Syntactic Typology of Clause Structure

In September 2019 I happened to be at the right place at the right time, so I had the opporunity to attend a lecture of the person whom my imagination projects as "the Indiana Jones of Linguistics": Denis Creissels. He presented an 90minute introduction to syntactic typology of the clause structure, that I intend to sketch here.

For the examples, please rerfer to the hand out.

Introduction to syntactic typology: Clause structure

1.       Basic transitive construction:

  • Verb 
  • Noun phrase 1: prototypical agent (A) – (roughly “the subject”)
  • Noun phrase 2: prototypical patient (P) – (roughly “the object”)

 

1.1.    Definition

Prototypical transitive actions (highest possible degree of semantic transitivity)

  • Verb: involve a change of state or position undergone by the patient
  • Agent: triggers the action involving the change.
    • Conscious and voluntary action
    • It aims at changing the state of the patient or controlling its position
  • Patient: undergoes the change of state 


1.2.    Extension of the basic transitive construction to the coding of situations that cannot be characterized as prototypical actions

 ·         The basic transitive construction can be used to encode not prototypical actions.

Ø  Example: the verb “to see”, expresses no action but uses the same construction as “repair”

·         There is cross-linguistic variation in this ‘transitivity prominence’

 ·         The subject can be coded differently depending on whether there is change of state or not:

·         In ergative languages, subject is marked as:

- ERGATIVE if the verb denotes a change of state (to break)          

  DATIVE if it does not (to see, to forget) 

·         Languages can also mark

-  only transitivity, where there is change of state

-  only intransitive constructions, if there is no change of state

 

1.3.    A terminological point

·         Use ‘agent’ and ‘patient’ for the semantic roles only.

·         Use A and P for NP1 and NP2 that, whatever their semantic roles, are encoded in the same way as the agent and the patient in the basic transitive construction (roughly, subject and object)

  

1.4.    Basic transitive construction and intransitive constructions of transitive verbs

 ·         In many languages, verbs expressing transitive actions have alternative constructions with the same denotative meaning:

-  Basic transitive construction:               The mechanic has repaired the car

-  Alternative construction:                      The car has been repaired by the mechanic 

·         How can we determine which construction is the basic transitive one?

-  Basic transitive construction:

o   Verb form is morphologically less complex

o   It has less restrictions

o   It is more frequent in spontaneous discourse

-  Alternative construction:

o   Agent and patient marking  have less characteristics typical for core nominal terms à evidence for detransitivization 

1.5.    Transitive coding typology

 

·         In some languages, if there is no morphological mark distinguishing subject (A) from object (P), their position may be essential for their identification à rigid word order

·         Two types of morphological mechanisms may contribute to their distinction:

-  Indicating (‘flagging’) the subject and / or the object. 2 procedures:

o   Marking case

o   Adjoining adpositions

 

§  Indexation of subject and / or object (agreement):  use of morphemes coindexed to them in a different position; often, within VP or in its immediate periphery

 

·         The contrast subject – object is maximized when one is indicated by case and not agreement, whereas the other is indicated by agreement and not case. The most common cases in world’s languages:

o   Objects are more commonly indicated by case marking

-  “Differential marking of object” is often conditioned by definiteness and / or animacy.

o   Subjects are more commonly indicatd by agreement

§  Differential agreement can also be conditioned

 

·         Transitive coding is not necessarily uniform: two or more formal types of coding may alternate according to various types of conditioning:

o   Transitive coding may vary depending on verb’s TAM value

-  Present: subject agreement & object case marking

-  Past: subject case marking & object agreement

 

·         In most languages, it is possible to describe the coding of A and P separately, but sometimes the coding of subject and the coding of object may interfere (co-argument sensitivity).

-  Ex: Hungarian only shows subject & object agreement if the object is determined

 

·         In some languages, the expression of TAM distinctions may differ in intransitive and transitive constructions:

-  Completive aspect marked in transitive clauses

- Zero-marked in intransitive clauses

 

1.6.    Optionality vs. obligatoriness of subject and object

 ·         Cross-linguistic variation in the possibility vs. impossibility of omitting the subject and object NPs (with either an anaphoric or non-specific Reading)

 ·         This variation is not correlated to the presence vs. absence of an agreement

-  Example: Both Bambara and Japanese lack agreement , but in transitive constructions

o   In Bambara, impossible to leave subject and object unexpressed

o   In Japanese, subject and object are commonly omitted

 

2.       Intransitive constructions

 2.1.    Valency and transitivity

 ·         Not all languages are obliged to express the participants denoted by the lexical meaning of verbs, in the transitive construction or else

 ·         Transitive (bivalent) verbs may have alternative intransitive constructions:

-  passive

-  antipassive 

·         Subjects or objects often receive the same marking as adjuncts 

·         In some languages, intransitive (monovalent) verbs can have transitive constructions

-  they can work as light verb compounds:

-  in Basque, ‘sleep’ and ‘speak’ are expressed literally as ‘do sleep’ (lo egin) and ‘do word’ (hitz egin)

-  The verb expresses subject agreement

-  The verb expresses object agreement with the nominal part of VP 

·         Intransitive (monovalent) verbs may also take transitive marking even to register the CCT 

2.2.    Alignment relationships between transitive and intransitive constructions 

·         In most languages, intransitive constructions mark their subject (‘single core term’) in the same way they are marked in transitive constructions (accusative or ergative). 

·         In most languages, intransitive and transitive constructions show the same case / agreement marking, but some languages have two classes of intransitive verbs:

o   Intransitive verbs with same marking as transitive ones

o   Intransitive verbs with different marking from transitive ones 

·         This similarity of marking between transitive and intransitive constructions can be conditioned grammatically, in particular, by TAM value 

·         Marking case in the subject of intranstive constructions does not necessarily correlates with ergativity.


3.       Trivalent verbs

3.1.    Extended transitive coding and double transitive coding

 

·         Trivalent verbs such as ‘give’, ‘show’, ‘send’, or ‘sell’ select for 3 arguments

·         They code the most agent-like participant as subject

·         Languages vary in the coding of the other two participants: recipient and theme

a) One is marked as object and the other as oblique (extended transitive coding):

-   Indirective: object is theme (give something to someone)

-   Secundative: object is recipient (provide someone with something)

b) No apparent hierarchy in recipient and theme coding (double-transitive constructions)

c) Syntactic hierarchy in double-transitive constructions is a complex question, since there may be asymmetries in the syntactic behaviour of the recipient and the theme:

-  Passivization: in some languages, only recipient or theme passivization is possible


 

4.       Non-core participants and circumstantials (obliques) 

·         Cross-linguistically, participants not selected by the verbal semantic structure are labelled  ‘obliques’

o   They are marked by adpositions or by a morphological case different from the one used in the quotation form.

o   The can be unmarked

-  Example: place names in locative function do not take locative morphological marks

o   They can be marked by a construction with a second verb which expands the argument structure of the first verb.

-  Example: ‘Buy it for me.’ Expressed as BUY 3SG GIVE 1SG

- where the marker for recipient (for) is expressed as the GIVE verb

 

 5.       Verbal valency

5.1.    The notion of valency alternation

 ·         ‘Valency alternation’: use of the same verb, or a derived form, in distinct constructions with the same denotative meanings

-  Example: active-passive alternation, with or without verb modification

·         Agreement morphemes can be coreferent with a given semantic role even if they realize different functions, like subject or oblique

 

5.2.    Morphologically marked valency alternation.

 

·         Morphologically marked valency alternation can often (but not always) be analyzed as derivation from a base form to a morphologically more complex derived form. 

·         The most widespread types of morphologically oriented valency alternations

-  Passivization: demotion of the subject

-  Antipassivization: motion of the object ("promoting" the accusative object to a nominative subject)

-  Reflexivization

-  Reciprocalization

-  Decausativization ou anticausativization

 Causativization: introduction of a causer in the syntactic role of subject

-  Applicativization: introduction of a participant as object

 

 

5.3.    Lability

             Labile verbs are verbs lending themselves to valency alternations involving no morphological marking. ‘Causal-noncausal’ lability:

-  As in The child broke the glass / The glass broke. It is particularly widespread

-  Many other types can be found in the languages of the world

 


6.       Clause structure and information structure

 

·         In some languages, linear order of NPs in the construction is bound rigidly to their semantic role, as in French John called Mary / Mary called John. 

·         In some other languages, the linear order of NPs is flexible (permutations are possible) without formal readjustment or changes in NPs semantic roles

o   This variation express differences in information structure 

·         In most languages, linear order is neither totally rigid nor totally flexible. For example, constituent order in Spanish is much less flexible than in Russian, Hungarian, or Basque, but more flexible than in French. 

·         In many of the languages with some flexibility in constituent order, the variation in constituent order crucially involves a ‘focus position’. Depending on the language, focus position may be:

o   Clause-initial (Wolof)

o   Clause-final (Russian)

o   Immediately before the verb (Basque, Hungarian)

o   Immediately after the verb (Makhuwa)

o   In some languages the use of the focus position requires special morphological marking (Wolof, Makhuwa)

 

·         Also, in languages with more rigid constituent order, focalization can be expressed by means of ‘cleft’ constructions in which the term in focus is introduced as the argument of an equative predicator 

-  It’s Mary who called John / It’s John who called Mary 

·      ·            Besides, some languages mark focalisation just intonationally (English)

 ·         Some other languages add a focus marker to the focalized constituent, without any change in the construction

 


Comentarios

Entradas populares