Linguistics 518: Introduction to Linguistic Typology

Report #6: Causatives (due 11 March).

Read Chapter 8 in LULT: 'causative constructions', pp. 165-184. In this
report you will be examining the way in which your language handles
regular changes in valency. By 'valency' is meant the size of the set of
NPUs which a given predicate requires (or presupposes as conceptually
present) in order to be used in a meaningful way. Thus, in English (and
most other languages), the predicate become dizzy requires a single NP
subject:  (a) John became dizzy.  Other NP's may occur in the same clause,
but they are not required by become dizzy for it to be meaningfully used:

(b) John became dizzy from the fumes, on the stairs, after his bath, etc. 

Typically, the valency of intransitive predicates is one; the valency of
transitives is two; the valency of ditransitive predicates (give, tell,
show, etc.) is three. However, the correlation is not always so
straightforward. The verb put, for example, is usually classed as a
transitive verb although its valency is three: an agent, a patient, and a
locative-goal:  (c) Mildred put the books *(on the table, down, away).
Many languages (especially verb-final ones) have regular morphological or
syntactic means for increasing and (less commonly) decreasing the valency
of a given predicate. It is in the context of such regular patterns of
valency-changing that the causative construction is to be examined.
1. Does your language use primarily morphological or primarily syntactic
(phrasal) means for increasing valency? For instance in English the normal
means is syntactic:

	(d) John made his brother fall.

However, there are residual fragments of an older morphological system:

	(e) Mary felled the tree. (Her cutting it made it fall.)

Such pairs in English are now limited to a very few: lie>lay, rise>raise,
rise>rouse, sit>set, rot>ret And there seem to be definite semantic
differences governing the choice between the phrasal versus the
morphological patterns: 

	(f) He makes the sun rise. - He raises the sun.

No matter how regular the pattern is, in every language there is a degree
of suppletion as well:

	(g) Will the parcel arrive on its own? Or must I deliver it?

2. One common pattern of valency change is typically termed the causative.
Different linguists have different conceptions of what the causative is.
In the context of this report, let us define the causative as a situation
in which X , rather than doing Z him/her self, has Y do it:

	(h) John had/made/got/(caused) Albert (to) peel the potatoes.

Terminology: In (h) John is the causal agent, Albert is the causee, and
have, make, get (to), and cause (to) are the causal morphemes. As you can
see from inspecting (h) there are often nuances in the degree of control
of the causal agent over the activity of the causee. Thus, get (to)
implies an element of persuasion or even cajoling that is absent when have
or make is used. In English we have only syntactic means to express the
(narrowly defined) causative. In other languages there are morphological
means:

  (i) 	maalii   phuul   tooR-egaa		
  	gardener flowers  pick-Fut
	'The gardener will pick the flowers.'

  (j) 	ham maalii se phuul tuR-vaa-Ege
  	we gardener from flowers pick-Caus-Fut
	'We'll have the gardener pick the flowers.'

In this example from Hindi-Urdu the causative morpheme -vaa- is affixed to
the basic root tooR- to form a derived causal stem tuRvaa-  and the causee
is indicated by the ablative postposition se "from". The marking of the
causee is an important part of the cross-linguistic study of causatives.
In some languages (like Hindi-Urdu), the causative of a transitive (2- or
3-place) base involves some oblique case or adpositional marking of the
causee. In other languages having a morphological causative, the causee
(which would have been the subject in the corresponding basic non-causal)
is made into the direct object of the derived causal stem:

  (k) koor-yi tshal-yi saaDyi     (l) mye tshal-ineev kuur saaDyi
      girl-Erg washed saaris          I-Erg wash-Caus girl saaris
    'The girl washed the saris.'     'I made the girl wash the saris.'

In (k) (from Kashmiri) the verb agrees with "saris"; in (l) it agrees with
the word for "girl".

3. Do your languages have regular morphological means for decreasing
valency, too?

(m) mistrii sab kamre nahII poot   sakaa 		(Hindi-Urdu)	
    mason   all rooms not   paint  could
   'The mason couldn't paint all the rooms.'

(n) sab kamre nahII put     sake			(Hindi-Urdu) 
    all rooms not   "paint" could
   'Not all the rooms could "paint".'  

Back to 518 Indexpage.