Structure and Morphology of Chamikuro Kinship

Chamikuro is an Arawakan language traditionally spoken in the headwaters of the Samiria River in the Loreto region of Peru, today with only a handful of speakers of varying abilities over the age of 80. The morphosyntax of the language was sparsely documented until last summer, in work Lev Michael and I carried out in Yurimaguas. In this post, I want to focus on the structure and morphology of the kinship system, which has received no systematic attention to date. The data comes from linguist Steve Parker (2010; speaker †Gregorio Orbe Caro, based on work in 1985, 1987, and 1993), a wordlist recorded by Lee Bendezú for the Ministry of Culture in 2019 (speaker †Natalia Sangama Orbe, 1932-2021), the unpublished notes of Christine Beier from 2022 (CLA 2022-08.005, 2022-08.010; speaker Antonio Inuma Orbe, b. 1938, first cousin to Natalia), and my collaboration with speaker Alfonso Patow Chota (b. 1925). I also make use of notes from anthropologist Tony Stocks, also based on work with Gregorio Orbe and his wife, Elisa Sangama (another first cousin to Natalia), from August 1975 and in process with the California Language Archive. Note that not all terms are equally documented across all sources.

I will walk through the terms in each generation for consanguines and affines, discussing what is known with confidence, what with some doubt, the details of conflicting information, and offering some hypotheses. I will note three suffixes attested only with kin terms, including a vocative -ko, a suffix -lawa, which seems to derive female-ego terms from male-ego ones, and a suffix -wakne, which seems to derive affinal terms from consanguineal ones. An alienability distinction also affects the Chamikuro kinship system. There are important ego-based distinctions at ego's generation and each generation above and below (indicated by ♂ and ♀), as well as same-sex age-grade terms among siblings. There are terminological traces of a former cross-cousin marriage practice, but it is worth emphasizing that this does not ramify clearly throughout the system. This is unsurprising given that Chamikuros were missionized by the Catholic Church beginning in the mid-17th century. I have written about similar issues in the case of Omagua and Kukama-Kukamiria (O'Hagan 2019). Unless mentioned otherwise, all described terms are referential; there does not seem to be a lexically distinct set of vocative roots as with some terms in the languages of the Nijagantsi branch of the family that constitute Chamikuro's closest relatives. I cite forms with their possessive prefix segmented. All segmental representations are IPA.

At ego's generation, there are ego-based distinctions that also affect the presence of age-grade terms. For male egos, there is a single term for 'sister,' and a term each for 'older brother' and 'younger brother.' For female egos, conversely, there is a single term for 'brother,' and a term each for 'older sister' and 'younger sister' (see below). The suffix -lawa is found on both female-ego sister terms. For Gregorio Orbe, 'older brother (♂)' is /tawu/, whereas for Alfonso Patow and Natalia Sangama it is /ʰtawu/. In general, Alfonso Patow's phonological forms are prioritized, as he pays careful attention to the production of laryngeal segments. However, Alfonso's form for 'sister (♂)' is /mameʔti/, with a different final vowel, contradicting all three other speakers and otherwise making it homophonous with 'be embarrassed.'

Reference of sibling terms

u-mameʔtu 'my sister (♂)'

u-koneʔte 'my younger brother (♂)'

u-htawu 'my older brother (♂)'

u-lotiaka 'my brother (♀)'

u-koneʔlawa 'my younger sister (♀)'

u-htawlawa 'my older sister (♀)'

At one generation above ego, it is known with confidence that there is an ego-based distinction for parents. It is possible that 'father (♂)' is /ˀʂaːma/, as -ne is an expected possessive suffix for alienable nouns, but so far alienable nouns have only been attested clearly for terms for grandparents. 'Father' is not currently attested in the vocative, which could provide other evidence, nor do there seem to be cognates in other Arawakan languages.

Reference of parent terms

u-ʔʂaːmane 'my father (♂)'

w-aʔjihku 'my mother (♂)'

u-skawna 'my father (♀)'

u-mawtuːke 'my mother (♀)'

Among parents' siblings, there is some doubt about the structure of the system. Parker (2010) documented a single term for 'uncle,' u-kohka, referring, one infers, to both father and mother's brothers of both male and female egos (for other terms he makes the latter distinction clear). The lack of an ego-based distinction is confirmed by Alfonso Patow. (The form is not documented in the speech of Antonio Inuma or Natalia Sangama.) However, the form w-eːkoneʔte is documented in the speech of Natalia Sangama, with the gloss of 'sobrino a tío' ('nephew to uncle'). Unfortunately nothing more is known about the elicitation context. The form is strikingly similar to 'younger brother (♂),' and, if the prompt were something like "How does my son refer to my brother?" one can imagine the gloss being spurious. Furthermore, Alfonso Patow suggested that the form u-ʃiliskaʔtuːke, glossed by Parker (2010) as 'stepfather,' could refer to a man's uncle, suggesting there is another uncle term in the system. Finally, it is worth noting that the cognate to /kohka/ in Nijagantsi languages refers only to one's mother's brother.

For 'aunt,' Parker (2010) does document an ego-based distinction, but not one that distinguishes father and mother's sisters (another piece of evidence indirectly suggesting there should be a similar distinction for 'uncle'). According to him, u-patiahka is 'my aunt (♂)' and w-ahka is 'my aunt (♀),' which is confirmed by Alfonso Patow. However, two other forms for 'aunt' are documented in the speech of Natalia Sangama: u-ʂaʔtapale 'my aunt (♂)' ('sobrino a tía') and u-ʂaʔtake 'my aunt (♀)' ('sobrina a tía'). The former is attested only in her speech, the latter also by Parker (glossed as 'stepmother') and in the speech of Alfonso Patow. In the latter case there is additional elaboration and also a specific example of use. Alfonso first claimed this word u-ʂaʔtake meant 'my aunt,' but clarified that it referred to another wife of one's father, which is consistent with Parker's 'stepmother.' He said that his mother, Tomasa Chota Alegría (c1896-1968), referred to an elderly woman named Martina Ojaicuro, who lived in Tomasa's household when Alfonso was a boy, with this term. Unfortunately it is not known how Martina was related to Tomasa. However, if there is some sense in which one's aunt is also one's stepmother (u-ʂaʔtake), then it suggests that one's aunt could become one's stepmother (either as a co-wife of one's father or after the death of one's mother), in turn suggesting that u-ʂaʔtake could not have referred to father's sister.

Chamikuro Kin Terms from A. Stocks's Field Notes
(courtesy of California Language Archive)

Thus we have multiple uncle-like words and multiple aunt-like words. That suggests either an ego-based distinction or a parallel-cross distinction or both, a maximum possibility space of eight distinct terms. Here Stocks's field notes, meant to be preliminary, shed some light (see preceding figure). He documents six distinct terms in this space. Here it is worth noting that Stocks worked with husband and wife together, making significantly easier the elicitation of ego-based distinctions. The form u-kohka indeed is 'my maternal uncle (♂),' but only of male egos. And the other two uncle terms are attested: u-ʃiliskaʔtuːke is 'my paternal uncle (♂)' and indeed for male egos; and w-eːkoneʔte appears in two places, with significant differences from what was documented by the Ministry of Culture. According to Stocks, this term is a female-ego term for either maternal or paternal uncle. Among aunt terms, there is the ego-based distinction between u-patiahka is (♂) and w-ahka is (♀), but these terms only refer to father's sister. For mother's sister, it is claimed there is no ego-based distinction, both being u-ʂaʔtake. However, the lack of an ego-based distinction here goes against both Alfonso Patow and Natalia Sangama's claims. It may be that Natalia's u-ʂaʔtapale (♂) should form a pair with u-ʂaʔtake (♀). This would bring a certain similarity to aunt terms (but it would contradict Alfonso's claim that u-ʂaʔtapale is a female-ego term). Each pair would share common roots: u-patiahka 'my paternal aunt (♂)' would seem to be formed on w-ahka 'my paternal aunt (♀)'; and both u-ʂaʔtapale 'my maternal aunt (♂)' and u-ʂaʔtake 'my maternal aunt (♀)' would share a common /ʂaʔta/. This leaves the case of w-eːkoneʔte, according to Stocks 'uncle (♀).' I suspect this is only a male-ego form, as it seems to be related to /koneʔte/ 'younger brother (♂).' This in turn raises the possibility that the age-grade distinctions at ego's generation, which Stocks failed to document, might also be present at one generation above ego, which would complicate the system more than is appreciated here.

Before moving on to the generation below ego, I note more about what Stocks's notes teach us at ego's generation, which is the extent of his documentation of the system. First, as noted previously, instead of documenting an age-grade distinction at ego's generation, Stocks documented u-lotiaka (given above as 'my brother (♀)')  as both 'my sister (♂)' and 'my brother (♀),' that is, a term for opposite-sex sibling. (Relatedly, if one reads the documentation from the Ministry of Culture carefully, this term has yet a different meaning, simply 'my sibling (♀)'.) Consequently, instead of u-mameʔtu 'my sister (♂),' it is 'my brother (♂).' Only u-htawlawa 'my older sister (♀)' is (partially) correctly given as a term for 'sister (♀).' I take these to be simply errors that would have been rectified had Stocks had more time.

What is notable in the preceding figure is that the terms for the children of father's brother are exactly the same as those for one's siblings, whereas those of father's sister are different. This suggests a distinction between parallel and cross-cousins, and thus likely cross-cousin marriage. From here, things get more complicated, and likely incorrect. Terms similar to those Stocks documents for the children of father's sister (cross) are found with the children of mother's sister (parallel). Central to the discussion here is the form u-hpuːʂana. Parker glossed this simply as 'sister-in-law.' Alfonso Patow claims it refers to a man's sister-in-law and a woman's brother-in-law. In a practice of cross-cousin marriage, one's cross-cousins become one's in-laws, and so one expects the terms that Stocks gives for certain cousins to have in-law meanings as well. For the time being, I will ignore Stocks's claim that mother's sister's children are referred to with these terms, focusing on father's sister's children. Here his claim is that u-hpuːʂana is 'brother-in-law,' regardless of ego, and that u-nahtona is 'sister-in-law,' also regardless of ego. I doubt this. The latter is clearly cognate to what in Nijagantsi languages refers only to a woman's female cross-cousin or sister-in-law, never a man's. (In Nijagantsi societies, men don't have sisters-in-law in a terminological sense, since they are all marriageable partners; and women don't have a brothers-in-law, for the same reason.) Furthermore, for Antonio Inuma, u-nahtona is explicitly 'sister-in-law (♀),' whereas u-hpuːʂana is 'sister-in-law (♂)' (pace Nijagantsi people). It is clear from Alfonso Patow that there is a male-ego brother-in-law term, namely w-aʔsahte. One way of reconciling the information from both Antonio and Alfonso is as follows. Alternatively there may be a distinct female-ego brother-in-law term, but, according to Antonio Inuma, there is no such term.

Possible reference of in-law terms

u-hpuːʂana 'my sister-in-law (♂), my brother-in-law (♀)'

u-nahtona 'my sister-in-law (♀)'

w-aʔsahte 'my brother-in-law (♂)'

Moving on to the generation below ego, there are distinct terms for children depending on the ego, as follows. With the exception of the fact that 'son (♂)' is derived from 'son (♀),' there is nothing to remark on here.

Reference of children terms

u-ʈʂomati 'my son (♂)'

u-mutle 'my daughter (♂)'

u-ʈʂoma 'my son (♀)'

u-hʃini 'my daughter (♀)'

Given that the parallel-cross distinction has held at the generation above ego and at ego's generation, it likely holds at the generation below as well. Parker gives the following forms for nieces and nephews without reference to a parallel-cross distinction. These are different from the forms he cites for 'son-in-law' (w-ohnawakne) and 'daughter-in-law' (u-htuʔnina), in this latter case without either ego-based or parallel-cross distinctions.

Parker's terms for nieces and nephews

w-ohnalawa 'my niece (♂)'

u-makʂoʔtuːke 'my nephew (♂)'

u-hsaliʃini 'my niece (♀)'

w-ahɲaje 'my nephew (♀)'

The facts that the female-ego niece term is based on 'daughter,' and that there are separate child-in-law terms leads me to believe that there is a parallel-cross distinction among nieces and nephews. Here we are on much shakier ground. These terms are not documented in the speech of either Natalia Sangama or Antonio Inuma (and Stocks did not document his own versions from Gregorio Orbe), and Alfonso Patow was regularly unconfident about terms for non-lineal ancestors and descendants. But there is a reasonable amount of circumstantial evidence to offer some hypotheses. First, if there is a parallel-cross distinction at this generation, then parallel nieces and nephews could simply be referred to with the terms for ego's corresponding children or ones derived from them, as with u-hsaliʃini 'my niece (♀)'. So in actuality we are discussing terms for cross-nieces and -nephews and possibly distinct terms for corresponding affines. Second, we know that -lawa is a productive suffix (cf. w-ohnalawa 'my niece (♂)'). Third, -wakne also seems to be a productive suffix (cf. w-ohnawakne 'my son-in-law'). A critical datum is attested only in the speech of Antonio Inuma: u-hkohkawakne 'my father-in-law,' based on /kohka/ 'maternal uncle (♂).' This suggests that terms for cross-aunts and -uncles are at least in some cases not identical to their affinal counterpart terms, but that the latter take the suffix -wakne. These facts all relate to the root /ohna/ in the list above. I suspect that w-ohna referred to a man's cross-nephew, not cross-niece, and that -lawa derived the female-ego counterpart, as it does for sibling terms (n.b., Alfonso Patow did not even recognize Parker's nephew term, u-makʂoʔtuːke). This would allow straightforwardly for w-ohnawakne 'my son-in-law,' to which we would add that it was only for male egos.

Proposed cross-niece/nephew terms

u-htuʔnina 'my cross-niece (♂)'

w-ohna 'my cross-nephew (♂)'

u-htuʔnina 'my cross-niece (♀)'

w-ohnalawa 'my cross-nephew (♀)'

The question then becomes, how to think of Parker's u-htuʔnina 'my daughter-in-law,' u-makʂoʔtuːke 'my nephew (♂),' and w-ahɲaje 'my nephew (♀)?' If u-htuʔnina is 'daughter-in-law,' and if it also has consanguineal reference, then it must also refer to a cross-niece. But male or female ego? In Caquinte (Nijagantsi), 'cross-nephew, son-in-law' is neutral with respect to ego, but 'cross-niece, daughter-in-law' is not. Could something similar be happening in Chamikuro? Furthermore, my proposed reorganization of the terms in this part of the kinship system means that, if u-makʂoʔtuːke refers to a man's nephew, then it must be a parallel nephew. Similarly, if w-ahɲaje refers to a woman's nephew, then it must be a parallel nephew. And we are left without a term referring to a man's parallel niece.

Proposed parallel niece/nephew terms

?          'my parallel niece (♂)'

u-makʂoʔtuːke 'my parallel nephew (♂)'

u-hsaliʃini 'my parallel niece (♀)'

w-ahɲaje 'my parallel nephew (♀)'

There is some reason to think that this is not entirely on the right track. First, this reorganization hinges on the fact that Parker was wrong to claim that w-ohnalawa referred to a woman ('niece'). We do not typically encounter errors of reference of this kind, rather ones of ego and parallel-cross distinctions. (But the alternative, that stems based on /ohna/ refer to 'son-in-law' and 'niece,' is similarly unlikely.) Second, the female reference of w-ohnalawa is weakly supported by Natalia Sangama's form w-ohnelawa, with a different vowel, which the Ministry of Culture gave the gloss 'primo a prima,' suggesting it is a male-ego term for a female cousin. This is difficult to square with this term otherwise being attested only at the generation below ego (and it may be a different root altogether, namely /ohne/), but it is suggestive. Third, there end up being two lexically distinct terms for parallel nephews, distinct from the terms for ego's corresponding children. This kind of lexical dissimilarity is not borne out with parallel cousin terms at ego's generation.

Two generations removed from ego things get much simpler. In the entire corpus only Parker documents a gender- and ego-neutral word w-atohka 'my grandchild.' Alfonso Patow could not remember a word for 'grandchild,' and indicated that his own maternal grandmother, a Chamikuro-speaking woman, simply addressed and referred to him by his name. At two generations above ego, there are distinct terms for 'grandfather' and 'grandmother' regardless of ego, namely u-ʔloːko-ne and u-ʔpaːja-ne, respectively. Here things get morphologically more interesting. The only vocative kin terms documented to date are ones for 'grandfather' and 'grandmother' (mea culpa), and they reveal a dedicated vocative suffix -ko (i.e., loːko-ko and paːja-ko, the corresponding terms of address) not attested elsewhere in the language. I suspect this suffix could have been used with at least some other kin terms, such as 'father (♂),' in which there appears to also be a possessive suffix. Perhaps it was even more general.

Christine Beier's work with Antonio Inuma reveals one final term of importance, namely u-hsuʔmahke 'my mother-in-law.' This form was also documented by the Ministry of Culture with Natalia Sangama (in whose speech the root-initial aspirate seems to be absent), but with the gloss 'son-in-law.' Instead of a reciprocal term, otherwise not found in the Chamikuro kinship system, I suspect that what was asked was how a son-in-law refers to his mother-in-law; it is the only gloss in this section of the wordlist not given as 'X to Y' (see other example of this cited above). Alfonso Patow did not recognize this word, nor is it documented by Parker or Stocks. Importantly, it would seem to be an affinal term distinct from the other cross-aunt terms reviewed above. This leaves open the possibility that there may be more pervasive cross-affinal distinctions across the system, which would have consequences for my hypotheses about the parallel and cross-niece/nephew terms and their affinal counterparts.

The most work remains to be done among terms for cousins, nieces, nephews, and affines. The prospects for learning more are not good, but are most promising with Antonio Inuma, who remembered a set of terms that Alfonso Patow could not. The latter is understandable: Alfonso's father was not a Chamikuro person or speaker, the only sibling of his mother's he knew was a half-uncle, and his maternal grandmother, the only Chamikuro-speaking grandparent he knew, died when he was about seven years old. His mother's family was more distant genealogically than the remaining (few) Chamikuro-speaking families of the Pampa Hermosa of his childhood. The other speakers knew more grandparents, and were relatively closely related cousins amongst themselves. Let's see what can be done in Yurimaguas this coming summer!