Otseqon

Otseqon (/oʊˈtseɪkɑn/; Otseqon: oceqon [očeqoŋ̇]) is a cluster of 5 dialects spoken by the Otseqon people living on 5 city-sized boats.

Phonology
Dialects differ somewhat in their phonology. Described here is the standard dialect, spoken on Toqo. See Otseqon phonology for discussion on dialect differences.

Vowels
/o/ is realized as [u] when the following syllable has /i/ or when followed by a coda /y/. Elsewhere it is a mid vowel [o].

/i/ lowers to [e] in the vicinity of uvular consonants.

Consonants

 * N Q| are coda archiphonemes that assimilate to the place of the following consonant.

/ʔ/ has a limited distribution, occurring word-initially and in the clitic /=ʔay/.

The dental consonants are laminal denti-alveolar, except for /d r/ which are apical alveolar.

The voiceless plosives are mildly aspirated.

Tone
Words in Otseqon are specified for tone on the first two mora, and then may have either a downstep or a peak at a lexically specified location in the word. The first two mora may each be either H or L. If the second mora is H, it spreads until the end of the phonological word or until a downstep. If the second mora is L, it spreads to the end of the phonological word, but a peak H tone may occur on one of the following moras.

Possible tone patterns for a two-mora word:
 * H-L
 * H-H
 * L-L
 * L-H

Possible tone patterns for a three-mora word:
 * H-L-L (H-L; no peak)
 * H-L-H (H-L; peak on 3)
 * H-H-H (H-H; no downstep)
 * H-H-L (H-H; downstep on 3)
 * L-L-L (L-L; no peak)
 * L-L-H (L-L; peak on 3
 * L-H-H (L-H; no downstep)
 * L-H-L (L-H; downstep on 3))

Possible tone patterns for a four-mora word:
 * H-L-L-L (H-L; no peak)
 * H-L-H-L (H-L; peak on 3)
 * H-L-L-H (H-L; peak on 4)
 * H-H-H-H (H-H; no downstep)
 * H-H-L-L (H-H; downstep on 3)
 * H-H-H-L (H-H; downstep on 4)
 * L-L-L-L (L-L; no peak)
 * L-L-H-L (L-L; peak on 3)
 * L-L-L-H (L-L; peak on 4)
 * L-H-H-H (L-H; no downstep)
 * L-H-L-L (L-H; downstep on 3)
 * L-H-H-L (L-H; downstep on 4)

Lexical suffixes may have a downstep or peak. This is deleted if the root has a specified downstep or peak, but surfaces if the root has no downstep or peak.

Note that two-mora roots may be specified for a downstep or peak on the third mora. This is realized on any affixes or clitics that follow.

Syllable structure
Syllables are CV(C); the coda consonant can be one of /ɾ j w/ or the archiphonemes |N Q|. Word-initially the onset may be /ʔ/. Codas do not occur after long vowels, though long [iː uː] could be analyzed as underlying /ij uw/, as those sequences do not occur.

Other coda consonants are found in ideophones, such as /tustustus/ ‘repeated sound of running fast’ and /kutɕkutɕkutɕ/ ‘chugging; sound of swallowing a liquid’.

Words must be at least bimoraic. Onsets do not count towards moraic weight. Vowels count for one, long vowels count for two, and a coda consonant counts as a mora. Thus /turu/, /kiː/, and /san/ are valid words, but /ta/ is not.

Phonological word versus grammatical word
The phonological word is distinguished from the grammatical word. A grammatical word is a free form that centers on a root and heads a phrase or functions as a member of a phrase (as a modifier, complement, auxiliary, etc). A word may be a bare root or morphologically complex with affixation and/or reduplication. A clitic is not a word, since it does not head a phrase or function as an internal member of a phrase. Rather, it is attached syntactically to a phrase, and phonologically to a word that is a member of the phrase. A phonological word is a morphosyntactic unit that consists of a ᴡᴏʀᴅ plus any clitics that attach to the host ᴡᴏʀᴅ. The phonological word is the domain of many phonological operations, and is the default meaning of "word" in this section.

/h/
/h/ assimilates features of the following vowel:


 * /hi he/ → [çi çe]
 * /hu ho/ → [ɸɯᵝ ɸɤᵝ], or [xᵝɯᵝ xᵝɤᵝ] particularly among older speakers
 * /ha/ → [χa]
 * /hə/ → [hə~xə]

Palatalization
Most consonants are palatalized before /i/: /w j/ do not occur before /i/, and uvular consonants lower it to [e] instead of palatalizing.
 * /m p pʼ β/ → [mʲ pʲ pʲʼ βʲ]
 * /n ʦ s/ → [ȵ ʨ ɕ]
 * /k kʼ x ɣ/ → [c cʼ ç ʝ]
 * /ɾ/ → [ɾʲ]
 * /h/ → [ç]

Postnasal voicing
Most consonants voice after the coda nasal |N|. Some additional changes occur:
 * N + /m n/ → [mm nn]
 * N + /p pʼ β/ → [mb]
 * N + /t tʼ/ → [nd]
 * N + /ʦ ʦʼ s/ → [nz~nʣ], or [ȵʑ~ȵʥ] before /i/
 * N + /k kʼ x ɣ/ → [ŋg], or [ɲɟ] before /i/
 * N + /q qʼ/ → [ɴq]
 * N + /h/ → [nz~nʣ]
 * N + /ɾ j w/ → [nn ɲj mm]

Gemination
Geminate consonants occur due to several morphological processes in Otseqon. Gemination de-ejectivizes ejectives and fortifies some fricatives.
 * /m n/ → /mm nn/
 * /p pʼ β/ → /pp/
 * /t tʼ/ → /tt/
 * /ʦ ʦʼ/ → /tʦ/
 * /s/ → /ss/
 * /k kʼ x ɣ/ → /kk/
 * /q qʼ/ → /qq/
 * /h/ → /ss/
 * /ɾ/ → /nn/
 * /j/ → /jj/ [ɟʝ]
 * /w/ → /mm/

/h/ historically developed from /s/, and retains its former value when geminated.

Vowel devoicing
High vowels devoice between a non-emphatic voiceless sounds. [e̞ ɤ̞ᵝ] lowered as a result of emphatic consonants do not devoice. The sequences /si su/ may optionally devoice to [ɕi̥ su̥] even when followed by a voiced sound unless the vowel is followed by a coda other than |Q|.

Vowels and sonorants always devoice utterance-finally.

Devoicing is somewhat unpredictable, and some vowels may not devoice even when expected. Devoicing is less likely when the previous vowel was devoiced. Otseqon dialects differ in the extent of their devoicing.

Lexical suffixation
When a lexical suffix is attached to a root, the initial consonant of the suffix undergoes one of the following alternations, conditioned by the root:


 * Glottalization: Aspirated stops become their ejective equivalents. In dialects with /mˀ nˀ/, /m n/ become /mˀ nˀ/. /s/ becomes /tsʼ/. Other consonants are unchanged. If the initial consonant is already an ejective, it remains ejective.
 * Gemination: The initial consonant of the lexical suffix is geminated. Some additional sound changes occur; see.
 * Insertion of -n: A coda nasal is appended to the root. This causes voicing and other changes to the initial consonant of the lexical suffix; see.

These are termed the glottalization, gemination, and n-grades. Roots that end with -n always trigger the n-grade, but no additional -n is inserted. Roots that end with -r, -j, or -w usually do not trigger any change. Some particular pairings of root + lexical suffix do not trigger any change.

Whether subsequent lexical suffixes trigger any change is unpredictable. Some trigger the gemination grade and some do not trigger any change.

Changes to coda /r/
Coda /r/ assimilates to /t/ before /t/, /tʼ/ (which loses its glottalization), and /ʦ/, /s/ before /s/, and /n/ before /n/, and dissimilates to /n/ before /r/. In other words, /rt/, /rtʼ/, /rʦ/, /rs/, /rn/, /rr/ become /tt/, /tt/, /tʦ/, /ss/, /nn/, /nr/ respectively.

Grassman's law for ejectives
Otseqon obeys a variation of Grassman's law that applies to ejectives. If an ejective is followed by another ejective in the next syllable, the first becomes the corresponding aspirated stop. This process can block the glottalization grade of lexical suffixation if the second syllable of the lexical suffix has an ejective.

Word classes
Words with concrete lexical meaning (i.e. non-grammatical) in Otseqon are divided into three main classes and three minor ones.

Roots
Roots are the primary and by far the largest class of words in Otseqon. Otseqon draws no distinction between nouns and verbs (see ), and in fact all roots in Otseqon are unaccusative verbs. Roots are inherently specified only for a single internal argument of an patient-oriented nature. This is the case even for verbs that strongly imply real-world agency, such as ‘to hit’, ‘to throw away’, ‘to cook’, etc. In Otseqon, the bare roots mean ‘to get hit’, ‘to get thrown away’, ‘to get cooked’, etc. More complex argument structures are derived; see.

Lexical suffixes
Otseqon has a very large class (several hundred items) of suffixes with concrete meanings (such as body parts, locations, etc), termed ‘lexical suffixes’. There is significant overlap between lexical suffixes and roots, but very few lexical suffixes are transparently related to standalone roots. Skilled speakers use lexical suffixes to background information and direct the listener's attention. The use of lexical suffixes instead of independent roots tends to denote routine, customary actions or states.

Lexical suffixes are divided into two types based on whether they incorporate into the root or incorporate the root into themselves. Incorporated lexical suffixes are far more numerous, with typically noun-like meanings and a function akin to noun incorporation in many languages. Incorporated lexical suffixes are used for backgrounded, non-specific referents, and their use is highly productive. Lexical suffixes that incorporate the root into themselves, termed predicative lexical suffixes, are fewer and generally restricted to locations and verbs of transfer. Such predicating suffixes can also function as independent predicates when attached to the morpheme mun ‘thing’, which functions in this use as an expletive host for the suffix, as shown in the pair below:

‘The man bought a house.’

∅-buy DET=man DET=house

‘The man bought a house.’

Predicating suffixes can only attach to objects. In the example below, the host for the predicating lexical suffix -heku ‘to buy’ must be interpreted as the object, even when semantically implausible:

? ‘The pigeon bought a man.’ (cannot mean ‘The man bought a pigeon.’)

Unlike typical noun incorporation, the incorporated root need not be a nominal:

qo-ici-heku

CL:round-three-buy

‘He bought three of them (small round objects).’

And it may be a wh-word:

kʷatta-heku na

what-buy Q

‘What did he buy?’

Interestingly, Otseqon lacks standalone roots with locative meanings. Locative notions are expressed exclusively by predicating lexical suffixes.

Classifiers
Classifiers are a large class in Otseqon, sharing some properties of both open and closed classes. The number of classifiers is fairly large, on the order of 200-300 words. It is mostly closed, but new words are occasionally introduced as ‘repeaters’ and may undergo semantic bleaching to widen their scope. However, like an open class, most Otseqon people do not know or use all of the classifiers regularly, and choice of classifier can be a stylistic matter like word choice.

Adverbs
Adverbs in Otseqon are a small closed class. Many of these have expressive meanings or otherwise add not-at-issue content (in the sense of Potts 2005, 2007). The location of adverbs within a clause is somewhat free, but they tend to occur immediately after the predicate, or after the final part of complex predicates.

Auxiliary verbs
Auxiliary verbs primarily express aspect (see ). They invariably come immediately before the main predicate.

Secondary predicates
Secondary predicates are predicates that only occur linked to another predicate with the linker morpheme /=ga/. They are all strong quantifiers.

Word order
Otseqon is rigidly verb-initial, with both VSO and VOS orders possible. Only VSO is admissible for clauses with both animate object and subject. However, pragmatically inferrable arguments are extensively dropped and including both subject and object is extremely rare in Otseqon sentences, generally acceptable only discourse-initially and if the communicative context cannot supply a subject for a transitive sentence. Once participants are established, they are referred to with pronominal prefixes and classifiers.

Nounlessness
Otseqon exhibits extreme predicate/argument flexibility. Any lexical item can serve as a predicate, including typically noun-like words like hanu ‘(to be a) person’, weak quantifiers like ken ‘(to be) many’ and ici ‘(to be) three in number’, wh-words, and proper nouns. Any lexical item can also serve as an argument, including typically verb-like words such as qata ‘to get hit (by a flying object)’. Verbs functioning as arguments (which must be preceded by a determiner such as ti) are equivalent to relative clauses with meanings like ‘the one who got hit’. One can reasonably say that Otseqon lacks nouns entirely, and all arguments are internally-headed relative clauses with meanings like ‘the one who is a person’, ‘the one who got hit’, ‘the one who I hit’, etc.

For example, given hanu ‘to be a person’ and kuusuti ‘to go into the interior’, ‘The person went inside’ could be expressed two ways:

This extreme flexibility is crucial to information structure in Otseqon. Sentences follow a comment-topic structure where new information always precedes given information. Focus is marked not with accent as in familiar languages such as English, but by its position as the primary predicate of the clause.

Argument structure
All Otseqon roots are associated with a single internal argument, an. The term is preferred to  because Otseqon groups the experiencer of perception verbs such as mita ‘to see’ with the patient of verbs such as qata ‘to get hit (by a flying object)’. Furthermore, any external argument is invariably interpreted as an instead of an, even for verbs that semantically entail the existence of an agent.

An interpretation of the oblique argument as an is impossible because the verb qata does not imply the presence of an agent argument. Introducing an agent is possible via a transitivizer marked by gemination of the final consonant of the root.

Now if the predicate is passivized an interpretation of the oblique as an is still possible because the word qatta ‘to hit’ entails the presence of an agent.

Transitivizers
All transitive verbs in Otseqon occur with a transitivizer. Otseqon has two primary transitivizers, termed direct and indirect.

The direct transitivizer is marked by reduplication of the final syllable of the word, minus any coda. Where possible the vowel is deleted, and this reduces to gemination of the final onset consonant. The direct transitivizer introduces an argument closely involved with the event, such as a volitional agent exerting control over the event. The direct transitivizer is what would be used to translate a sentence such as ‘I broke the glass [intentionally, for example in order to grab a fire extinguisher]’.

The indirect transitivizer is used when the introduced argument is less involved with the event. The introduced argument does not have control over the event, or is an inanimate force or stimulus. The indirect transitivizer would be used to translate ‘I broke the glass [accidentally]’. It is marked with the -ya suffix.

Antipassives
Otseqon has two antipassives, -n and -tsu. They are syntactically identical, demoting the patient to an optional non-core argument. Ergative personal prefixes appear on antipassivized verbs. The event is interpreted as atelic and the patient usually as nonspecific. If it occurs with an overt patient, that patient takes the e= preposition. Either can also yield a reflexive reading.

While the antipassives are syntactically identical, they're semantically distinct. -tsu has ‘job-like’ semantics, appearing on a subset of verbs that denote activities that tend to take time and effort. These are often conventionalized activities where the patient is understood as part of the cultural context. The word kusu ‘to be fermented/aged/ripened’ has a basic unaccusative argument structure, and with the addition of -tsu it is unergative: kukusutsu ‘I am fermenting’. However, this actually means that one is making tenna, a type of fermented bean cake, because the implied object of a -tsu antipassive is understood to be the culturally "normal" one, and tenna is a very common food to make.

In general, the choice of antipassive is mostly determined by the verb, with job-like verbs taking -tsu and everything else taking -n.

One Nominal Interpretation rule
TODO update examples

Otseqon syntax follows the One Nominal Interpretation rule: if a transitive verb is followed by a single DP, that DP must be interpreted as the object of the verb. This rule is very strong, and applies even in cases where it would be pragmatically ruled out.


 * 1) anqʼazzu wa ti hira

For the DP to be the agent, the verb must be antipassivized:

anqʼazzun wa ti hira

Thus the antipassive is commonly used for topic tracking because of this constraint.

Tense
There is no obligatory distinction between past and present tense in Otseqon, but there is an optional past tense clitic /=ta/. This is a second-position clitic, appearing after the predicate or after an auxiliary verb. In the absence of overt tense marking, the Aktionsart class of the root determines the default temporal interpretation. Stative predicates prefer a present tense interpretation, and everything else is interpreted as past by default. The continuous and habitual auxiliaries coerce the event into a state, and so yield a default present tense reading. The default temporal interpretation can be overridden by context.

The use of past tense clitic on non-stative predicates that would already be interpreted as past tends to imply background information. This is particularly true in narratives. In a story, it is not used for things that are actually happening in the story, but rather background information prior to the story's timeframe that may be relevant.

The past tense clitic may freely appear on typically nouny predicates (as these are just states). On animates, this usually indicates that the referent is deceased, or in the case of job titles, no longer holds the position.

Aspect
TODO examples need to be updated to the recent phonological changes

Aspect is marked by a pre-verbal auxiliary verb. Like adverbs, these auxiliary verbs have no morphological possibilities, but unlike adverbs they are restricted to pre-verbal position. Interestingly, they're all transparently related to non-auxiliary verbs. Some of the auxiliaries retain overtones of non-aspectual meanings.


 * wa < niwa ‘to be standing; to exist’. The only word that begins with [w]. Orthographically it is written 〈wa〉, but the vowel is lengthened to fill the two-mora constraint on prosodically independent words. It is rather high-frequency and denotes events that are currently in-progress at the topic time. It's not used with inherently stative predicates.

(1) ∅-suda ti gime za-∅-cii ni sara detta na?
 * gime "excessive continuous" < gime ‘to live’. Has overtones of "all the time" or "to excess", for example in (1)

3ᴀʙꜱ-what ᴅᴇᴛ ᴇxᴄᴇꜱꜱɪᴠᴇ.ᴄᴏɴᴛ 2ᴇʀɢ-3ᴀʙꜱ-happen<ᴄᴀᴜꜱ> ᴏʙʟ ᴅɪꜱᴛ ᴍɪʀ ɪɴᴛ

‘What are you doing there all the time‽’


 * ttu < niitu ‘to be known’. One of a handful of words with an initial geminate. A rather general auxiliary that aspectually coerces events into states (which happens to generally give a habitual reading). May also be used for characteristic properties of something, even if that property is already a state.

(2) mode ∅-kuu Taru ni Aritta
 * mode < mode ‘to be thrown away’. Used for events or states someone/something experienced which happened at least once without respect to a particular location in time and is repeatable.

ᴘᴇʀꜰ 3ᴀʙꜱ-go Taru ᴏʙʟ Aritta

‘Taru has been to Aritta.’ (note: not ‘Taru has gone to Aritta’. The implication is that he isn't still there.)


 * kate "finality perfect" < kate ‘to be put away’. Has overtones of "got it over with", or the general involuntariness of what happened. Similar to the experiential perfect, but implies that the speaker would rather it not happen again.

Frame of reference
Otseqon prefers absolute and intrinsic frame of reference. There are no words that mean ‘left’ or ‘right’.

Polarity pairs
Otseqon has a large number of pairs that differ only in polarity (e.g. ‘to have’ vs ‘to lack’, ‘good’ vs ‘bad’). In general the negative forms are somewhat weaker than using the positive form + negative marker. [Litotes] is not uncommon. Thus there are basically 4 levels:
 * positive (e.g. good)
 * not negative (not bad) (can also be very good via understatement)
 * negative (bad)
 * not positive (not good)

Seat of emotion
Otseqon consider the stomach to be the seat of emotions, like the heart in English. Additionally, while being warm-hearted is a good quality, in Otseqon being warm-stomached means one is angry.

Many emotional expressions use the lexical suffix *-su* ‘stomach’:


 * buunzu √buu ‘hot’ + -su stomach ⇒ ‘to be passionate, to be rageful’
 * oqosso √oqo ‘warm’ + -su ⇒ ‘to be mad’
 * saissu √sai ‘lukewarm’ + -su ⇒ ‘to be indifferent’
 * saassu √saa ‘cool’ + -su ⇒ ‘to be one who has cooled off from being angry’
 * kurutsʼu √kuru ‘cold’ + -su ⇒ ‘to be deliberate or thinking hard about something’

Back-channeling
These are conventionalized expressions that the listening party in a conversation interjects periodically. If you don't use these, the speaker assumes you're not paying attention.
 * 1) Often repeating the reportative evidential sai (with variants depending on formality register), with or without the demonstrative ka, and often with a key phrase that the speaker just said.
 * 2) sotohai (lit. “It IS understood”), using the =(h)ai polarity clitic.
 * 3) Just ai
 * 4) The adverb sira ("against expectations"), in this context meaning more or less “really?”
 * 5) ka ne “it is that?”

Conceptualization of time
The Otseqon conceptualize time as moving from east to west. Instead of the past being behind them and the future in front of them (or vice versa), the past is always to the east and the future is to the west. This is partly motivated by the direction of the sun during the day (rising in the east and setting in the west, like earth's) and partly because the Otseqon believe time is totally invariant with regard to humans, so one's orientation doesn't affect its path. The past = east future = west association extends to locations of things and the direction one faces during certain celebrations, ceremonies, traditional locations of things in the house, etc.

Additionally, in Otseqon, people and things don't move through time, time is what moves. So you'd say stuff like "The deadline is fast approaching" instead of "We're fast approaching the deadline" (English permits either, but the latter would be really unnatural in Otseqon). English also pervasively has the metaphor. Some of the time this is blatant like “How do you spend your time these days?” or “I've invested a lot of time into […]” or “You need to budget your time better” or “I saved some time by not brushing my teeth this morning”. More generally, however, we consider and understand and experience time as a thing that can be spent, wasted, budgeted, saved, have free time, etc. The Otseqon do not think like that. You don't spend time or save time. They don't say "hey, I have some free time, want to hang out?" rather, "hey, do you want to hang out?" because the justification of "I have some free time" is simply unnecessary because they don't think of having possession of time at all.

Metaphors
Like any language, Otseqon often uses words in a non-literal sense. For example, English has a somewhat pervasive metaphor of having control, force, or influence = up and being subject to control = down, e.g. ‘His power rose’, ‘His popularity is declining’, ‘He fell from power’, ‘I am on top of the situation’, ‘He's in a superior position’, etc. In Otseqon, the metaphor associated with power and influence is age: you'd say things like ‘His power aged’, ‘His popularity is like it was when it was new’, ‘I have experienced the situation before’, ‘He's in an older position’, etc. This is because in Otseqon culture authority is a product of age. You're supposed to respect elders and generally do what they ask. Being older is generally a good thing. This also extends to the popularity of things, and generally anything else associated with influence. In the US at least we seem to think of “new” as a feature, but the Otseqon are typically skeptical of anything new (which is expected to prove itself before it becomes widely popular).