Conlang Relay 18/Sabasasaj
Aahagansis tuam tumiguduata gi tatagauman hitabi kuiis bihtau wudinsiuwas piugiuaan sidumindaba khal auuliil mia thiuni.
Haam takuiita gi agaalidih thaw hjuuiisi phuibi. Hwusuu ahtadih sarpi mu ihiirkur ha bihtau. Tatagauman gi ubariu jati khadu taratiga ulawpi, wus ulas pulti naiphaundawsim, us wus tjidais iaa siahir ba thawkul. Tuam gi ikhaumu hirju gia thitumuw, iun a hajdala im huala hwiakansajta naitia siwiail tatasaa khaaah dauiaiam tihismi. Indarba danu thiuwisma sasaihhu khai thiugua sabakuirissim im hwiakuita dauitah ha piumtanah.
A smooth rendering
May it please you that I might tell you how the statue atop a corner of the villa came to life.
There was a fox, sleeping in a man's arms. Suddenly, the fox awoke and began to cry out. A few clouds of smoke were converging on the statue, which then coalesced and began to whirl around it, but dissipated a short time thereafter. We then saw atop the villa a woman, covered in snow; her face appeared as smooth as apple skin, beautiful and unforgiving. However, she did not know the habits of our people; she only smiled and wiped her face with her sleeve.
Stops in the coda lenite, p(h) t(h) k(h) b d g > s s h w r j.
The first of two hs separated by one syllable nucleus deletes; this includes hs in ph th kh.
s + h yields hh.
Stops p t k voice to b d g between vowels, or between a nasal and a vowel, in some morphological circumstances. As a rule of thumb (completely accurate in this text?), this process is in force from the right edge of stem part 1 to the left edge of stem part 2 (see below), inclusive, except that t in an agreement marker doesn't voice.
There are also some morphemes with their own allomorphy; if the form I cite has parens, then the material in parens is only introduced to avoid clusters of more than two Cs.
Morphology and syntax in overview
Sabasasaj is AdvSOV and head-final. Adpositions are postpositions.
It's verb-happy, with weak morphosyntactic distinctions between noun and nonfinite verb and finite verb. The most preferred arrangement of a Sabasasaj sentence is a single long unbranching chain of subclauses, give or take some stubby tendrils.
The boundary between S and O, both of which may be heavy phrases, is mostly cued by prosody. In the text I have used comma (not exclusively) for important cases of this boundary (which is not in accord with standard Western punctuation rules).
Most underived Sabasasaj words are verbs, and all verbs have bipartite stems, with two components (of typical length one syllable each).
The verb is structured as below. All the non-stem pieces may be empty. Its two basic TA categories are progressive and perfect, and these have slightly different formation.
- body: [preverb]-[iteratives]-[stem part 1]-[directional]-[agreement 1]-[agreement 2]
- progressive: [body]-( [negation])-[stem part 2]
- perfect: ([perfect marker]-XOR-[negation])-[stem part 2] [body]
The spacing is intentional. Negation splits the progressive into two phonological words. The perfect also consists of two phonological words, and a few things can intervene between them. To be more precise, the order is
- first word of the perfect, if there is one
- mood: this text has irrealis khal and volitive mia
- some pronoun-like elements, including ha, the argument 1 form of a pronoun referring to a topical 3rd person entity, and a & u re which see below
- verbal word, containing stem part 1
The negative morph is kha.
nai particulate matter; solid which doesn't hold its shape phu animal saba mouth thi woman
Preverbs mostly serve to classify an argument of the verb (this is required in a few constructions and useful in a few others), or behave similarly to a semantically weak incorporated argument.
The iterative, which can have a gnomic sense, is formed by reduplicating the first CV of the verbal word.
Directionals (at least the compositional ones)
aal between, among hi on the (top) surface of i outward; can have the force of a 2nd person goal/recipient ri upward s(a) inward ul coming together um along the surface of
Sabasasaj gets away with a very few adpositions by having directionals carry most of the semantic load, but an adposition are still necessary if one wants to introduce an argument for a directional. The adposition used for this is gi.
hir having gone onto sth.; can be used for experiential perfects i (the sematically empty perfect marker) ul having come together
Verbs agree with two core arguments, which I've just called 1 and 2 here. The only valences verbs may have are intransitive or transitive; that is, either one or two agreement markers must appear. If a verb is lexically transitive, it may freely intransitivise by dropping either agreement marker. However, actual argument phrases need never appear (a verb is a complete clause).
Approximately, the argument that gets agreement 1 is more controlling than the one that gets agreement 2, but the choice is lexical. In particular, intransitives have split-S marking.
Here are the agreement markers that occur here.
agt. 1 agt. 2 (i)n n 1 sing m 1 plur li 2 th 3 sing human s 3 plur human k i 3 sing animate (but nonhuman) u d(a) 3 sing inanimate (a)w 3 plur inanimate a ta relative sing i ti relative plur
Actually, the relative markers use agreement 1 forms even in agreement position 2, unless agreement position 1 is empty.
Also exceptionally, the 3rd singular inanimate in agreement 2 appears with a t, not a d, if agreement 1 holds a 3rd person animate (or human) agreement marker. If in this event agreement 1 holds 3rd person human singular, it is dropped outright.
When a finite clause is an argument of another verb and it's not the relativisation construction, below, the verb uses inanimate agreement. (Though watch out for a topic-introducing cleft construction, where essentially this happens under the verb "exist" and the subclause does not take agreement.)
Mass nouns tend to take plural agreement.
The agreement markers I have labelled "relative" mean that the argument in question corefers with an argument of the verb to which this one is subordinate. The syntax doesn't distinguish which argument of the higher verb it is. However, relative verbs never refer to a speech act participant argument, or one marked with a topic pronoun, unless they're followed by the particle mu.
Relative verbs generally can't modify a noun. (Structures where you might want to have a relative verb modifying a noun instead put the verb on the outside; this has a sort of internally headed relative clause vibe. Sometimes these can even be nested!)
There are two pronoun-like elements, a and u, which have similar relativisation semantics. They're more powerful, though; they can corefer to a verb higher than one clause above this one in the syntax. In fact, they always do, unless they have a prefixed adposition, like gia (~= "in/at/to/from/on/... which"). a is used if the argument of the higher clause is argument 1, u if it's argument 2.
Stuff for the noun
Nouns don't mark much. All we get here is number, which I've just left in the lexicon in the one place it appears, and possessor markers.
ma 1 plur possessor ta 3 sing possessor
Unfortunately, some nouns take possessor markers (for the present purposes) infixally. These don't show the final a above.
This lexicon repeats most of the particles &c; several are given some more explanation above.
Many of the forms listed here are derived with more or less noncompositional semantics (to do otherwise would be pretty confusing).
In a verb listed x-y, x is [stem part 1] and y is [stem part 2]. In the glosses, 1 and 2 refer to the core arguments, and A to a non-core argument expressed in an adpositional phrase.
If a noun is listed containing ^, that means it takes its possessor marking infixally at that point.
-pi 2 exists, goes -- yes, stem part 1 is empty! This verb always has either a preverb or a directional, though. a -- arg 1 relative pronoun aahagansis a. tuam: villa, ~manor house ak-tih 2 sleeps -- note the metaphor "X's sleeping goes in" = 'X wakes up' au-il 2 finds 1 favourable / pleasing ba (adpos.) at (the time), under (the circumstances) bi-u 1 begins to do 2 -- note that arg 1 is agreed with again in the subclause. also, watch out for allomorphy irregularities near the -u danu only daui^ah face daui-am 1 puts 2 forward -- literally, though be mindful of the evident derivational relationship with 'face' gi (adpos.) -- semantically empty gia = gi + a (though you'd expect the other order) ha -- more or less a pronoun: argument 1 of this clause is current topical entity, established in a previous sentence haam man = male person hajdal-un (in the perfect) 2 is smooth hiir-ur 1 shrieks hjuuiisi thaw h.: fox huala apple hwiakansaj hide, skin hwiakui^a sleeve hwusuu all of a sudden ikhaumu fallen snow (mass) im and indarba however ja-tu 2 are many khal -- irrealis particle ku-is 2 is alive kui-sim 1 bends 2 (along its whole length) mia -- volitive particle mu -- used in some relativisations na-tia 1 are similar to each other pi-nah 1 drags, pulls 2 piugi-an 2 pertains to, is about 1 pul-as 2 comes together, unites, coalesces phaunda-sim 2 rotates about an axis sai-hu 1 does 2 siahi-aa 2 extends to, reaches its end sidum-pa 1 narrates 2 to A siwi-il 1 is beautiful takui^i plur. of takui^a arm tara-ka 2 is distributed as clouds tasa-aah 1 shows mercy to, forgives 2 tatagauman statue tjit-is 1 is small tuam house tumigudua corner (of a structure) tumu-ju 1 covers A with 2 tha-kul 2 fades, vanishes, dissipates thaw red thi-i 2 sees 1 thiuk-i 2 learns 1 thiuwis people (mass) u -- arg 2 relative pronoun ubariu smoke (mass) us but wudinsiuwas way (in which sth. is done), method wus and then