AUTHOR: Olawsky, Knut J.
TITLE: A Grammar of Urarina
SERIES: Mouton Grammar Library 37
PUBLISHER: Mouton de Gruyter
Carolina González, Department of Modern Languages and Linguistics, Florida State
This book is a comprehensive study of Urarina (previously known as Shimacu or
Tucale), an Amazonian language spoken by between 2,000 and 3,000 speakers in the
area of Rio Chambira (Loreto, Peru). Since its lexicon and many of its
grammatical characteristics differ from those of neighboring languages, it is
considered a language isolate. As other languages in the area, it is endangered
by the inroads made by Spanish, which is impacting its lexicon and some aspects
of the phonology.
Olawsky's description of the language is based on his fieldwork in Nueva Unión,
a remote area with about 100 speakers of Urarina, for over five years. It
includes data from 126 recorded texts, in addition to elicitation from native
Urarina speakers and New Testament translations. Statistics regarding the
frequency of sounds, word classes and grammatical constructions come from a
lexical database of 3,650 entries compiled by the author.
The book is over 900 pages long and divided into 23 chapters. Chapter 1
introduces Urarina, the scope and methodology of the investigation, and some of
the unusual typological characteristics of this language. These include its
OVA/VS constituent order, the existence of a three-way distinction for person
marking in all verbs, and the importance of word length in its phonology and
morpho-phonology. Chapters 2-4 describe the phonology and morpho-phonology of
the language, and chapters 5-21 its morphology and syntax. The two last chapters
are devoted to discourse strategies and linguistic variation, respectively.
Urarina has 16 consonantal phonemes, five phonemic vowels, and four diphthongs
(Chapter 2). Long vowels, although rare, are contrastive, as are nasal vowels.
The most common syllable structure is CV. Onsets are optional, and codas are not
allowed, except for [h]. Monosyllabic words are attested, and most of them have
long vowels or diphthongs. However, most words are polysyllabic, and there is a
tendency for nominal and verbal roots to be three syllables long. Phonological
phenomena include rightward nasal assimilation, which targets vowels and [h];
vowel copying; post-vocalic aspiration before voiceless and lateral consonants,
and insertion of the glottal stop before two identical vowels when the first
vowel is short and the second long.
Chapter 3 discusses some aspects of the morpho-phonology of Urarina. Several
morphemes undergo alternations between vowel initial and onset initial
allomorphs, which tend to be velar stops or palatal affricates, or
shortening/lengthening of vowels across certain morphological boundaries. One
interesting characteristic is that several alternations seem sensitive to word
length and syllable weight. For example, the nominal plural marker –kuru occurs
after trisyllabic roots, and the allomorph –uru elsewhere. However, –kuru is
also found in monosyllabic and disyllabic roots with a CVV syllable, and also
occasionally on quadrisyllabic roots.
Chapter 4 outlines the major tonal features of Urarina, which is described as a
pitch accent language contrasting high and low tone. Each morpheme has maximally
one tone, and most final syllables have high tone. Nouns fall into four tonal
classes, depending on the tonal effect they have on the verb, adjective or
postposition to their immediate right; verbs fall into three different classes.
Tonal classes are partly based on word length and syllable structure. Tone is
also connected to syntactic structure; for example, each constituent tends to
have only one high tone. Stress is not analyzed in detail, but two facts noted
are that it correlates with higher amplitude/intensity, and that syllables with
long vowels attract it.
Chapter 5 introduces Urarina word classes, considered in detail in subsequent
chapters. Urarina is a verb-prominent language, which is typologically unusual
(over 55% of words are verbs and 30% nouns). Verbs differ from nouns in that (i)
they take arguments (ii) they are inflected for many categories, not just person
and number, and (iii) their citation form ends in /a/, while that of nouns can
end in any vowel. Verbs and nouns are linked by derivational processes of
verbalization and nominalization, as discussed in chapter 10.
Demonstratives and pronouns are subtypes of nouns. Demonstratives indicate three
degrees of spatial reference and can be used as nominal heads. They can combine
morphologically with other elements that indicate location and direction.
Pronouns, marked for person and number, are used for emphasis and are optional.
Adverbs refer to the categories of time, place, manner, degree and epistemics.
Except for time adverbs, they are derived from verbs through the participle
suffix, and generally, they precede and modify verbs. Adjectives are always
derived and do not seem to form an independent class.
Other categories in Urarina include postpositions (usually indicating location
or temporal concepts); conjunctions (subordinate or coordinating, and always
clausal or sentential); interrogatives; and clause introducers (of clause type,
negation, or mood). Since clause introducers co-occur with verbal morphology
they seem to form a distinct word class, which is typologically unusual.
Quantifiers, interjections and ideophones (onomatopoeic words) are also
discussed in this chapter.
The morphology of Urarina is complex, and is mostly nominal or verbal. However,
it is somewhat flexible and not restricted to specific word classes. For
example: multipurpose words are attested; some verbs are able to express
semantic concepts not typically associated with verbs in other languages
(enumeration, the verbs 'be like that', 'be in vain'), and morphology used for
nouns or verbs can be assigned to other word classes.
Chapter 6 covers the structure of noun phrases (NPs), and chapter 7 possession.
There is a tendency for demonstratives, possessives and numerals to precede the
head noun, while other types of modifiers, such as adjectives, can precede or
follow it. Possessive pronouns and proclitics are common within the NP. The
possessor precedes the possessed, and there is an optional possessive marker
between both. At the clause level, the verb 'have' indicates closer or long-term
possession, while the copula, used with free or bound pronouns and followed by
the possessive marker –raj, indicates short-term position. Affixes indicating
possessive derivation are attested but rare.
Urarina seems to be in a transitional stage between synthetic and analytic use
of possessives; a trend towards free pronouns can be observed. The distinction
between alienable and inalienable possession (for body parts and kinship terms)
is still maintained in the traditional language, but it is being lost in the
Chapter 8 focuses on number marking. Interestingly, the unmarked number seems to
be the non-singular. Two different plural suffixes are used: one for most verbal
forms, and another for nouns, postpositions, and verbs in the third person
singular. Agreement between noun and verb is not obligatory, and plural marking
tends to be optional, especially in the third person. Within the NP, nouns
referring to humans are more likely to be marked for plural than other nouns, or
than numerals. The first person plural is marked verbally for the exclusive,
inclusive and dual; the last two tend to be neutralized in favor of the dual. In
intransitive verbs indicating size, color and shape, a distinction is made among
singular, dual, paucal, and plural.
Verb classes are taken up in chapter 9. Urarina verbs can be transitive,
intransitive, and reflexive (discussed in chapter 16). There are no
ditransitives, and the only ambitransitive is the verb 'to burn'. The copulative
verb ''nia'' indicates attribution, location, existence or temporary presence, and
even possession (with the possessive marker –raj). Verbless clauses occur in
questions involving demonstratives. Special verbs include interrogative verbs,
enumeratives, and the verb 'to be in vain'.
Transitive verbs have an obligatory object, unmarked in the third person.
Intransitive verbs can be active, stative, or referring to position, shape, and
color (PSC verbs). PSC and 'affect' transitive verbs -which refer to processes
that reduce or change the shape, integrity or size of an object- follow a
special derivational system for size (big or small) and various postures, shapes
or colors. Stative verbs have a special plural suffix.
Word formation (chapter 10) is as complex as in other Amazonian languages (p.
430). Derivation changes word class and mainly results in nominalization of
verbs and verbalization of nouns. The only exception is derivation used to
create complex verbs. Five types of total reduplication are attested in verbs,
with functions ranging from aspectual interpretation, spatial distributive, and
motion. Some types are more common with PSC and 'affect' verbs and are
distinguished for big and small items. Compounding, which always involves nouns,
can be of nine types, each semantically different.
Urarina boasts a distinction of three person inflection classes, named
'D-forms', 'E-forms', and 'A-forms' (chapter 11). These inflection classes are
expressed morphologically by separate sets of person markers (especially for
first/third person singular) and by differences in mood and polarity. 'D-forms'
are obligatory in finite dependent clauses. 'A-forms', so called because the
third person singular is –a, are used in greetings, citation forms, polar
questions, and with the negative introducer ''kwatia'' 'don't'. 'E-forms', so
called because the third person singular is –e, are used in sentences with focus
markers, and are also typical in narrative styles, and after or before dependent
Chapter 12 covers verbal morphology. Urarina is a highly polysynthetic language
with a rich morphological system for verbs. Only one prefixal position occurs,
which can be occupied by the intransitivizing prefix ne– or the 1st /2nd
singular object proclitic. In contrast, 17 suffixal slots are identified: (1)
causative (intransitive), (2) causative (intransitive/transitive), (3)
impossibility, (4) continuous aspect, (5) impersonal passive, (6) habitual
aspect, (7) distributive/plural object, (8) velocity/politeness, (9)
diminutive/counter-expectation, (10) completive aspect, (11) in-law talk
(politeness), (12) plural (1st person), (13) epistemic modality/probability,
(14) mood/future tense, (15) negation, (16) person, and (17) 2nd person plural.
The maximum number of filled slots attested in the language is six, although on
average only between two and four verbal positions are filled.
Verbal slots are followed by enclitic positions, which are unstressed and can
attach to non-verbal words. Seven enclitic positions are identified: (18)
politeness, (19) assertive, (20) witness evidential, (21) reportative
evidential/remoteness, (22) reassurance, (23) negative question/ interrogative,
and (24) attitude/emotion/rhetorical question.
Politeness in Urarina (chapter 13) is gender/family related, and does not depend
on context. It is obligatorily expressed in conversations between members of the
opposite sex or between male in-law, and in reference to people within specific
in-law relations. The latter, expressed with the verbal suffix –ana, is
typologically rare. Other expressions of politeness are optional.
Negation is expressed grammatically with a varied range of strategies (chapter
14). Negative verbal suffixes have different allomorphs for different moods.
Negative interjections, conjunctions and clause introducers are also common.
Some of these clause introducers involve both a clause-initial particle and a
final negative marker; the latter are incompatible with negative inflection.
Interrogative pronouns are used negatively as indefinite pronouns or negative
quantifiers. The negative copula expresses an adverbial function, meaning 'not
at all' when it co-occurs with a negative verb. The scope of negation is
clausal, and several negative strategies can be employed in the same clause,
which reinforces negativity.
The imperative is a separate clause type in Urarina (chapter 15). Also marked
for number and person, it has different forms for the 2nd person (positive
imperative), 1st person singular (hortative) and 3rd person (jussive). The
imperative cannot occur in dependent clauses; but a clause that follows an
imperative clause might be interpreted as dependent and consecutive to it.
Valency changing mechanisms are discussed in chapter 16. Among the suffixes that
may modify the valency of a verb are causatives: –a, which implies direct
personal involvement of the causer, is not very productive and occurs in
intransitive verbs only, while –erate, which implies indirect causation, is
productive and may apply to both transitive and intransitive verbs. No separate
passive form occurs in Urarina; rather, the passive is expressed via a
nominalizer suffix on transitive verbs. The agent may occur before the
passivized verb with no formal marking; constituent order indicates its
function. Other valency changing suffixes are the reciprocal –ita, the
intransitivizer –ne, and the valency increaser particle –ke.
Serial verb constructions (chapter 17) are one-predicate structures that
indicate direction or movement. Typically, they convey a point of change
regarding a previous event, or an emphasis on motion or movement towards/away
from the point of reference. They are formed by two adjacent verbs, both sharing
the same subject and referring to one event (or part of an event). The first
verb is the semantic head of the construction, is marked with the neutral marker
–a, and can only host causative and continuous aspect morphemes. The second verb
indicates the direction or result of the event, hosts all other verbal morphemes
in the construction, and is typically the verb 'come', 'go', or 'wander around'.
Urarina has a typologically rare constituent order: OVA in transitive clauses,
and VS in intransitive ones (chapter 18). Since the O or S arguments are
frequently omitted and are interpreted from context, especially in dependent
clauses, the most common order is V (58%). The OVA constituent order occurs in
21% of cases, and VS in 10% of cases. Other clause types (AOV, SV) make up 7% of
cases; generally these involve focused structures and special forms such as
negation and content questions.
Adjuncts are optional and usually located in the periphery of main clauses.
Post-positional phrases tend to occur before main verbs; in dependent clauses,
they are only attested pre-verbally. Adverbs usually occupy the clause-initial
position. Except for temporal adverbs, they don't require a focus marker.
Locational adjuncts are also optional and occur pre-verbally with verbs like
'go' or 'arrive', as do regular O arguments. Dependent clauses are usually
placed before main clauses, unless they have a purposive or consecutive function.
Prominence is marked in Urarina by focus (chapter 19); its most common function
is contrastive focus, although topicality may also occur. Focus marking always
involves the 'E-forms' on the verb, and an enclitic indicating co-reference with
the subject: –ne or –na for the first person, and –te for all other persons. The
enclitic attaches to an emphasized constituent in sentence-initial position, and
fronting applies if the constituent is not sentence-initial. The focus enclitic
carries high tone in most contexts and is typically followed by a pause.
There is one focus marker per clause, and it is rare to find more than one focus
marker per sentence. In principle, all constituents can be focused, although
objects and serial verb constructions rarely are. The focusing of dependent
clauses indicates special emphasis. In some instances NPs are focused with
adjuncts, adverbs, or conjunctions; Olawsky suggests that in this case, both
constituents share topic and focus.
Multiclause constructions are discussed in chapter 20. Subordinated clauses,
which generally appear before the main clause, involve D-marking and can host a
limited number of suffixes and enclitics. They can be temporal/conditional,
complements, or participial. Complement clauses take the place of the O argument
and are marked with the infinitive suffix –na (if the subject coincides with the
main clause) or with –ne. Although both clauses can be introduced by verbs of
linking, thinking, knowing, be afraid, –na clauses also occurs with verbs of
ability and beginning/ending, and –ne classes with perception and saying verbs,
and with 'must'. Other complementation strategies include the use of the
complementizers, quotatives, and juxtaposition of questions.
The participle clause is the most common type of dependent clause. It is
non-finite and shares the same subject as the main clause. It is not restricted
to any specific verb or class of verbs, and morphologically, it is marked with a
suffix. Semantically, it is independent, and can have a range of functions,
including adverbial modification, sequencing of events or actions, and
overlapping events, actions or states. Although clause coordination is attested,
it is not very common. Other coordinating strategies include ellipsis, finite
clause juxtaposition and the use of adversative conjunction 'pero', loaned from
Questions are discussed in chapter 21. Content questions involve the 'E-form'
and a fronted interrogative pronoun. Polar questions require the 'A-form', have
a declarative constituent order, and are marked intonationally by a pitch rise
towards the end of the question. Negative questions are introduced with special
clause-initial forms; negative, clause final markers also occur. Rhetorical
questions indicate surprise or curiosity, and are marked with the 'A-form' and a
sentence-final enclitic. Indirect questions do not differ from direct questions.
They are expressed as independent clauses, juxtaposed to the main clause, and
have the same constituent order of the main clause.
Chapter 22 outlines the main discourse strategies used in narratives.
Phonologically, ideophones and consonant gemination are used to emphasize
dramatic events; extra lengthening of long vowels conveys intensity.
Morphologically, the 'E-form' is used, due to the abundance of clauses with
focus markers. Reduplication and triplication of verbal roots are used to convey
liveliness. Syntactically, ellipsis of arguments and quotative verbs is common.
Verbs can be repeated iconically to convey emphasis on the continuity of the
action or the event. Long sentences with many dependent clauses are also common,
as well as clause chaining, where events are related to each other by chaining
the last and the first clause of adjacent sentences.
Reference tracking is problematic in discourse because Urarina lacks any formal
marking for switch-reference and has no syntactic pivot. Together with omitted
arguments, this results in ambiguity. In discourse, switch reference can be
indicated by (i) the insertion of the relevant NP, which is rare; (ii) the use
of the demonstrative ''nii'' instead of an NP; (iii) subject marking on the verb;
(iv) the context; (v) certain conjunctions, and (vi) ideophonic discourse
particles, especially when the quotative verb ''naa'' is omitted.
The final chapter discusses dialectal variation and some of the differences
between traditional and innovative language. Four dialectal areas are
distinguished: Western dialects (including Tigrillo and Espejo), lower Chambira,
Upper Chambira (the largest dialect area), and the Corrientes River. The main
differences among these dialects, all of them mutually intelligible, are
phonological and lexical. Olawsky makes a distinction between traditional vs.
innovative language; the first is not used by younger speakers. The innovative
language tends to neutralize certain phonemic contrasts and leans towards a more
analytic morphology. It lacks some attitudinal markers still used in the
traditional language, and is also losing the contrast between alienable and
inalienable possession. Syntactically, it shows the use of other types of
subordinate clauses, involving non-finite verb forms of verbs with temporal
The book includes several appendixes. Appendix A glosses five texts, which are
available in audio form in the enclosed CD-Rom. Appendix B is a list of
scientific names, and appendix C includes some pictures of the Urarina. A list
of references and an index follow.
This is a comprehensive description of the phonology, morphology and syntax of a
little known Amazonian language, and should be of interest to typologists and
linguists interested in Amazonian and Native American languages. Olawsky's
description of Urarina is objective, clearly written, and extremely detailed,
and it provides multiple examples of all sounds, morphemes, word classes and
syntactic constructions discussed. Although the description may seem daunting
(it is over 900 pages long), the organization and clarity of the writing make it
more manageable. Each chapter outlines the main topics, and then discusses each
in detail with the help of glosses, summary tables, and statistical information.
Cross-reference also helps the reader to refer to other sections or chapters
were more or related information is given about a particular topic.
Perhaps one of the most intriguing issues that Urarina poses is its relationship
to other Amazonian languages. Although previous authors classify Urarina as
Panoan, Tupian, Macro-Tucanoan, or Andean (see Dean 1994 and references
therein), Olawsky considers Urarina an unclassified language, based on its
unusual typological characteristics and differing lexicon from neighboring
languages. One intriguing feature of Urarina not discussed in this book is the
similarity of its phonology and morpho-phonology to that of Panoan languages.
Urarina has various phonological and morphophonological phenomena that Olawsky
describes as partly dependent on word length; in Panoan, several phenomena seem
based on syllable count (Loos 1999). For example, Urarina has an optional
process of postvocalic aspiration that resembles that described for Huariapano
(aka Panobo, Gordon 2005) in Parker (1994). Panobo was once spoken close to
Urarina (p.8). In both languages, aspiration occurs before voiceless consonants;
in Urarina, it may occur before laterals as well. Aspiration occurs on non-high
toned syllables in Urarina, only once per root, usually in the first syllable.
It is in no case attested in two syllables in a row. Similar restrictions are
found in Huariapano, where aspiration occurs in odd-numbered syllables counting
from the beginning of the word (Parker 1994). Hopefully, more research will help
elucidate the extent to which Urarina is connected to Panoan and other languages
in the Amazonian region.
Dean, B. 1994. The poetics of creation: Urarina cosmogony and historical
consciousness. _Latin American Indian Literatures Journal. Review of American
Indian Texts and Studies_, Vol 10 (1): 22-45.
Gordon, Raymond G., Jr. (ed.), 2005. _Ethnologue: Languages of the World_, 15th
edition. Dallas, Tex.: SIL International. Online version:
Loos, E.E. 1999. Pano. In R.M. W. Dixon and Alexandra Y. Aikhenvald (eds.), _The
Amazonian languages_. Cambridge: Cambridge University Press, 227-250.
Parker, Steve. 1994. Coda epenthesis in Huariapano. _International Journal of
American Linguistics_ 60 (2): 95-119.
ABOUT THE REVIEWER
Carolina Gonzalez (Ph.D., Linguistics, University of Southern California) is an
Assistant Professor of Spanish and Linguistics in the Department of Modern
Languages and Linguistics at Florida State University. She specializes in
phonetics and phonology and is interested in rhythmically-based consonantal
phenomena, Amazonian languages (especially Panoan), and Spanish dialects.