Pilar M. Valenzuela

Metsá Rama

Metsa Rama


"Noa shipibobo iki nonbiribi joiya".                                                        
"Nosotros los shipibos tenemos nuestra propia lengua".

Laqueando la ceramica

"El shipibo-konibo (en adelante, SK) es una de las cerca de cuarenta lenguas habladas actualmente en la Amazonía Peruana; en este número no se incluye a las lenguas aimaras ni a la mayoría de lenguas quechuas. Las lenguas amazónicas del Perú han sido agrupadas en catorce o quince familias distintas (dependiendo del estatus asignado al conjunto bora y huitoto). A esta gran diversidad lingüística debemos agregar dos lenguas que permanecen bajo la categoría de "aisladas" (Pozzi-Escot 1998).

El SK pertenece a la familia lingüística pano, que fue establecida como tal en la literatura a fines del siglo XIX mediante el trabajo de Raoul de la Grasserie (1890). El pano es una de las agrupaciones lingüísticas más conocidas de las tierras bajas de Sudamérica conjuntamente con las familias arawak, tupí, karib y tukano. Se ha estimado que actualmente unas veinticinco lenguas pertenecientes a esta familia serían habladas en las zonas fronterizas de los bosques amazónicos del Perú, Brasil y Bolivia" (Valenzuela, Pilar M. 2001).



The information given below is taken from Valenzuela, P.M. 2003, Transitivity in Shipibo-Konibo Grammar, University of Oregon Ph.D. dissertation.

                                                         TYPOLOGICAL FEATURES

Here I will briefly categorize SK with respect to a few major, well-known typologies and highlight some characteristic features; in this way, the reader can get a sense of the kind of language we are dealing with from the beginning.  I have also included some diachronic observations, a short list of negative characteristics (i.e., what SK does not have), and a summary of semantic and cultural points that may be of special interest to some readers.


2.1. Morphological Typology


SK can be characterized as a predominantly agglutinating language:  words tend to be plurimorphemic, there is usually a biunique relationship between form and function, and in most cases morpheme boundaries are easily identifiable.  However, it is common to have unmarked nominal and adjectival roots (since the absolutive case is generally zero and there is no agreement across NP constituents), as well as sequences of morphemes lacking clear-cut boundaries (e.g., certain nominal plus ergative, or root plus detransitivizer sequences).  In addition to this, there are a few instances of stem changes and suppletion.  Finally, the verb may carry one or more deictic-directive / adverb-type suffixes, in what can be described as a polysynthetic tendency.   

In terms of affixation, SK is suffixal, except for a closed set of noun roots that occur in a reduced form and may be prefixed to verbal, adjectival, and other nominal roots; this set of noun roots generally refers to parts of the body.  In addition, nouns ending in an open unstressed syllable form their vocative by shifting the stress to the last syllable; i.e., an instance of suprafix or suprasegmental marking. 

SK also has postpositions and different types of enclitics, some of which are more precisely “endoclitics” given that they may precede suffixes and thus occur word-internally.  The border between postpositions and enclitics is sometimes fuzzy; the same is true of serialized or compound verbs and verb plus suffix sequences.


2.2. Basic Constituent Order


As to main clausal constituent order, SK is of the basic head-final type.  The following strict word order correlations can be attested:  the possessor precedes the possessum and as mentioned above there are postpositions rather than prepositions.  Nevertheless, SK exhibits a great deal of word order flexibility not only with respect to major clausal constituents (except for nominalized and reference-marked clauses which are strictly verb final), but also within the noun phrase.  For instance, adjectives, some quantifiers, the numerals ‘one’ and especially ‘two,’ and relative clauses may either precede or follow their head nominal with no obvious semantic consequence.  Additionally, most auxiliaries can either precede or follow the semantically main verb.  The free intensifier kikin ‘extremely, truly’ always precedes the adjective and its modified head, but there are additional intensifier enclitics (namely -yora and -shaman); manner adjuncts and postpositional phrases precede or follow the verb; copulas precede or follow their predicate; and keen- ‘want’ precedes or follows its subordinate clause when functioning as a manipulation or modality verb.


2.3. Expression of Arguments


Verb arguments are expressed through free lexical case-marked nominals, with no coreferential pronominal marking on the verb or auxiliary (subject plural marking is obligatory when plurality is not expressed in the nominal).  Hence, it can be said that SK is a lexical argument language (as defined in Jelinek 1984) for both subject and object functions.  Omission of required subject and object is normally understood as a zero third person singular form.  There is no systematic morpho-syntactic means of distinguishing direct from indirect objects, or primary versus secondary objects. 



2.4. Head-Marking versus Dependent-Marking Features


The use of free case-marked lexical nominals and the absence of pronominal marking on the verb or auxiliary are characteristic of a dependent-marking type language.  Other dependent-marking features are the marking of nominal and pronominal possession on the possessor rather than on the possessum, the presence of absolutive or oblique case-marking on the objects of postpositions, and the use of inflectional morphology on certain adjuncts in agreement with the syntactic function of their controlling nominals.  As for relative clauses, these are non-finite, obligatorily verb-final, and may be incomplete in terms of arguments; all these are clear dependent-marking features which are also shared by other types of subordinate clauses. 

         Head-marking features are also found in SK; for example, a relative clause may be internally-headed.  In addition, certain types of negation are marked on the head and not on the dependent element of a constituent:  negation of indefinite pronouns and universal nominal negation (such as ‘no child came’) are marked on the verb rather than on the nominal; negation of attributes indicating what a nominal referent is made of (such as ‘not a wooden house’) is marked on the nominal rather than on the attribute.  Finally, the verb may carry adverbial information such as whether an action is performed all night or all day long, completely, again, by two participants, in a distributive fashion, up or down the river, from or towards a specific point set in the discourse, etc.  These last features may be interpreted as head-marking.  In attributive phrases both the noun and the modifying adjective occur unmarked. 


2.5. Case-Marking


SK exhibits a fairly rigid ergative-absolutive case-marking system.  The ergative case is marked by the enclitic -n that attaches to the last element of the corresponding NP; the absolutive is -a for some pronouns, and zero in all other cases.  Unlike the situation most commonly found in languages of this type, including most of those in the Panoan family, there are no instances of case-marking splits of the familiar sorts.  A possible exception is found with regard to emphatic pronouns, whose case-marking is governed in a nominative fashion.  In addition, desiderative clauses and clauses with overt progressive marking may mark A arguments as either absolutive or ergative depending on the pragmatic status of the O argument. 

The phrasal enclitic -n exhibits a rich allomorphy and represents an interesting instance of both case syncretism and polyfunctionality.  Besides the ergative, it also codes genitive, instrumental-means, locative-allative, temporal, and other oblique functions.  Pronominal paradigms can be said to be asymmetrical, since certain pronouns exhibit more case distinctions than other pronouns and nouns.


2.6. Participant Agreement


Certain adverbial-like words, phrases, and clauses (referred here as “adjuncts” since they are not arguments of the verb) are semantically oriented towards one core participant or controller and receive a marking in accordance with the syntactic function this participant plays.  Given that the agreement markers in question display an overall tripartite distribution, this feature can be analyzed as a type of split-ergativity which might be exclusive to Panoan.  


2.7. Clause-Chaining and Switch-Reference System


Having a preference for the combination of two or more clauses of different rank only one of which exhibits fully finite verb inflection, SK can be characterized as a clause-chaining language; in turn, the non-finite clauses in the chain carry same- or switch-reference marking (i.e., “reference-marking”).  Nevertheless, it is uncommon to find the lengthy chains typical of some Papuan languages (e.g., Wambon, de Vries ms.; Haruai, Comrie 1998).  In SK reference-marked clauses are strictly verb-final; usually lack an overt subject argument; carry no obvious nominalizing morphology; and may precede, follow, or be embedded in their matrix clause.  Furthermore, reference-marked clauses need not be adjacent to their matrix clause and two or more marked clauses may share the same matrix clause. 

Besides coding coreferentiality or non-coreferentiality of arguments and the relative temporal or logical order of the two events, same-reference markers may reveal the transitivity status of the matrix verb.  This is so, given that most same-subject markers are identical to the participant agreement morphemes mentioned above and hence correlate with the S or A function played by their controller in the matrix clause.  Same- or switch-reference marking may also be used to encode different types of discourse (dis)continuity.


2.8. Transitivity Agreement


There are other types of transitivity-associated phenomena in SK.  Phasal verbs  like ‘start,’ ‘finish,’ and ‘stop (doing something),’ for example, must agree in transitivity with their semantically main verb; a comparable situation is found with pro-verbs or auxiliaries, serialized verbs, certain verbal suffixes, and suppletive verb stems. 


2.9. Split-Ergativity


As mentioned above, SK has a fairly consistent ergative-absolutive case-marking system.  Syntactically, imperative clauses and internally-headed relative constructions also exhibit this type of alignment; i.e., only NPs in S/O function may be overtly expressed in the former, or be interpreted as coreferential with a matrix clause argument in the latter.  On the other hand, case-marking on emphatic pronouns, the plural verb suffix -kan, and most interclausal control properties operate in a nominative-accusative fashion.  As mentioned above, a tripartite distribution is attested with respect to the participant agreement morphology; in addition, the overall shape of emphatic pronouns also shows distinct forms when in S, A, or O functions.   


2.10. Evidentiality


The coding of evidentiality takes place at two different levels.  First, a major distinction between first-hand information and second-hand information is established.  Upon this basic distinction, two other markers may be added to indicate inference or speculation. Evidentiality is not a verbal category in SK but is expressed through clitics generally appended to the first major sentence constituent; however, if the first major constituent is a finite verb, evidentials precede the plural and finite aspectual morphology.  In narratives, the direct and reportative evidentials do not occur on every clause or sentence, but are used to mark certain discourse units.  Reported imperative expressions are possible in SK to transmit someone else’s orders or requests.  Evidentials have epistemic and mirative functions. 


2.11. Other Salient Features


The phoneme /ö/ is one of the four vowels SK has; this segment has been posited as one characteristic feature of Amazonian languages (Aikhenvald and Dixon 1999:8).  Probably the most interesting phoneme is the retroflex approximant 

/?/, which is a highly variable segment.  Some morphemes exhibit alternate allomorphs depending on whether they attach to an element with an even or an odd number of moras.  This phenomenon has been called “alternate mora-timing” (Lauriault 1948).  However, the answer as to whether codas or only nuclei should count as moraic in SK is not straightforward as can be drawn from the examination of stress patterns and morpho-phonological distributions.

SK exhibits several instances of case-stacking, so that a nominal may receive up to four (reconstructible) case-markers.  Ablatives are always formed upon locatives.

A few verbal suffixes indicate specific tense distinctions such as different degrees of past and a ‘tomorrow’ future.  Other interesting verb categories are those coding deictic-directive / adverb-type meanings.  With regard to valency-changing mechanisms, SK has a few causativizers encoding different degrees of force-dynamics, three applicatives one of which is restricted to transitive stems, a middle marker (with reflexive, anticausative and sometimes passive senses), and a dedicated reciprocal.  Verb stems can contain body-part prefixes, as well as onomatopoeic roots which combine with transitive and intransitive semantically generic verbs.  Short answers to polar questions are interesting since they involve the semantically generic verbs functioning as pro-verbs, as well as same-subject marking.  First person plural inclusive exhortatives and negative imperatives are biclausal.  Relative clauses can be prenominal, postnominal, circumnominal, or adjoined. 


2.12. Diachronic Observations


Certain verbs have become or are in the process of becoming bound verbal modifiers while simultaneously retaining their former function; e.g., the negative existential verb yama- is also the negative verbal suffix, the transitive verb pake- ‘cause to fall’ may be appended to another transitive verb to code distributive action, the verbs ‘come’ and ‘go’ seem to have combined with a second marker to become venitive and andative suffixes, and the verb toshi- ‘pop up, burst’ may be added to another verb stem to encode a sudden event.  As mentioned above, when added to a finite verb, second position clitics obligatorily precede the plural -kan and aspectual morphology, thus suggesting a relatively recent grammaticalization of the plural and aspect markers.  There is identity in form and semantic compatibility between applicatives and same-subject markers; this may be due to a diachronic connection (Valenzuela 1999).  A common origin is also posited for certain pairs of formally identical clauses, which nevertheless exhibit distinct finite versus nominalized syntactic status.  There are a few periphrastic finite verb forms such as the “narrative past,” whose structures reveal an obvious nominalized origin.

Some postpositions have developed from nouns denoting body parts to which an oblique marker has been added; e.g. pekáo ‘behind’ < peká ‘back’ + the approximative locative -o.  The comitative postposition betan has developed into a noun phrase conjunction while retaining its former function.  A shortened version of the comitative is also found in combination with the prefix ra- ‘body, person’ to form the numeral rabé ‘two.’  The intensifier -yora probably developed from yora ‘person, body’ (see also Semantic and Cultural Observations below).

Participant agreement on adjuncts exhibits traces of an older tripartite case-marking system; Chapter 20 provides a synchronic and comparative analysis of this typologically salient feature.  Reference-marked demonstratives have developed into interclausal and intersentential conjunctions.  The medial demonstrative ja has become the third person singular pronoun, and may also function as definite article, whereas the numeral ‘one’ may be used as indefinite article.  


2.13. Negative Characteristics


SK has no gender or noun class categories, no classifiers, no obvious alienable versus inalienable possession distinction, no first person plural inclusive versus exclusive distinction (but see hortatives), no dual number (although the verb suffix -bekon indicates action carried out by two participants), and no honorific pronouns.  As for verbs, these do not exhibit cross-referential pronominal morphemes.

Another negative characteristic is the absence of voice mechanisms of the canonical sort, such as agentive passive and antipassive. 


2.14. Semantic and Cultural Observations


Most perception, cognition and utterance events are morpho-syntactically encoded through prototypical transitive clauses.  Emotions are generally expressed via extended intransitive clauses, with the experiencer coded as S and the stimulus argument coded as oblique.  Forces, illnesses and other inanimates depicted as affecting other entities behave morpho-syntactically as highly transitive subjects.  For example, the only way to express the equivalent to the English ‘drown (intr.)’ is through a transitive construction involving the verb ‘kill,’ with the flowing water coded as A and the affected participant coded as O.  In turn, to express the equivalent of the English ‘drown (tr.)’ a causative construction with the literal translation ‘X caused the flowing water to kill Y’ would be required (see exx. (8)-(9) in Chapter 16).  The same holds for ‘to die suffocated’ or ‘to die from vomiting’ which involve the nouns join ‘air’ and kinan ‘vomit,’ respectively.  Also, body parts are encoded as transitive subjects in order to express the meaning ‘ache.’

Compounding is a very productive process.  Animal and plant names at the generic biological level tend to be monomorphemic, while most categories at levels below the genus are compounded.  It is common to employ the name of a highly salient animal or plant as a modifier that refers metaphorically to a distinctive characteristic of the species.  For example, ino-rono ‘jaguar-snake’ is a kind of snake whose head is said to resemble that of a jaguar (Valenzuela 1998b).  Traditional Shipibo proper names are also compounded.  As a general rule, the first element is common to women and men, and at least some may be related to former clan names; the second element is usually specific to each sex.  Ethnonyms are generally based on animal names; e.g., Kashibo [kaSi ‘bat’ + -Bo collective / plural] ‘Bat People.’  

Onomatopoeia, often in combination with the intransitive and transitive semantically generic verbs or auxiliaries, is a common lexicalization strategy.  This may be used in the coining of new vocabulary; e.g., to’-a-ti [ONOM-do.T-INF] ‘shotgun’ (lit. ‘sth. to make “to” to so./sth. else’), to’ ak- [ONOM-do.T] ‘kill with a shotgun,’ to’ ik- [ONOM-do.I] ‘kill oneself with a shotgun.’ 

The adverb bakish is best translated as ‘one day from today’ since it corresponds to both ‘yesterday’ and ‘tomorrow.’  There are two different words for English ‘water;’ while jene refers to flowing water, onpax refers to contained water.  Therefore, when bathing in the river one must take care that little children do not swallow jene, but one can only drink a glass of onpax.  Jene also means ‘sap’ or ‘juice;’ when preceded by xoma ‘breast,’ it means ‘milk.’

The derived noun piti is composed of the root pi- ‘eat’ and the irrealis nominalizer or infinitive suffix -ti; therefore, piti literally means ‘food.’  However, piti is primarily interpreted as ‘fish,’ which is the Shipibo food par excellence.  Consider the following excerpt from Valenzuela (2000a:4) (see also 1.3.6):  “The impressive resources of the major rivers and adjacent ox-bow lakes constitute the principal source of protein for the Shipibo, for whom “a day without fresh fish is a day of starvation” (Lathrap op.cit., p.19)”.  There is an additional term, yapa, which means ‘fish.’    

Color terms are joxo ‘white,’ wiso ‘black,’ chexe ‘sharp (color), black,’ yankon ‘blue / green,’ panshin ‘yellow,’ joshin ‘red’ (also ‘ripe’).  Note that the last two terms share the ending -shin.  An additional term, emo, is best translated as ‘faded white, colorless’ (like an old piece of cloth that has lost its original color).  SK is relatively rich in categories describing olfactory perception; some ten different terms describing kinds of smells have been identified (Tournon, p.c.).

Native numerals are only westíora ‘one’ and rabé ‘two;’ further on Quechua numerals have been borrowed.  Westíora might derive from westí ‘few, sparse’ + yora ‘body, person’ > intensifier suffix; and as mentioned above, rabé derives from ra- shortened form of ‘body, person’ + -bé shortened form of the comitative beta(n) also attested on pronouns.

The terms wetsa and poi refer to ‘same-sex sibling’ and ‘opposite-sex sibling,’ respectively; i.e., wetsa designates a man’s brother or a woman’s sister, and poi a man’s sister or a woman’s brother.  The term rayos means either a man’s parent-in-law, or a man’s or woman’s son-in-law but not daughter-in-law.  According to Shipibo culture, relatives holding the relationship of rayos must not speak directly to each other, but communicate through their daughter or wife.  This socio-cultural behavior triggers the relatively frequent use of jussive constructions (e.g., ‘May your husband bring firewood’ instead of ‘Bring firewood!’).  The term chichi means both ‘maternal grandmother’ and ‘highest authority.'

Web Hosting Companies