|
2000 East Nusantara Linguistics Workshop
Valency Workshop
Co-ordinator: Marian Klamer klamerm@let.vu.nl
Valency
Valence or valency is a semantic and a syntactic notion. As a semantic notion it refers to the number of participants 'on stage' in the scene expressed by the verb. Valence as grammatical notion concerns the number of arguments present in any given clause, and canonically they are expressed by a noun (phrase) or a pronominal element. The semantic and syntactic notions of valency need not coincide: for example, in He is eating the Patient/Theme is semantically present but not grammatically expressed.
The study of verbal valency, the processes that alter it, and the expression of arguments, gives us information about what exactly constitutes the interface between the lexicon and the syntax of a language. For this reason it has been central to much recent linguistic debate.
Questions that are often addressed include the following: Where exactly does a 'word' end and becomes a 'phrase'? What are the principles behind the derivation of certain types of verbs? What are the principles that govern the expression of verbal arguments: when can they be covert, and when are they obligatorily overt? What is it that triggers the difference?
The study of verbal valency also helps us to discover more about the structure of the lexicon. For example: Which semantic notions are expressed by root words, and how do these differ (if at all) from the notions expressed in morphologically derived forms? Can we make crosslinguistic generalisations here, or does every language exhibit its own ideosyncratic lexical patterns? Note that the latter is assumed in theories that consider a language's lexicon to be the inventory of arbitrary signs Ð how valid is this assumption?
In the workshop, we will compare the languages with respect to the semantic valency of their verbs, the manipulation of verbal arguments by derivational morphology, and the expression of verbal arguments in syntax. We hope that this will provide us with an interesting typology of valence and valency in Eastern Nusantara, which can be used to contrast the characteristics of those languages with Austronesian languages of other areas. It may even be used to compare traits of Austronesian languages with those of IndoEuropean languages, in order to evaluate some of the claims in linguistic theories on their crosslinguistic validity.
A. Valency and morpho-syntax: the expression of verbal arguments
1. What are the grammatical relations of this language?
2. What is the evidence for the grammatical relations, e.g.:
3. Is the system of grammatical relations in basic (affirmative/declarative) clauses organised according to a nominative/accusative, ergative/absolutive, tripartite, or some other system?
Kambera
4. Is there some split in the marking of the grammatical relations, e.g:
The operating principles for the system of verb agreement are different from those of full NPs, but the split is not in terms of Nom/Acc vs. Abs/Erg. Grammatically definite NPs (containing an article) are crossreferenced on the verb and may be omitted. Grammatically indefinite NPs are not crossreferenced on the verb. A or S may or may not be crossreferenced when they are indefinite.
Indefinite NPs may or may not be absent (implied) depending on their pragmatic/discourse saliency. In other words, whether or not arguments are expressed through pronominal crossreferencing depends on the Ôgrammatical definitenessÕ of the argument, while their expression in an NP depends on pragmatic/discourse saliency. There is thus a clear definiteness effect. This effect is visible in all object markings, and in many subject (S/A) markings.
5. Exemplify some simple intransitive, transitive (and ditransitive) clauses, illustrating the expression of S, A and O (direct, indirect).
Kambera
S = Nominative, Genitive or Dative
1. Provide an overview of the basic verbal classes, according to their valency.
2. Are there semantic classes of verbs with special morpho-syntactic properties?
Kambera
Yes, deictic verbs, verbs of emotion and appearance and ideophonic verbs:
3. Are there semantically transitive verbs that always have an overt object? (pronoun, pronominal clitic, inflectional affix)
Kambera
Yes, in principle all transitive verbs have an overt object (NP or pronominal clitic or both). In contexts where the object is understood and unambiguous, it can be omitted, as in:
Are there semantically transitive verbs that never have an overt object?
Kambera
No. But some transitive verbs are often attested without an object, these are also mentioned under 4 below.
4. Does the language have ditransitive verbal roots (morphologically underived verbs with three arguments), or are ditransitive verbs always derived?
Kambera
Always derived, see data above.
5. Are there structurally motivated subclasses of intransitive verbs, e.g. the distinction between 'unaccusative' (fall, melt) and 'unergative' verbs (run, dance)?
Indicate the type of structural distinction: argument marking, derivational morphology, other...?
Kambera
There are no distinct unaccusative/unergative lexical classes of intransitive verbs. However,
there is a special class of deictic intransitive verbs (4 verbs in total): they have a
special type of argument marking: their single argument is always marked as an object.
There is another class of intransitives that can be separated from the rest of the
intransitives because they are derived from transitive roots (by prenasalisation):
pata 'break X'> m.bata 'be broken'. This derivational process is no longer productive.
6. Are there also verbs of which it is not clear whether they are transitive or
intransitive because they appear just as frequently with an overt object as without one?
Kambera
Yes, about 25 verbs occur as frequently with an object as without one. For example:
Also, verbs derived from numerals allow both an intransitive and a transitive interpretation:
7. Are predicates of possession and location transitive or intransitive?
Kambera
Intransitive.
Is there a verbal form used in the expression of ((in)alienable) possession:
I have a garden/a child (NOT: That is my garden/my child)? If so, is the verb
transitive or intransitive? If not, provide the translational equivalent of
the examples, with glosses.
Kambera
There is no special possessive verb. The deictic (locational) verb ningu
'be here' is used as such:
Is there a verbal form used in locational predicates: He is at home/There are
fish in the sea? If so, is the verb transitive or intransitive? Provide the
translational equivalent of the examples, with glosses.
Kambera
Yes, intransitive ningu 'be here' is the locative verb used in such predicates:
8. Is there a verbal form used as existential predicate: I am a woman/She's a teacher.
If so, is the verb transitive or intransitive? If not, provide the translational equivalent
of the examples, with glosses.
Kambera
No. The nominal predicate is used as such, and its argument marked with an Accusative clitic:
C. Valency and derivational morphology: Valency-increasing devices
In what follows a checklist is given for derivational processes that result in valency
increase or decrease. For each of the processes it is important to distinguish between
morphological derivations (die > die-CAU), and analytic/periphrastic derivations
(die > cause to die). In the latter case, there may be an overlap with serial verb
constructions and/or clause combining strategies such as complementation. It will probably
lead us too far afield to discuss the non-morphological constructions in detail, especially
because it would require quite detailled background knowledge about the syntax of the language.
However, if your language employs periphrastic contructions for the valency changing operations
discussed below, it would be good to note this and to provide some illustrations, plus an
indication of how frequently that type of construction is used (seldom/sometimes/always). The
same applies if your language is largely isolating and employs lexical substition to express e.g.
causatives (die > kill).
Causative
Comment: A causative construction can be symbolised as CAUSE(x, P) = 'x causes P', where x is the argument introduced by the causative derivation, and P the caused predicate. Causative predicates standardly involve one more argument than the caused predicate. Therefore, if the caused event is intransitive, the causative is transitive (Wayan made [John laugh]). If the caused event is transitive, the causative is ditransitive, etc. (Wayan made [John eat nanga])
1. Form: Describe the word class and the morphological type of the possible bases for a
causative derivation. Provide illustrative examples.
Kambera
Root morphemes as base
Morphologically complex (=derived) forms as base
Is there a distinction between productively and unproductively derived forms?
Kambera
Unproductively derived verbs are often the base of productive derivations.
2. Function: Describe the semantic and grammatical function of the causative derivation.
Semantic: Does the causative always add an argument to the event structure of the verb?
Provide illustrative examples where appropriate.
Comment: The causative derivation does not necessarily result in a valency change, esp. if
the base is transitive: Leti va- is also used for 'intensive iteration', Tukang Besi hoN-
takes a transitive verb as its base and derives a transitive verb with a 'stronger sense of
result arising from the action'.
Kambera
The causative prefix pa- in Kambera has different effects. It may change the valency of
the verb by adding one argument. In that case, a causative or permissive verb with a
causing/permissive argument is derived. When the base is a noun, the added argument is not
necessarily causative or permissive: lunggi 'hair' > pa-lunggi '(cause to) have
hair'. And finally, the derivation may also result in a change in the semantic features of the
verb, by making the Agent/Actor of the base verb more volitional. In this case there is NO
argument added.
The causative derivation in Kambera should therefore NOT be characterised in terms of valency
change, nor as a derivation adding a causing argument, but rather as a morphological derivation
that adds notions of [control] or [volitionality] to the semantics of the verb. The semantic
link between the causative/factitive function of pa- and its intensive/volitional reading is
that both interpretations need a controlling, volitional Agent/Actor.
In Kambera, the valency change caused by prefixing pa- is thus just a possible, secondary
effect of the overal semantic change caused by prefixing pa-.
Describe the semantics of the causer (the 'added' argument): is it always agentive/controlling,
or is it variable depending on the word class or morphological type of the base? Provide
examples where appropriate.
Kambera
No. tila 'kick X' > pa-tila (1) 'kick habitually', (2) kick each other'
Grammatical: Does it add a syntactic argument to the clause?
Describe the grammatical function of that argument: subject or otherwise?
How is the grammatical status of the original arguments altered: base S > derived O,
etc.? Check this for distinct types of morphological bases. Note in particular the distinct
argument alternations in: (i) causative derivations of verbal versus nominal bases and (ii)
causative derivations of transitive versus intransitive bases. Provide examples where a
ppropriate.
Kambera
3. Does the causative derivation have other functions, e.g.:
Kambera
Intensity, and reciprocality, see examples under 2.
4. Describe how the arguments of a causative verb are expressed in the clause:
A: usually a clitic with an optional NP crossreferencing it
Is there a difference between the expression of arguments of causative verbs and those of
'common' transitive verbs? Provide examples where appropriate.
Kambera
No.
Applicative
Comment: An applicative derivation is a valency increasing operation. It adds an object argument
that is (in the canonical case) semantically a Goal (Beneficiary, Recipient, Location):
arrive > arrive-APP my mum 'arrive at my mum's' dance > dance-APP the teacher 'dance
for the teacher'. It may also bring a peripheral participant onto 'center stage' by making it
into an object of the verb: give food [to Suharto] > give-APP food Suharto; climb [on the
hill] > climb-APP hill.
1. Form: Give the applicative morpheme(s).
If there is more than one, indicate the distinction, e.g. in terms of the semantics of the type
of argument that is added.
Kambera
Suffix -ng
Describe the word class and morphological type of the possible bases for the applicative
derivation(s)
Kambera
2. Function: Describe the semantic and grammatical function of the applicative
Semantic: Does the derivation 'upgrade' a participant in the event structure of the verb, does
it 'add' a participant, or can it do both? Describe the semantics of the upgraded/added
participant. Note the possible variations depending on the word class or morphological type of
the base.
Kambera
Applicative argument can be both added and upgraded. Semantics: Recipient,
Beneficiary, Maleficiary, Goal, Source, Location. (Not: Theme, Instrument,
Companion, ...)
Grammatical: Does it add a syntactic argument to the clause? Describe the grammatical function
of that argument: O (indirect?) or otherwise? How is the grammatical status of the original
arguments altered? Check this for the various types of morphological bases. Note in particular
the distinct argument alternations of the applicative argument in: derivations of verbal vs.
nominal base and of transitive vs. intransitive base.
Kambera
3. Does the applicative object argument entail other semantic roles besides the standard
ones of Beneficiary, Recipient, Location?
Comment: in many languages, the applicative derivation treats all non-Patient object arguments
alike, but there are also languages where the introduction of an Instrument or a
Companion/Comitative argument involves (a) separate morphological derivation(s).
Kambera
In Kambera an Instrument is introduced with a serial verb construction where a verb is
combined with a second verb, the instrumental verb wàngu 'use'.
4. Does the applicative derivation have other functions where it does not change the valency of the base verb?
For example, when it has (secondary) functions as a:
Kambera -ng:
general verbalising morpheme: N ngia 'place' > V ngia-ng 'be located')
morpheme indicating aspectual notion of 'continuous' action and/or
to derive process-oriented verbs: winggir 'besiege/surround X' >: winggiru-ng
'to be laying a siege' (* 'surround X for Y')
morpheme marking animate/inanimate contrast: rongu 'hear X (inanimate)' > rongu-ng
'hear X (animate)'
5. How are the arguments of an applicative verb expressed:
Definite argument: Pronominal marker n the verb, with optional coreferent NP
Possessor raising or external possession
1. Does the language have a derivation where a nominal possessor is raised to grammatical
object position, e.g.: I ate the white man's beans > I bean-ate the white man
Kambera
No.
2. Does the language have a derivation where a possessor is raised to grammatical
subject, e.g. : My heart is bleeding > I heart-bleed
Kambera
Yes, when emotion verbs are derived, a verb is combined with the noun eti 'liver',
the seat of emotions. This noun may constite a separate NP, functioning as the
grammatical subject of the predicate, or be incorporated. In the latter case, the
possessor of the noun is 'raised' to function as the grammatical subject of the derived
predicate.
Example:
3. Give an overview of the type of predicates/clauses in which PR can occur
Kambera
PR is only used to derive emotion verbs. It generally incorporates nouns with
intransitive verbs.
4. Does PR always cooccur with noun incorporation? If so, are there
(semantic/syntactic) restrictions on the types of nouns that may be incorporated? Are there
restrictions on the incorporating verbs?
Kambera
In Kambera, it generally incorporates body part nouns (especially eti) with intransitive verbs.
Mutual interaction of valency-increasing derivations and their functions
1. Does one and the same base allow various derivations?
Kambera
Yes, provided it makes sense semantically.
2. Do the functions of the various affixes show overlap, e.g. when an applicative also
has a causative function (I go-APP Wayan 'I go with Wayan: I make Wayan go (with me)'?
Kambera
Not in Kambera.
3. Can PR constructions be the base for causative/applicative derivations, or vice versa?
Kambera
Derived emotion verbs can be the base for causative/applicative derivations in Kambera.
In such derivations, the possessor can be raised, cf. the contrast between the following
two examples. Note that the raising of the possessor in the second example is not accompanied
by incorporation of eti.
4. Is there evidence for an intrinsic ordering of any of the processes?
Kambera
No, the processes can apply in either order, depending on what is semantically most appropriate.
D. Valency and derivational morphology: Valency decreasing devices
Passive
1. Does the language have a passive construction? (A clause containing a verb with
special morphology to mark the promotion of a 'deep' Patient/Theme to grammatical subject, and
the demotion of the original Agent into an oblique/adpositional phrase)
Kambera
There is no passive construction in Kambera.
2. If so, do all verbs, including the applicative/causative/etc., always allow passivisation? Or is passivisation restricted to certain classes of verbs?
Kambera
N/A.
3. Do verbs with passive morphology also appear in other contexts, with different functions? E.g. as modifiers of nouns (the lecture was given by him > the given lecture).
Kambera
N/A.
Detransitive
1. Does the language have a detransivising derivation with a prefix tV(C)- ?
In Kambera, the process seems partly productive, partly unproductive: Many derived forms with
ta- do no longer have an independently used root form, but there are also many derived
forms with an independently functioning root form.
In Kambera ta' can be applied to both transitive and intransitive verbs:
In transitive verbs, the prefix expresses that the Patient/Theme argument becomes the
single argument of a non-intentional, involuntary, accidental or unexpected event.
Derivation with transitive base : O > S
The sole argument of the derived forms is expressed as S, not as O, though the morpho-syntactic
marking of this S is more limited than a 'normal' S. For example, this S is often only expressed
as an NP, and not crossreferenced on the verb.
2. Does the language have another detransitivising morphological process?
Prenasalisation combined with initial consonant mutation. This process detransivises a
transitive verbal base. The original O becomes the S of the derived verb. Examples:
kunggulung 'roll X' > nggunggulung 'roll over'
In contrast with ta- prefixation, this process is no longer productively used, and cannot
take any transitive base, because it is limited to those verbal stems that have initial
consonants that may be prenasalised. It also does not invoke the semantic notions of
unexpectedness, suddeness etc. that are typical of ta- derivations. As simple detransitivising
process it seems like it the mirror image of the causative; a kind of 'anticausative'.
Reflexive
1. Does the language employ a specific construction to express reflexive notions such as to
dress herself? Is there more than one construction expressing reflexiveness?
Kambera
Kambera uses a construction where the verb is combined with a possessed noun wiki 'self'
2. Is the object of the reflexive expressed as a special pronoun (as in English), or as a noun
(liver, body, self)?
Evidence for considering wiki as a noun:
The noun is not found in any other contexts in Kambera.
3. Is there an extra morpheme on verbs in reflexive constructions?
A Kambera verb in a reflexive construction has the prefix pa-. This prefix has many
functions, among others deriving causatives (cf. above).
Reciprocal
|