Current status

Regarding the vitality of the language, the number of speakers indicate that this language is critically endangered, at moribund levels, rating an (8A) on the Ethnologue scale.^[3] This has to do with many factors, including the fact that in some communities, for example those living on the Terra Indígena Rio Guaporé, intergenerational transmission has ceased entirely due to a complete shift to Portuguese.^[2]

History of the language and population

This Brazilian state, like many others, was once strictly belonging to indigenous people of Brazil. The large-scale colonization of Rondônia happened only in the second half of the 20th century. At this time many people died from the introduction of preventable diseases to which they were not immune, and as a result, in the 1950s the Tuparí people were nearly wiped out entirely by measles. According to Caspar (1956),^[4] there were only 67 survivors of these epidemics, who were then forced to work for slave wages. The group eventually recovered and increased in their numbers. Now population growth is no longer an issue due to their large family sizes.^[5] However, their language is only spoken by approximately 350 people on the two main reserves (Rio Branco and Rio Guaporé). The cultural and linguistic situation varies by village, with some partaking in trans-generational linguistic transmission and others adopting Portuguese (Singerman, 2018, pp 6).^[5]

Existing literature on Tuparí

Tuparí is documented in the literature to some extent, with most publications being in Portuguese and English. It has also been discussed in German.

Major contributions to the literature include that of Franz Caspar, mentioned above, who wrote extensively from an ethnographic point of view in the mid 20th century.^[4] Several linguists have also produced materials on the Tuparí language. Poliana Alves has written on the phonology of the language (Alves, 1991)^[6] and has produced a bilingual dictionary (Alves, 2004).^[7] Aryon Dall'Igna Rodrigues collaborated with Franz Caspar to publish a sketch of the grammar (Rodrigues and Caspar, 2017).^[8] Lucy Seki, in 2001, wrote about the morphosyntax of nouns in Tuparí (Seki, 2001).^[9] In 2018, Adam Roth Singerman produced an extensive description and analysis of Tuparí for his PhD dissertation at the University of Chicago, focusing on the morphosyntax of the language (Singerman, 2018).^[2] As well, Singerman also worked to create a literacy workbook, Wan Tupari Ema’en Nika! (Tupari et al., 2016)^[10] to help boost vitality. His work adds information about the syntax and expression of evidentiality in Tuparí as hadn't been covered in previous research. In regards to non-linguistic references, there are Tuparí stories, written in Portuguese, as part of anthropologist Betty Mindlin's collection of indigenous myths (Mindlin, 1993).^[11]

Other major contributions to the literature on Tuparí include Monserrat (2000),^[12] Galucio (1993,^[13] 2011^[14]), and Moore (1994).^[15]

Documentation: Caspar and Singerman

Linguistically, the description of Tuparí was lacking prior to the publication of Singerman's dissertation in 2018,^[2] especially compared to others within its linguistic family and region (Singerman, 2018).^[5] However, Tuparí was one of the first languages and cultures of Rondônia studied in detail. This research was due to the Swiss ethnographer and explorer, Franz Caspar, who took many field-notes in the 1940s and the 1950s, as he lived within the community (Singerman, 2018).^[5] He is known in Tuparí as Toto Amsi Tàn (the long-nosed grandfather).^[5] Caspar's notes were passed on to Aryon Dall’Igna Rodrigues, a Brazilian linguist who wrote his dissertation in German on Tupinambá at the University of Hamburg. He wrote an eighty-paged description on Tuparí based on Caspar's notes, however it was never published.^[5] It is believed to be because, while he had come up with a description of the language, he had never done any fieldwork in the community himself.^[5] This was enough for him to deem that Caspar, being an anthropologist, could potentially have some linguistic errors within his field-notes and documentations (Singerman, 2018, pp 7–8).^[5] Even though the description of the Tuparí language prior to that done by Singerman was fairly limited, the cultural documentation was among the best in the region. With that being said, the Tuparí people are also currently experiencing a loss of that culture through a language shift, due to the overwhelming modern pressure for them to switch to Portuguese.

Consonants

Table adapted from Singerman (2018), pp. 370.

Vowels

Table adapted from Singerman (2018) pp. 367.

Valency manipulation

Valency manipulation refers to the grammar's ability to manipulate how many and what kind of arguments a verb can take. The Tuparí language uses four prefixes for this, causative m- /õ-, comitative-causative ete- /ite-, intransitivizing e-, and reciprocal e-. The causative morpheme makes the subject of an intransitive verb the object of a transitive verb. The distribution is limited to mostly unaccusatives and verbs of the semantic classes of verbs of motion, nonvolitional states or changes in state, and some verbs of thought or emotion. There are a few transitives that can take this morpheme. The comitative-causative morpheme promotes an intransitive subject to a transitive subject by introducing a new direct object. This morpheme requires that both the subject and the object are undergoing the acting together (i.e. coming). Singerman reports that his consultant approved the morpheme on a wide range of verbs of motion. It may also attach to some auxiliaries to express temporary possession. The intransitivizing morpheme e- is attested on a few intransitive verbs derived from unmarked transitive base. Common examples are the verb command transforming to speak, kill transforming to die. Intransitives derived with e- may also interact with other valency changing morphemes, such as causativization. Lastly is the reciprocal eue-. This morpheme can only occur on transitive roots. It requires plural subjects to be acting upon each other, i.e. fighting each other. Another notable feature is that the reciprocal can combine with other valency morphemes indicating that it probably occupies a position further left. This is further evidenced by the fact that it does not always fall under the domain of reduplication.

Singerman (121):^[25]

Singerman (150):^[26]

Evidentiality

Grammaticalized evidentiality (marking of the source of information) is functional, rather than lexical, and thus is morphologically expressed in a grammatical manner, rather than as an adverbial. Additionally, it has semantic characteristics. It is typically marked in Tupí languages by free particles rather than by bound morphemes, however in Tuparí, evidentiality is marked by a bound verbal suffix. This suffix agrees in number with the subject of the verb, and is required on a clausal level. In Tuparí, only past tense contexts mark a difference in witnessed and unwitnessed utterances, and evidentiality must be marked in these contexts. The sole purpose of the -pnẽ/ -psira morphemes is to mark this difference of witnessing. Evidential markers in Tuparí can easily be considered separately from epistemic markers since their strict syntactic positioning and clausal organization mean that they can only mark sources of evidence, and not other elements such as speaker attitude. They also must be marked in every clause, for example:

Singerman (347, 315):^[27]

There are six allomorphs of the evidential suffix in Tuparí, as seen in the table below:

The evidential marker in Tuparí always attaches to the highest verbal head. For example:

Singerman (313, 343):^[28]

In this case, since there is no auxiliary present, it attaches to the lexical verb 'arrive'.

Due to always agreeing with sentential subjects, -pnẽ only works with singular subjects and -psira only with plural ones.

Example: (345) Evidential marking on the lexical verb ‘come’ Singerman (315):^[29]

a. te- ‘3C’ + s ‘come.SG’ + EV → tèynẽ

b. te- ‘3C’ + ã’ẽ ‘come.PAUC’ + EV → teã’emsira

c. te- ‘3C’ + ip’anẽ ‘come.PL’ + EV → teip’anemsira

The variations depend on whether the lexical verb distinguishes between singular, plural and paucal (two), or just between singular and plural.

The Tuparí case system

Tuparí has four cases, as seen in the table below.^[30] Note that Nuclear and Locative cases can be stacked.

As demonstrated in the table above, there are four case types in Tuparí: nuclear, locative, instrumental-lative, and oblique. Additionally, under certain conditions, Nuclear and Locative cases are able to stack (for reasons that will be discussed below). As mentioned by Singerman (2018), the case markings are used in complementary distribution, unless they are referring to time, in which case LOC (on Portuguese loan-words), INS (on Tuparí temporal expressions) and OBL (in finite embedded clauses marking time) may be used.