Kanien’kéha

• Dataset 1: Obstruent voicing in Kanien’kéha (phonology, phonemes, voicing) 
• Dataset 2: Word order and information structure in Kanien’kéha (word order, information structure, focus, topic)
• Dataset 3: Verbal morphology in Kanien’kéha (morphology, verbs, templates, polysynthesis)
• Dataset 4: Nominal morphology in Kanien’kéha (morphology, nouns, templates, polysynthesis)
• Dataset 5: Noun Incorporation and Information Structure (noun incorporation, information structure, focus, word order) (DeCaire et al. 2017)
• Dataset 6: Pronominal Prefixes in Kanien’kéha (syntax, morphology, transitivity, argument structure)
• Dataset 7: The meaning of verb suffixes in Kanien’kéha (semantics, verbal meanings) 
• Dataset 8: Applicatives in Kanien’kéha (syntax, argument structure, applicatives)
• Dataset 9: Accent in Kanien’kéha (phonology, stress, tone)
• Dataset 10: Epenthesis in the history of the Iroquoian language family (phonology, historical linguistics, epenthesis)

These data show that Kanien’kéha obstruents lack a voicing contrast. The stops /t/ and /k/ are voiced when immediately preceding vowels, and voiceless everywhere else. The fricative /s/ also lacks a voicing contrast, but behaves in a more restricted manner than the stops: It is only voiced inter-vocalically or pre-vocalically in word-initial position, and is voiceless everywhere else. That is, contrary to the stops which only need the following segment to be vocalic in order to be voiced, /s/ cares both about the following and the preceding segment (unless there is none, when /s/ is word-initial), as both must be vocalic for /s/ to be voiced. This dataset can be used in a basic introduction to phonology course, in order to demonstrate the lack of a phonemic voicing contrast, and thus the presence of phonemes which correspond to two distinct phonemes in English, which can be useful to emphasize the non-universality of contrasts which English speakers may consider basic and common. 

• Oh nahò:ten   tesatonhontsió:ni? 

what                  you.want.it 

‘What do you want?’ 

•  Kanonhsá:se   tewakatonhontsió:ni. 

new.house         I.want.it 

‘I want a new house.’ 

•  Ónhka   wa’ena’tarón:ni’   ne   kí:ken   tekahswà:ne,   Só:se   káton   Wá:ri? 

 who        they.baked.it        NE  this       pie                      Joe       or          Mary 

‘Who made this pie, Joe or Mary?’ 

•  Wá:ri   wa’ena’tarón:ni’   ne   kí:ken   tekahswà:ne. 

Mary     she.baked.it         NE   this       pie 

‘Mary made this pie. 

•  Wahshní:non’   ken   ne   kahonwé:ia’? 

you.bought.it      Q      NE   boat 

‘Did you buy the boat.’ 

•  Iah,   wa’kón:ni’   ne   kahonwé:ia’. 

no      I.made.it     NE   boat 

‘No, I built the boat.’ 

• Oh nahò:ten   ninón:wes    ne   sheién:’a             tánon’   ne   tsién:’a? 

what                 they.like.it    NE   your.daughter   and       NE   your.son 

‘What do your daughter and son like?’ 

• Riién:’a   raon’wéskwani   ne   ahá:raste,                tánon 

my.son   he.enjoys.it          NE   he.would.draw.it   and 

       kheién:’a           iakaon’wéskwani   ne   aióntswa’te’            tewa’á:raton. 

       my.daughter   she.enjoys.it             NE   she.would.play.it   lacrosse 

‘My son likes to draw, and my daughter likes to play lacrosse. 

• Oh nahò:ten   ísenehre           énseneke’? 

what                  you.two.want   you.two.will.eat 

‘What do you two want to eat?’ 

• Akaónha   kítkit       én:ieke’,             tánon’   í:’i   só:ra    én:keke’. 

she             chicken   she.will.eat.it   and        I      duck   I.will.eat.it 

‘She will have the chicken, and I will have the duck.’ 

This dataset shows that word order is relatively “free” in Kanien’kéha, with no basic default ordering of S, V, and O. Instead, the linear order of core sentence constituents is determined by information structure: Focus elements (bolded) always occur in initial position. Topic elements (underlined) are also fronted to initial position. When the two co-occur, both types of elements are fronted, but (contrastive) topics occur first, outside of focus elements, as shown in the last pair. This is consistent with cross-linguistic trends, where topics are usually higher than / outside of foci. This dataset can be used in more advanced syntax classes, to demonstrate the influence of information structure on word order in certain languages, which is very different from the more fixed word order of English, in which information-structural categories are usually marked instead by intonation or special constructions. 

This dataset illustrates key aspects of Kanien’kéha’s polysynthetic verbal morphology. It could be a good exercise for an intermediate to advanced morphology class, in which students are asked to study the form-meaning pairings in order to propose a template of Kanien’kéha verbal morphology. The complexity of this template is a good example of polysynthesis, and can thus be used to broaden the students’ typological horizon (as most of them may only be familiar with highly isolating languages, such as English or Mandarin). The target template is provided below: 

NEG-MOOD-ARG-VOICE-NOUN-NMZ-ROOT-DER-ASP1-ASP2 

DER suffixes: applicative, purposive, distributive 

This dataset is similar to the previous one, but focuses this time on nominal morphology. We can see that nominal morphology is not as rich and complex as verbal morphology. This is typical of polysynthetic languages, which are usually “verb-centric”. The target template that students could be asked to construct in an intermediate to advanced morphology class is the following: 

EXPL/POSS-ROOT-NSF/LOC-CLTC 

CLTC clitics: plural, augmentative, prototypical, decessive 

• 1a) Oh       nahsátiere’   onwáhstsi’?

      what   you.did         now 

‘What did you do today?’ 

• 1b) Wa’kenaktahní:non’. 

      I.bed.bought 

‘I bought a bed.’ 

•  2a) Wahsena’tarón:ni’   ken? 

        you.made.bread      Q 

‘Did you make bread?’ 

• 2b) Iah,   wa’kenaktahní:non’. 

       no     I.bed.bought 

‘No, I bought a bed.’ 

•  3a) Wá:ri    ken   wa’enaktahní:non’? 

        Mary    Q       she.bed.bought 

‘Did Mary buy a bed?’ 

• 3b) Iah,   í:’i   wa’kenaktahní:non’. 

        no    I      I.bed.bought 

‘No, I bought a bed.’ 

• 4a) Wahsehonwahní:non’   ken? 

        you.boat.bought            Q 

‘Did you buy a boat?’ 

• 4b) Iah,   kanákta’   wa’khní:non’. 

        no     bed            I.bought.it 

‘No, I bought a bed.’ 

• 5a) Wahsenaktón:ni’   ken? 

       you.bed.made       Q 

‘Did you build the bed?’ 

• 5b) Iah,   wa’khní:non’   ne   kanákta’. 

       no     I.bought.it       NE   bed 

‘No, I bought the bed.’ 

These data show that noun incorporation is not a mere stylistic choice between two equivalent constructions, but is instead strictly governed by information structure: In cases of predicate focus (pairs 1 and 2) or subject focus (pair 3), noun incorporation is obligatory where possible; in cases of object focus (pair 4) or verb focus (pair 5), noun incorporation is impossible. That is, noun incorporation is the obligatory default everywhere, except when focus targets one of its two constituents (the object or the verb), which disrupts and prevents incorporation. This is a more advanced dataset for students interested in more specific features of the language, or for advanced morphosyntax or information structure classes which aim to broaden students’ typological horizon. 

This is a bit more difficult dataset for advanced morphosyntax classes focusing on argument structure and verb classes. The data aim to show that Kanien’kéha has three series of argument tracking morphemes: (a) a “subjective / agentive” series, used with human > non-human transitives (1-4) and unergatives (5-8); (b) an “objective / patientive” series, used with animal (see note below) > human transitives (9-12) and unaccusatives / statives (13-16); and (c) a “transitive” series, used with human > human transitives (17-20). As marking is identical between unergatives and transitives with a human agent, and between unaccusatives / statives and transitives with a human patient, the broader generalization is that these morphemes only formally track human arguments (agents for the agentives, patients for the patientives, both for the transitives), and not non-human arguments. The dataset is designed in a way that allows users to ignore the debate regarding whether these morphemes are the arguments themselves, or agreement markers.

Note: Northern Iroquoian languages are well-known for having “two feminine genders”. The feminine-indefinite (not featured here) is used for impersonals and some human females. The feminine-zoic (the first “it” argument in 9-12) is used for other human females and animals. 9-12 could thus also be translated as “she likes/sees me/you”. The choice between the feminine-indefinite and feminine-zoic when referring to a woman is subject to a complex interplay of several sociocultural factors, which we cannot discuss here. To simplify, it is assumed in the dataset above that the first argument in 9-12 is an animal, and not a woman.

These data show a range of representative examples for three verbal suffixes commonly used in Kanien’kéha, namely the “distributive” (first table), the “reversive” (second table), and the “purposive” (third table). These data make the meaning of each suffix clear: The distributive expresses that the action denoted by the verb is distributed over many times, objects, or places; the reversive inverts the action denoted by the verb; and the purposive indicates that the subject is on the way to do the action denoted by the verb. These data can be used in introductory semantics classes, as good examples of the generalizations one can propose to determine the lexical semantics of specific morphemes. 

These data provide examples of the so-called “benefactive” suffix in Kanien’kéha, which is essentially a regular high applicative. It can be used with any verb in order to add a general “affectee” argument. That this morpheme can be added to any verb (whether transitive or intransitive) and that it adds a general “affectee” argument rather than a specific “goal” argument suggests that it is a high as opposed to a low applicative in Pylkkänen’s sense. This dataset could be used in an intermediate syntax class when applicatives are introduced, as a useful test case for distinguishing high versus low applicatives. For instance, the first two or three pairs might lead one to think that this is a low applicative, given that we only have transitive verbs with a kind of “transfer-of-possession” interpretation, but the rest of the data then shows that this morpheme can be added to intransitives, and is therefore a high applicative. 

This is a more advanced phonology dataset that can be used as an example for pitch accent systems when discussing the range of typological variation in stress systems. Kanien’kéha is a pitch accent language, in which every word has one stressed syllable (but stress is not lexically contrastive). Stressed syllables can receive one of three “accents” (combination of lengh and tone, concept specific to Iroquoian): high accent (high tone on short vowel, 1-7); rising accent (rising tone on long vowel, 8-14); and falling accent (falling tone on long vowel, 15a-21a). Exercises of various difficulty can be based on this dataset. An easy task would be to determine the position of the stress or accent: It is systematically penultimate (with some surface exceptions straightforwardly explained by historical reasons, but excluded here for simplicity). A more difficult exercise would be to ask students to determine the distribution of the three different accent types. High accents are found before consonant clusters. Falling accents are found before singleton consonants when there is an underlying glottal (glottal stop or /h/) between the accented vowel and the following consonant, as demonstrated by the resurfacing of the glottal when accent shifts (15b-21b). Rising accent is found elsewhere, namely before singleton consonants, when there is no underlying glottal.

This dataset aims to provide some historical perspective by exemplifying one major sound change process that occured in the shift from Proto-Northern Iroquoian to modern Kanien’kéha: All obstruent (k, t, s) – sonorant (r, n, w) clusters were broken up by an /e/ epenthetic vowel (except for /k/-/w/, given that /kw/ is a single labio-velar segment). These data could be used in an introductory or intermediate historical linguistics class, as an exercise in which students would be asked to explain this process of sound change as concisely as possible, using phonological natural classes like “obstruents” and “sonorants”. For a more advanced exercise, students could be asked to provide an hypothesis explaining why this new phonotactic constraint does not apply to /kw/, or to make the connection with the previous dataset on accent / stress: Some modern Kanien’kéha words have antepenultimate instead of regular penultimate stress on the surface, because the penultimate vowel is an epenthetic /e/ which was added after the establishment of the penultimate stress system in the history of Northern Iroquoian (i.e., the stress system was established in Proto-Northern Iroquoian first, and then epenthesis was added later in Kanien’kéha). This can be seen in the fact that Proto-Northern Iroquoian words all have regular penultimate stress once epenthetic vowels are removed.