|Date: Sat, 14 Feb 2004 06:06:02 -0800 (PST)
From: Patrycja Jablonska
Subject: A Grammar of Tariana
AUTHOR: Aikhenvald, Alexandra
TITLE: A Grammar of Tariana
SERIES: Cambridge Grammatical Descriptions
PUBLISHER: Cambridge University Press
Patrycja Jablonska, University of Tromsoe/CASTL
The book consists of 26 chapters plus Appendix on the dialects of
Tariana, original texts, Vocabulary list and a useful index of authors,
languages and subjects.
The first chapter contains basic information about the language and its
speakers. Tariana is a polysynthetic language of the Arawak family,
spoken by around 100 people in the linguistic area of the Vaupes river
basin in the territory of the Upper Rio Negro in northwestern Brazil.
Due to multilingualism ubiquitous in the area, Tariana combines the
features inherited from Proto-Arawak (e.g. head-marking in the sense of
Nichols 1986) with influences from East Tucano languages (e.g.
dependent-marking). Bibliography on the language is provided, as well
as the description of the social organization, ceremonies and beliefs.
Every informant is described individually and the source of materials
Chapter 2 is devoted to the phonology of Tariana. It starts with issues
of segmental phonology, including an interesting feature concerning
variation in phonotactic restrictions depending on the morphological
status (i.e. roots vs. affixes vs. clitics). Syllable structure, stress
assignment, basic phonological processes, and intonation patterns are
also described. Morphemes are divided into classes according to their
Chapter 3 is a rudimentary description of open and closed word classes.
Category-internal subgrouping of verbs, nouns and adjectives is
introduced, together with the type of morphology each of the categories
takes. Semi-closed classes (i.e. manner adverbs and time words) and
selected closed classes are also briefly reviewed.
Nominal morphology is investigated in greater detail in chapter 4,
where the polysynthetic character of nouns manifests itself in
immensely complicated nominal derivation and inflection (up to 16
slots). In particular, the property of double marking of grammatical
functions (case markers and noun classifiers) at different levels of
embedding results in the phenomenon of endoclisis (cf. Harris 2002).
The intricacies of the classifier system are presented in chapter 5.
The discussion includes variation in classifiers (henceforth, 'cls')
depending on the morphosyntactic context (agreement markers on
modifiers, derivational affixes on nouns, numeral cls, verbal cls,
etc.), typology of cls based on semantic criteria (animate, shape and
form, function, specific, quantifier-like cls), as well as stacking
possibilities related to 'double marking' of grammatical functions. Cls
occurring with modifiers of closed classes are shown to vary in
properties like animacy restrictions on the generic form,
obligatoriness of the feminine form (Tariana has feminine vs. non-
feminine gender distinction), number distinction, double plural
marking, individuation effect, etc. Finally, the productivity of the
system is highlighted by the discussion of 'repeaters' (i.e. head noun
repeated on modifier) - the main source of newly grammaticalised cls.
The system of possession marking described in chapter 6 is a
prototypical example of interrelation between genetically inherited and
areally diffused patterns (borrowed from Tucanoan). Tariana
distinguishes between alienably and inalienably possessed nouns, with
the latter obligatorily prefixed by a system of four persons cross-
referencing prefixes. It displays considerable restrictions on the use
of indefinite prefix in comparison to other Arawak languages, as well
as gradual expansion of the group of optionally possessed nouns.
Furthermore, there are two possessive constructions with cls: the
construction with possessive -ya- (pref-ya-cl) and the construction
with generic possessed classifier -ya rupe 'thing, manner'.
Case marking in Tariana (chapter 7) operates on a nominative-accusative
basis and correlates with the discourse status of an NP. There is a
focused A/S (terminology due to Dixon 1994) marker -nhe/-ne. In other
(non-A/S) functions pronouns are obligatorily marked by -na when non-
topical. Topical non-A/S NPs and pronouns are marked by -nuku/naku.
Oblique cases include instrumental-comitative -(i)ne and locative -se.
The pervasive Tariana feature of 'double marking' is also conspicuous
within the system of Case marking: (i)'double case marking' consisting
in stacking a topical non-A/S marker -nuku on top of an oblique case
marker (filled positions 11/12 and 15); (ii) 'double marking of
syntactic function' resulting from the deletion of the embedded
predicate and subsequent stacking of the case marker of an NP in the
subordinate clause with the case marker of the whole subordinate clause
(subordinate clauses in Tariana are nominalizations).
Chapter 8 is devoted to Number marking. Subclasses of nouns are
delineated according to different ways they can(not) show number
distinctions. Among the usual pluralia tantum and uncountable nouns,
there is an interesting singulative morpheme and associative plural ('X
and whoever is with him'). Certain enclitics (e.g. diminutive) are
obligatorily marked for number, which results in multiple number
marking. Human nouns trigger obligatory number agreement both on
modifiers and on verbs. Non-human animates may trigger plural agreement
if individualized, whereas inanimate nouns never trigger verbal plural
Chapter 9 deals with further nominal morphology like nominal tense (and
the possibility of embedding nominal tense under tense-evidentiality
marking), extralocality, contrast, coordination (participant contrast
and action contrast), approximative, diminutive, augmentative and
Chapter 10 contains a brief discussion of derivational affixes (mainly
gender sensitive and nominalizing), as well as non-productive
compounding in Tariana.
In chapter 11 the author reviews certain closed word classes like
personal pronouns, the specifier article, demonstratives, interrogative
pronouns, numerals, distributive individualizer 'napada', quantifiers,
connectives and adpositions. They differ w.r.t. the syntactic function
(head or modifier or both), a pre/post-head position, number
distinction, possibility of taking classifiers, and obligatoriness of
plural agreement on the head. Adpositions are mostly postpositions
(with one exception) and are derived from nouns or verbs. Yet, some of
them take nominal cross-referencing prefixes, locative case marker,
some can be topicalized and can take phrasal enclitics.
Chapter 12 introduces verb classification based on their argument
structure (intransitive 'Sa' (Agent argument cross-referenced on V)
,'So' (Theme argument not cross-referenced),and 'Sio' (with 'oblique'
subject), ambitransitive, transitive, ditransitive, and extended
transitive). Main serialising verbs are also discussed, as well as
seven (!) copula verbs. In addition, the structure of a predicate
consisting of 21 positions is given.
Valency changing derivations are the topic of chapter 13. The
discussion concerns valency reducing mechanisms like passive,
reciprocal, and (marginal) reflexive. A rich system of Tariana
causatives is also analysed and includes: morphological (non fully
productive) causativization, causative serial verbs (curious due to
double marking of the Causer on both a serialising and a lexical
verb), periphrastic (clearly biclausal) causative with and without a
dependency marker. Furthermore, Tariana causative morpheme displays an
interesting property of making a peripheral argument obligatory (though
not core, as in Applicatives). Finally, topic advancing derivation
(resembling Austronesian Voice morphology) is discussed.
A complicated tense-evidentiality system in both affirmative and
interrogative clauses is described in chapter 14. Present and two types
of past (recent and remote) display a distinction between visual, non-
visual, inferred and reported evidentiality. Reported evidentiality is
not used in interrogative clauses, but it occurs in imperative. First
person future distinguishes between certain and uncertain.
Chapter 15 contains a discussion of aspectual markers, degree and
emphatic markers, as well as a semantically heterogeneous semi-open
class of Aktionsart enclitics. The latter can specify manner of action,
direction, associated action, extent of action, and a type of object or
Different types of imperative mood and various modality markers
(frustrative, intentional, apprehensive ('lest'), uncertainty,
conditional, purposive, counter-expectation and declarative-assertive
are described in chapter 16.
Chapter 17 investigates predicate negation, negative imperative,
derivational negative prefix, inherently negative verb stems, and
emphatic negation used in double negation contexts. Chapter 18 features
serialising devices in Tariana. Asymmetrical (aspectual, directional,
modal, causative, etc.) serial verb constructions (SVCs), symmetrical
SVCs and ambient SVCs are discussed and a clear demarcation line is
drawn between the different types.
Complex predicates characterised by the presence of a subordinator and
often containing identical verbs (as opposed to SVCs) are the topic of
Participle formation with both prefixed and prefixless verbs, and
different types of productive nominalization, including the use of
classifiers in a nominalizing function, are discussed in chapter 20.
Chapter 21 gives a general 'syntactic' view of the structure of noun
phrases and predicates. Order of modifiers within NP, different
sentence moods and different types of predicates are discussed.
Additionally, basic grammatical relations, including the recalcitrant
notion 'subject', are also analyzed.
Different types of subordinate clauses (chapter 22) vary in the degree
to which they are nominal and properties like switch reference
sensitivity, positioning w.r.t. main clauses, possibility of taking
different case markers and independent tense/evidentiality marking.
Relative clauses and subordinate complement clauses are the subject of
chapters 23 and 24 respectively, where the latter are necessarily
nominalised in the Subject function.
Chapter 25 contains remarks on discourse organizing devices. Chapter 26
dealing with some residual issues like etymology and lexical semantics
concludes the book.
The book is undoubtedly the result of very careful and meticulous
investigations into the Tariana language spanning the period of several
years. The level of attention to detail is best illustrated by the fact
that the author provides Latin names for various species. All the
examples are carefully glossed and translated and for many of them the
pragmatic context is provided, which makes it easier for the reader to
understand the nuances of the information structure. Given the
polysynthetic nature of the language, the list of abbreviations seems a
necessity and indeed a detailed one has been provided. Many chapters
are concluded by very useful tables consolidating the information
spread throughout the chapter, e.g. Table 20.1 comparing properties of
different types of nominalizations or the summary of similarities and
differences between SVC and complex predicates (19.2).
In general, the book can be treated as a touchstone of the way one
should compile reference grammars. The quality of data presentation
makes it possible for linguists of different theoretical orientations
to use it. However, certain shortcomings have to be pinpointed.
On a purely terminological level, it seems that sometimes the way the
author uses certain terms which have been reserved in the linguistic
literature conflicts with their standard use. Thus, e.g. 'stative
verbs', according to the author, include verbs like 'break', 'wake up',
etc. (p.54). What the author means probably is inchoative variants of
the so-called causative-inchoative alternation. These, however, are not
stative (cf. Vendler (1967)). The same concerns the use of the term
'Aktionsart'. This is an aspectual notion which has a long tradition in
German and Slavic linguistic literature (cf. Isacenko (1960)), as well
as in a recent strand of research where it is equated with 'lexical
aspect' (cf. Ramchand 1997 inter alia). What the author calls
'Aktionsart' in Tariana, however, comprises a very heterogeneous group
of adverbial-like clitics covering notions like manner of action,
associated action, the type of objects, directionals, etc.
Certain other terms like 'generic' and 'indefinite' are frequently
used, but never clearly defined. This leads to a paradoxical situation
when in example 7.24 (p.146) the noun 'nawiki' (people) which is
specific in reference ('all the people') is described as generic.
Similar imprecision in the treatment of the notion 'event' (which has
become increasingly important recently and has been carefully defined
e.g. by Travis (2000)) leads to contradictory statements: on p. 423 the
author describes SVCs as referring to a single event. On the next page,
when dealing with symmetrical SVCs she states: ''They often refer to
several events closely knit together.'' Furthermore, the term
'secondary predicate' (used in connection to example 16.116) has also
been reserved for depictives and resultatives, and so has the term
'resumptive pronouns' standardly used to refer to pronouns inserted
after relativization, but here (example 11.25) naming a kind of free-
Beyond the terminology, the discussion of certain phenomena would
profit from making it more precise. Thus, when discussing the variable
order of enclitics (ex. 12.90 REPETITIVE-COMPLETIVE and 12.91
COMPLETIVE-REPETITIVE) the author concludes it boils down to 'the
aspect of the activity to be focused'. The glosses indicate, however,
that the two sentences might have different truth conditions, which
might be due to scope differences (stemming from structural
differences). Sentences with different focus properties are
indistinguishable with respect to truth-conditional semantics.
Analogously, the author ascribes the ambiguity of negation in ex. 17.7
to the possibility of negation scoping over individual components of a
SVC. This, however, cannot be true if SVCs are monoclausal, as the
author repeatedly emphasizes. What seems to be at stake in 17.7 is the
scope interaction between the two quantifiers (Neg > 'many' or 'many' >
Neg). The reading preferred in 17.7 with 'many' outscoping Negation
seems to have to do with the Contrastive marker on 'many'.
In case the 'Aktionsart' clitic has an adverbial meaning (and a lot of
them do), the reader would benefit from knowing the scope of the
adverbial used in a SVC (as e.g. in ex. 15.145). This remark concerns
not only scoping over individual components of a SVC, but also, in the
case of morphologically causative verbs, scoping over the causing vs.
caused event. (e.g. in 15.148)
As far as prosody is concerned, the fact that proclitics are said to be
able to form both grammatical and phonological words (p.54) seems to
belie a definitional property of clitics, namely their prosodic
deficiency. That fact makes one wonder what are the exact criteria for
distinguishing free forms from clitics, especially in view of the
statement (p.46) that no phonological processes occur on proclitic
boundaries. Is it only the possibility of inserting a 'pausal' form
then? It might also be contingent on the stress properties. These,
however, are not indicated in crucial minimal pair examples 2.6 and
2.7. One grows even more suspicious when it turns out that certain
proclitics can form independent phonological words with enclitics (e.g.
In the discussion of verbal valency the author states that the only
obligatorily transitive verbs in Tariana are those containing the
transitivizer -i(ta). Yet, there are examples (e.g. 13.60) where the
verb contains the causative morpheme, but there is no object. This
seems to force the author to soften the requirement a bit (p.280 in
connection with purpose and instrument arguments) to the effect that
the objects have to be 'easily recoverable from the context'. This
formulation makes Tariana object drop look similar to East Asian
languages phenomenon of topic drop. If that is so, however, the
pragmatics should have been clearly divorced from the syntax. Otherwise
the notion 'obligatory transitivity' becomes rather vacuous.
Furthermore, it seems that the analysis of certain grammatical
phenomena needs to be made a bit more fine-grained. Thus, for example,
there is only one slot (17) reserved for Aktionsart clitics in spite of
the fact that they do cooccur, as noted by the author herself (examples
15.179. and 15.180). The same remark applies to Mood markers, where the
by proxy imperative marker can attach on top of the detrimental one
(ex. 16.33). Moreover, Purposive clearly stands out as distinct from
other Modality markers since it is nominalizing in nature. It also
occurs in a totally different place in the verbal template (cf. 12.1
and 12.2), namely in position 8, exactly where the Passive morpheme
occurs. This is interesting in view of the frequent cross-
linguistically connection between the passive participle and
nominalizing morphology (e.g. in Slavic or Bengali). It is also not
immediately obvious that the semantics of all the examples quoted in
16.7 is purposive.
Some of the features of Tariana that the author considers very rare
cross-linguistically might turn out not to be so exotic in the end. For
instance, the generalization popular in the generative tradition seems
to be that the Tense morpheme defines the root as verbal projection.
From this perspective it is surprising that nominals in Tariana should
take tense markers. However, judging from the glosses the author
provides for examples, the future marker usually ends up translated as
'future' ('prospective', e.g. ex. 9.2 and 9.3) and the past marker as
'dead/late' (e.g. 9.11). Thus, it might turn out that these markers are
simply modifiers of sorts - hypothesis which seems to be confirmed by
the lack of homophony between verbal and nominal Tense markers.
Alternatively, the exoticness of the construction might simply consist
in the fact that nominalization is allowed to attach on top of the full
The other side of the same coin is that certain very interesting and
rare phenomena are a bit underappreciated. Thus, Tariana might
potentially be argued to display an overt split intransitivity
diagnostic (cf. Perlmutter 1978) in the form of cross-referencing
prefixes on unergative and (standard) transitive verbs. The only
argument of So (unaccusative) verbs does not get the prefix. This type
of head marking might actually be treated as traces of ergativity (cf.
Dixon 1994 pp 71-78). Yet, considering the widely assumed
generalization that the presence of an external argument is crucial for
passivization, it is unexpected that prefixless So verbs should be able
to undergo passive formation (cf. ex. 13.14 and 13.15). An additional
quirk complicates the picture, to wit causativized So verbs take cross-
referencing prefixes only when the ''result is achieved
intentionally''. One might treat this fact as suggestive of a fluid
ergative system (in the sense of Dixon 1994). It might also be
comparable to two types of transitivizers in Salish languages (cf. e.g.
Davis, H. (2000) Salish evidence o the causative-inchoative
alternation. Morphological analysis in comparison, ed. by Wolfgang U.
Dressler, Oskar E. Pfeiffer, Marcus Pochtrager and John R. Rennison.
Current Issues in Linguistic Theory 201. John Benjamins.
Dixon, R. M. W. (1994) Ergativity. Cambridge University Press.
Harris, A. C. (2002) Endoclitics and The Origin of Udi Morphosyntax.
Oxford University Press.
Isachenko, A. V. (1960) Grammaticheskij stroj russkogo yazyka v
sopostavlenii s slovackim. Morfologia. Izdatelstvo Slovatskoy Akademii
Nichols, J. (1986) Head-marking and dependent-marking grammar.
Perlmutter, D. (1978) Impersonal passives and the unaccusative
hypothesis. Proceedings of the Fifth Meeting of the BLS. UC Berkeley.
Ramchand, G. (1997) Aspect and predication: the semantics of argument
structure. Clarendon Press.
Travis, L. (2000) Event Structure in Syntax. Events as Grammatical
Objects, ed. by Carol Tenny and Pustejovsky James, pp. 145-185. CSLI
Vendler, Z. (1967) Linguistics in Philosophy. Cornell University Press.