Vowels

The vowel system of Bororo is somewhat cross-linguistically unusual in that it distinguishes roundedness only in its back vowels (although Crowell (1979) analyzes unrounded “back” vowels as central^[4]).

The mid vowels /e ɤ o/ alternate with the open-mid (or "lax") vowels [ɛ ʌ ɔ] in apparent free variation.

The unrounded back vowels /ɯ ɤ/ become central [ɨ ɘ] word-finally. Furthermore, [ɘ] (i.e. word-final /ɤ/) is not distinguished from, and often surfaces as, front [ɛ]. As a result, /ɤ/ and /e/ are at least partially merged word-finally.

Consonants

Unusually for a South American language, Bororo has no phonemic fricatives.

The velar stop /k/ becomes aspirated [kʰ] before nonback vowels /a e i/; the labiovelar approximant /w/ becomes labiodental [ʋ] in the same environment. Plain and labialized velar stops /k g kʷ ɡʷ/ merge as labialized [kʷ ɡʷ] before rounded vowels /u o/. The voiced velar stop /g/ lenites to fricative [ɣ] between vowels.

Syllable structure and stress

The canonical syllable structure in Bororo is (C)V: that is, a mandatory vowel nucleus (or diphthong), optionally preceded by a single onset consonant. Aside from unmodified loanwords from Portuguese (which are quite common, and becoming more so), Bororo syllables never have onset consonant clusters or codas.

Stress in Bororo occurs generally (again with the exception of unmodified Portuguese loanwords) on the penultimate mora. Since diphthongs contain two morae, this means that a diphthong in either penultimate or final position will generally be stressed: [ˈbai̯ɡa] "bow", [kaˈnao̯] "pimple, scale". Before compounding, however, diphthongs are reanalyzed as a monomoraic syllable, and stress is assigned again to the penultimate syllable of the whole compound: [keˈrakae̯] "finger", from /kera/ "hand" and /kae̯/ "digit".

Morphophonology

Bororo has a process of stem-initial mutation.^[5] When a stem (whether a noun, verb, or preposition) beginning in a consonant takes on a prefix itself containing a consonant, the stem-initial consonant is voiced; for example, kodu "to go" becomes ikodu "I go" with the first person singular (prefix i-) but pagodu "we go" in first person plural inclusive (prefix pa-). In addition, /b/ becomes /w/ and /dʒ/ becomes /j/ when following a prefix, regardless of whether the latter has a consonant or not: bogai "for" becomes iwogai "for me". Stems beginning with /d g m n r j w/ do not undergo any change. This process is part of a wider harmony restriction in Bororo, rare among the world's languages: no word (excluding clitics and compounds) may contain more than one voiceless obstruent.^[6]

Meanwhile, a stem beginning in a vowel requires the insertion of a "thematic" consonant to attach a prefix. The consonant used varies both by the prefix and the stem-initial vowel: the first-person singular, first-person plural exclusive, and third-person plural prefixes insert the thematic consonant -t-/-d- before stems beginning in a- or o-, while all other combinations (i.e. other personal prefixes or stem-initial i-; no native stems begin in e- or u-) insert -k-/-g-. For example, whereas ore "children" becomes i-t-ore "my children", iwögü "steal" becomes i-k-iwögü "I steal". A few irregular stems which would be expected to take -t-/-d- instead take -n-: oika "anxiety" becomes inoika "my anxiety", rather than *itoiga or similar. Finally, since vowel-initial stems never take a third-person singular prefix, no thematic vowel is inserted. This process is summarized in the table below.

Nouns

There is no category of obligatory nominal affixes in Bororo; a noun phrase may therefore be composed of a single unmarked root.

Verbs

Bororo verbal morphology is basically divided along lines of transitivity. Both transitive and intransitive verbs take the same set of bound pronouns to convey subjects and objects (e.g. i-reru-re, a-reru-re "I danced, you danced" / i-re a-reru-dö "I make you dance"), but each has an exclusive set of suffixes. "Intransitive" suffixes naturally apply to intransitive verbs, for which there is only one core argument – for example, the neutral aspect -re in i-reru-re – but also to the agent of a transitive clause, hence i-re a-reru-dö; meanwhile, "transitive" suffixes can only be applied to the main verb of a transitive clause – e.g. the causative -dö in i-re a-reru-dö. For this reason, the suffixes listed in this section cannot properly be called "verbal" suffixes alone; they are grouped in this section merely for convenience.

Pronouns

Personal pronouns are divided into free and bound forms. Free forms are used to strongly mark a topic (marked by the "fronting" suffix -re: homophonous with but not identical to the declarative re-): imi-re i-tu-re "(as for me,) I'm going". They are also used in "verbless" sentences (for which see below): ema-re-o "here it is / this is it". Bound forms, meanwhile, are used to mark verb subjects and objects, inalienable possession, and postpositional inflection. The full list of free and bound forms (but not alienable possessors, for which see "Possession" above) is shown in the table below.

1. Crowell (1979) references a first-person singular bound form yo- occasionally, but does not provide any description of this form or its use; it may be a typographic error. Nonato (2008) does not record it.
2. U- is used as the third-person bound pronoun as an agentive subject, as the subject of some irregular intransitive verbs (most commonly -tu- "to go") and as the inalienable possessor of some irregular nouns. Otherwise, third person singular is not marked.

Transitive and intransitive clauses

As seen above, a different set of suffixes applies to intransitive verbs and transitive subjects on the one hand, and verbs containing transitive objects on the other. It follows, therefore, that an intransitive clause must contain only a single verb of the former type (e.g. i-reru-re "you dance"), whereas a transitive clause must contain one of each. In the latter case, the "subject" word precedes the "object-verb": for example, a₁-re i₂-wiye "you₁ advised me₂". The use of bound pronouns is required even if the sentence contains a full noun in either subject or object position: ime e₁-re areme e₂-wiye, word-for-word "the men they₁-did the women them₂ advise" – "the men advise the women".^[9]

Since ambitransitive verbs are at least exceedingly rare in Bororo, however, and intransitives with postpositional focus are very common, the frequency of strictly "transitive" constructions is far fewer than that of English or Portuguese.

Yes–no questions

Yes–no questions ("did/do..." etc.) follow the same syntactical rules as declarative and imperative clauses in Bororo. Indeed, an interrogative clause may differ from a declarative one only by intonation: ature "you went", ature? "did you go?" However, a yes–no question may also be indicated explicitly by placing the clitic =na immediately after the questioned element: aturena? "did you go?" This clitic is mandatory after a question that contains only a single uninflected word, for example with pronouns: imina? "me?"^[10]

Informational ("wh-") questions

Bororo has six question words, all ending in the interrogative morpheme -ba: kabo-ba "what", yogüdü-ba "who", kakodiwü-ba ’which', ino-ba ’how', kai-ba ’where' and kodi-ba 'why'. If the question element is modified in some way (e.g. by a preposition), the following modifier must itself be marked with -ba: kabo₁-ba tabo₂-ba imedü maragodü-re "what₁ did the man work with₂?"

As this example illustrates, Bororo makes use of wh-movement. An exception to this rule is pa, an alternative form of kai-ba "where", which follows rather than precedes its subject: imedü pa? "where is the man?"

Copular clauses

Unlike in English, many well-formed Bororo clauses lack a verb entirely, in which case the first noun (the subject) takes the "intransitive-ergative" suffixes described above. These clauses usually correspond to those formed using the verb "to be" in English: pöbö-re wöe "there is water here", ema-re-o "here it is", etc. Crowell (1979) divides these clauses, which he calls "copulative", into "existential", "equative" and "identificational".^[11] Unlike clauses with verbs, these use free pronoun forms exclusively (with the exception of inalienable possession).

Existential clauses denote the existence of their subject, optionally qualified by time and place. These are formed simply by applying the relevant inflection to the noun phrase: karo-re "there are fish / fish exist", karo-re pöbö tada "there are fish in the water", etc. If the subject of these clauses is modified by possession, this is taken as a full possessive statement: i-ke-re, "I have food" (lit. "my food is").

Equative clauses denote an identity between two noun phrases. These are formed by the juxtaposition of two noun phrases, only the first of which is inflected: imedü-re imi, "I am a man". If these noun phrases are modified outside of inflection (e.g. are possessed by a full noun rather than a prefixed pronoun), they require the "complex equative" particle rema (singular) or remage (plural) to end the clause: Kadagare onaregedü-re Creusa rema "Creusa is Kadagare's child".

Identificational clauses are used to present an object within a particular space. These are formed identically to existential clauses, followed by the addition of one of a set of suffixes -o, -no, and -ce, which correspond directly to the deictic prefixes a-, no- and ce-: karo-re-o "here / this is a fish".