English-like. Maybe this?

mai shing be hyo king gri
mother-S have-A love-C sleep-C be.in-A house
'mother loves to sleep in the house', literally 'mother has love of sleeping in the house'

I think I like that better. Though both constructions can coexist in the language, of course.
Porphyrogenitos
sinic
sinic
Posts: 313
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Laashu - The Five Mutations

Post by Porphyrogenitos »

Syntax: Basic constructions - continued

Construct - modifier

Okay, so...the construct-modifier construction is perhaps the most multifunctional construction in Laashu. So I'm not sure if I'll get to everything in this post.

As we saw above, it can be used in a manner akin to an adverb or prepositional phrase:

ao glaa byu
3-S run-C fast
'he/she/they run fast'

hrae zhe gla
woman-S go.out-C run
'a/the woman came out running'

mai shing gre king hya
mother-S have-A house-C be.in-A settlement
'mother has a house in the city/town'

It also plays the role of noun+adjective in English. Contrast the following subject-predicate construction...

snyai kri
man-S old
'a/the man is old'

...with the following construct-modifier construction:

snyaa kri
man-C old
'a/the old man'

Construct-modifier is also used for possession:

gre snya
house-C man
'a/the man's house'

ko u
country-C 3
'his/her/their country'

kre snya
old-C man
'the elderliness/old age of the man'

And it also plays the role of a relative clause - all of the above "adjective" phrases could also be translated as relative clauses: 'the house that is of the man', 'the country that is of him/her/them', 'the old age that is of the man'.

snyaa tlung sku
man-C hit-A dog
'a/the man who hits a dog'

This could equally be translated as 'the man of hitting the dog'.

sko sting snya
dog-C bite-A man
'a/the dog that bites a/the man'

There is more to be said about relative clauses in Laashu - in general, objects and other arguments are not as accessible to relativization as they are in English. But that will be for another post.
Porphyrogenitos
sinic
sinic
Posts: 313
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Laashu - The Five Mutations

Post by Porphyrogenitos »

Syntax: Basic constructions - continued

Topic - comment

I will briefly discuss the Laashu topic-comment construction. It's pretty much much it says on the tin, it's basically the same as topic-comment constructions in many East Asian languages.

I decided to add it to the other three constructions in part because there didn't seem to be a great way to front material - I'm sure there could be a way, but I haven't thought of how that might work yet, and in any case it would probably be wordier than the topic-comment construction. I also didn't want to use bare parataxis, I wanted it to be explicitly encoded in syntax.

Some simple topic-comment constructions would be things like:

skua sta
dog-T big
'as for the dog, it's big'

mau re mla
mother-T sibling-C bad
'as for mother, she's a bad sister'

snyau gre li
man-T house-C two
'as for the man, there are two houses'

The first two of these could express essentially an identical meaning with a subject-predicate construction. In cases like that it's really just a matter of emphasis and information structure.

The topic-comment construction can also be used to increase the prominence of a semantic patient, thus creating a sort of "passive":

skua snyai tlu
dog-T man-S hit
'as for the dog, the man hit [it]'

However, there is a more formal syntactic passive in Laashu which I will discuss later.

Topic-comment constructions are usually the matrix for a whole sentence, but they can sometimes be embedded:

ao string snyau pai mla
3-S believe-A man-T father-S bad
'he thinks that, as far as men go, father is bad [a bad one]'

Something like that. That example seems a bit contrived. But that sort of embedding of topic-comment constructions would, I imagine, be more common in formal written language.
User avatar
gestaltist
mayan
mayan
Posts: 1597
Joined: 11 Feb 2015 11:23

Re: Laashu - The Five Mutations

Post by gestaltist »

Quite a cool idea. Gave me food for thought. :)
Porphyrogenitos
sinic
sinic
Posts: 313
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Laashu - The Five Mutations

Post by Porphyrogenitos »

gestaltist wrote: 16 Jan 2021 23:14 Quite a cool idea. Gave me food for thought. :)
Thank you!

If I post more about grammar, I think I'll talk about parataxis (i.e. the "unmarked construction") or the passive construction (itself built out of the basic five constructions).

But I'm feeling hesitant about the phonology. It's very unaesthetic, especially with the orthography I'm using. I might relent and switch to a tone-based system. I was also thinking at one point of having a tonal system exist side-by-side with an alternative suffixal system (perhaps the putative sources of tonogenesis). But certain suffixes could create phonological ambiguity - for example, [maʔ sa] [maʔ tsa] and [ma tsa] could be difficult to distinguish from each other. And it is important to be able to include thing like affricates as possible onsets, since I need to make sure there are a good amount of possible roots, at least 200 or so.

I'm currently thinking of the following system:

Unmarked: Low a
Subject: Falling à
Action: Nasal ã or an
Construct: High ā
Topic: Rising á

There would be just 5 vowels, maybe a couple of diphthongs, so there would need to be at least 40 onsets to get a good amount of root forms.
Post Reply