The Arawak language family is the largest in South America in terms of its geographical spread, from Central America (Belize, Honduras, Guatemala, and Nicaragua) to as far south as Bolivia (and formerly Argentina and Paraguay). Within South America, Arawak languages are spoken in Lowland Amazonia and adjacent regions, covering Guyana, French Guiana, Surinam, Venezuela, Colombia, Peru, and Brazil, in at least ten locations north of the River Amazon, and at least ten to the south of it. There are over forty extant languages and a few dozen extinct ones. The genetic unity of Arawak languages was first recognized by Father Filippo Salvadore Gilij as early as 1783, based on a comparison of pronominal prefixes in Maipure, an extinct language from the Orinoco Valley, and in Moxo from Bolivia. The limits of the family were established by the early 20th century. Proposals to include Arawak languages in putative macro-groupings such as “Arawakan” or “Macro-Equatorial” have proved spurious and unsubstantiated. The heritage of Arawak languages survives in such common words as hammock, hurricane, barbecue, guava, and tobacco.
Arawak languages are synthetic, predominantly head-marking and suffixing, with a closed and historically stable set of prefixes—bound pronouns on verbs, the relativizing prefix ka- and its negative counterpart ma-. Personal prefixes distinguish first, second, and third person, and also impersonal and indefinite forms. Prefixes mark the subject of a transitive verb and of an intransitive active verb, and the possessor on nouns. In at least two thirds of the languages, personal suffixes or enclitics express the object of a transitive verb (o), and the subject of stative verbs (s
o) or the subject of non-verbal predicates. A few highly synthetic languages (including those from the Kampa subgroup in Peru) employ suffixes or enclitics to cross-reference the object and also the recipient or an oblique. There is typically a number of locative cases which can be stacked in one word. The majority of Arawak languages do not employ cases for marking core grammatical relations. The only exception is Tariana, from the multilingual Vaupés River Basin linguistic area. Here, core cases were developed under the influence of the neighboring Tucanoan languages. Inclusive–exclusive distinctions were developed in Resígaro and Palikur as a result of language contact.
Open classes are verbs and nouns; adjectives tend to form an open class, and share some features with nouns, and some with verbs. Verbal roots tend to be exclusively monosyllabic. Noun roots can contain more than one syllable. Derivational processes include affixation, compounding, and various kinds of reduplication. Just a few languages have single-word serial verb constructions. The order of suffixes within a word can be variable, reflecting the scope of the morphemes.
Nouns divide into obligatorily, or inalienably, possessed and optionally, or alienably, possessed. Obligatorily possessed nouns are body parts, kinship terms, and a few important possessions, for example ‘name’ and ‘house’. If the possessor is not specified, these nouns take an unpossessed form marked with a suffix, also used as a nominalizer on verbs in many languages. Alienably possessed nouns take a possessive prefix and an additional suffix (chosen based on the meaning of the noun). Most languages distinguish masculine and feminine genders in third person singular personal pronouns, demonstratives, nominalizations, and also as agreement markers on adjectives. More than half the languages have complex systems of classifiers on number words, and also on verbs, in possessive constructions, and on nouns themselves. They categorize the noun in terms of its shape, consistency, and animacy. Singular and plural numbers are fairly uniform across the family; dual has developed in Resígaro, as a consequence of language contact with the unrelated Bora. Other nominal grammatical categories include nominal tense, augmentative, diminutive, and approximative.
The verb is the most complicated part of the grammar of every Arawak language, and the only obligatory constituent in a clause. Typical verbal categories include tense, aspect, evidentiality, numerous modalities (including a frustrative meaning ‘do in vain’), and valency-changing derivations—passives, reflexives, reciprocals, causatives, and applicatives. Some Kampa languages have up to six applicative derivations, including comitative, benefactive, goal, presential, separative, and instrumental. Highly synthetic languages, such as Kampa and Palikur, have patterns of noun incorporation.
Many Arawak languages are located next to speakers of languages from other families. They take on their features, in grammar and sometimes also in lexicon. Tariana, the only Arawak language spoken in the multilingual Vaupés River Basin area surrounded by Tucanoan languages, has a distinct Tucanoan flavor to its grammar. Mawayana, Garifuna, and Palikur, in contact with Carib languages, have acquired a few Carib features. Resígaro has been affected by Bora, and Amuesha bears traces of contact with Quechua and other languages that are hard to identify. The interaction of genetic inheritance, language contact, and independent innovations makes Arawak languages dauntingly diverse.