Você está na página 1de 13

DISTINGUISHING RAISING AND CONTROL IN SUNDANESE1

Eri Kurniawan
University of Iowa
Indonesia University of Education
eri-kurniawan@uiowa.edu

Just as in most of the world‘s languages, Sundanese 2 contains constructions that appear to be
structurally identical such as raising and control. The present paper attempts to provide a
descriptive account of the syntax and semantics of raising and control in Sundanese, identifying
the crucial ways in which the two constructions differ from each other. I will then suggest that,
unlike claims that have been made for other languages, the complement of a raising predicate is a
‗larger‘ constituent than is that of control. To shed some interesting light on the debate whether
raising and control involve two disjoint processes within the minimalist approach of Chomsky
(1995) or they should be treated in a grammatically uniform fashion as propounded by Hornstein
(1999), I examined a number of facts that distinguish raising and control in Sundanese. The facts
converge in suggesting that the two constructions in question are syntactically and semantically
different.

1. Background

A great deal of attention has been paid to pairs of sentences such as (1-2), which are generally
assumed to be raising and control constructions, respectively.3

(1) Ahmad di-anggap (ku)4 abah indit ka kota.


A PV-assume by grandfather go to town

1
I would like to say hatur nuhun ‗thank you‘ to William Davies, Roumyana Slabakova and Alice Davison for their
invaluable comments and feedbacks. Also, thanks to the audience at the 17th AFLA for their helpful comments and
input concerning some of the issues elaborated here. I am also grateful to all Sundanese facebookers for their
comments, feedbacks and judgments for the data.
2
Sundanese is one of the many Indonesian indigenous languages predominantly spoken in the west part of Java,
Indonesia. It has some 27 million speakers, making it the third most widely spoken language in Indonesia, after
Indonesian and Javanese.
3
The abbreviations used in the glosses are: AV: actor voice, PV: passive voice, SING: singular, PL: plural, COMP:
Complementizer, REL: relativizer, PAR: particle, DEM: demonstrative, DEF: definite, PROG: progressive, FUT:
future, PERF: perfect, RED: reduplication, APPL: applicative.
4
The preposition ku ‗by‘ in passive constructions is optional only if the agent adjunct is third person. When it is first
or second person, the omission of ku results in ungrammatical sentences, as shown below.
(i) *Buku éta di-baca kuring.
book DEM PV-read I
‗The book was read by me.‘
(ii) *Buku éta di-baca manéh.
book DEM PV-read you
‗The book was read by you.‘
‗Ahmad was assumed by the grandfather to go to town.‘

(2) Ahmad di-titah (ku) abah indit ka kota.


A PV-order by grandfather go to town
‗Ahmad was ordered by the grandfather to go to town.‘

Despite the virtually identical surface representation, the two constructions have been argued in
other languages to exhibit some syntactic and semantic differences. Cross-linguistically, some of
the differences pertain to the finite or non-finite nature of the complement clauses, which is not a
noticeable property in a language like Sundanese, as Sundanese does not have outward signs of
finiteness. Hence, it is unclear how raising and control constructions differ in this language.
The split between raising and control has been under debate in recent years. Among the
many frameworks people have proposed and adopted, two have enjoyed a lot of attention;
namely the traditional view from the Government and Binding Theory (Chomsky 1981), adopted
in most minimalist treatments (Chomsky 1995), and the Movement Theory of Control (Hornstein
1999). The former holds a belief that raising and control are two disjoint processes, whereas the
latter insists that raising and control should be treated in a grammatically uniform fashion. One
of the focuses of this paper is, therefore, to determine which of the two approaches can best
explain the raising and control phenomena in Sundanese.
The goals of this paper are as follows: (i) to provide a descriptive account of the syntax
and the semantics of Sundanese raising and control; (ii) to propose an analysis for the structural
distinctions among raising, control and prolepsis in Sundanese; and (iii) to assess whether the
movement analysis of control holds up in accounting for the syntactic differentiation between
raising and control in Sundanese.
This paper is organized as follows. Section two sketches out how raising and control
differ in Sundanese by means of various traditional diagnostics. It also provides additional
peculiar characteristics that are able to tease apart the two constructions in question. Section
three offers an analysis to account for the syntactic distinction between raising and control in
Sundanese. The last section presents the implications of the raising and control structural
distinction in Sundanese for the current analyses of raising and control.

2. Raising and Control in Sundanese

This section delineates syntactic and semantic facts of raising and control in Sundanese.
Importantly, through the application of the well-established diagnostics and some additional
properties peculiar to Sundanese, I show that raising is syntactically and semantically different
from control.

2.1 Traditional Diagnostics

Raising is differentiated from control through various traditional tests such as thematic roles,
embedded passives, selectional restrictions and complement extraction.
2.1.1 Thematic Roles

Raising and control structures have distinct thematic structures. Observe the following sentences.

(3) Ahmad tangtu mariksa barang bawaan-na.


A certain AV.examine thing carry.NOM-DEF
‗Ahmad was certain to examine his belongings.‘

(4) Ahmad ati-ati5 mariksa barang bawaan-na.


A careful AV.examine thing carry.NOM-DEF
‗Ahmad was careful in examining his belongings.‘

At first glance, the difference between (3) and (4) seems to lie merely in the use of matrix
predicate, i.e. tangtu ‗certain‘ and ati-ati ‗careful‘. However, when we dig deeper, it becomes
apparent that the two sentences are significantly different in terms of the thematic role of the
subject Ahmad. In (3) Ahmad receives only one single role, that is, the agent of mariksa
‗examine‘. On the contrary, Ahmad in (4) seems to receive two thematic roles: the experiencer of
ati-ati and the agent of mariksa.
A similar difference can be found with transitive raising and control verbs. Observe the
following sentences.

(5) Ahmad di-anggap nulis surat éta.6


A PV-assume AV.write letter DEM
‗Ahmad was assumed to write that letter.‘

(6) Ahmad di-paksa nulis surat éta.


A. PV-force AV.write letter DEM
‗Ahmad was forced to write that letter.‘

Ahmad in (5) is solely the agent of nulis ‗write‘. Ahmad in (6), on the other hand, appears to be
the patient of dipaksa and the agent of nulis.

5
Ati-ati ‗careful‘ is indeed a (adjectival) predicate, not an adverbial such as sigana that corresponds to the English
raising predicate seem. Observe the distinct distribution of ati-ati and sigana in the following examples.
(i) (Sigana), Ahmad (sigana) maca buku ieu.
probably A. probably AV.read book DEM
‗(Probably), Ahmad (probably) read this book.‘
(ii) (*Ati-ati), Ahmad ati-ati maca buku ieu.
careful A. careful AV.read book DEM
‗*Careful Ahmad read this book.‘
‗Ahmad was careful to read this book.‘
The fact that ati-ati cannot precede the subject provides empirical evidence for its predicative function. In fact, ati-
ati can act like an adverbial by obligatorily combining with a preposition such as kalawan ‗with‘.
6
The letter é in the Sundanese orthography represents the tensed mid front unrounded vowel, which is distinct from
the regular e that is an orthographic symbol for a schwa. And the vowel sequence eu as in meuli ‗buy‘ represents a
lax central unrounded vowel, which is articulatorily produced higher than the schwa.
2.1.2 Embedded Passive

With raising predicates, sentences with a passive and active complement are synonymous. This
is exemplified in (7) and (8).

(7) Ahmad nga-harep-keun paraji (jang) mariksa pamajikan-na.


A. AV-expect-APPL midwife to AV.examine wife-DEF
‗Ahmad expected the midwife to examine his wife.‘

(8) Ahmad nga-harep-keun pamajikan-na (jang) di-pariksa (ku) paraji.


A. AV-expect-APPL wife-DEF to PV-examine by midwife
‗Ahmad expected his wife to be examined by the midwife.‘

In both cases, Ahmad‘s expectation remains the same; that is, the midwife will examine his wife.
Conversely, with control predicates like ngolo ‗persuade‘, a sentence with a passive complement
is not synonymous with the active counterpart. Compare (9) and (10).

(9) Ahmad ng-olo paraji (sangkan) mariksa pamajikan-na.


A. AV-persuade midwife so that AV.examine wife-DEF
‗Ahmad persuaded the midwife to examine his wife.‘

(10) Ahmad ng-olo pamajikan-na (sangkan) di-pariksa (ku) paraji.


A. AV-persuade wife-DEF so that PV-examine by midwife
‗Ahmad persuaded his wife to be examined by the midwife.‘

In (9) Ahmad persuaded the midwife of the need for examining Ahmad‘s wife, while in (10) it
was Ahmad‘s wife that was persuaded to be examined by the midwife.

2.1.3 Selectional Restrictions

In raising constructions, when the selectional restrictions of the embedded predicate are met, the
sentence is perfect (11). If they are violated, though, the whole sentence becomes semantically
odd (12).

(11) Ahmad percaya ucing éta maok lauk di dapur.


A believe cat DEM AV.steal fish in kitchen
‗Ahmad believed that cat to have stolen fish in the kitchen.‘

(12) #Ahmad percaya ucing éta bisa ng-omong basa Inggeris.


A believe cat DEM can AV-speak language English
‗#Ahmad believed that cat to be able to speak English.‘
The verb ngomong ‗speak‘ must felicitously combine with a sentient, volitional agent, and a cat
is not such an agent.
The situation is different with control sentences (13-14). Both sentences are semantically
odd regardless. This is due to the semantics of maksa ‗force‘, which assigns a thematic role to its
object and requires it to be sentient and volitional—a requirement that can never be met by a cat.

(13) #Ahmad maksa éta ucing maok lauk di dapur.


A. AV.force DEM cat AV.steal fish in kitchen
‗#Ahmad forces the cat to steal fish in the kitchen.‘

(14) #Ahmad maksa éta ucing ng-omong basa Inggeris.


A. AV.force DEM cat AV-speak language English
‗#Ahmad forces the cat to speak English.‘

2.1.4 Complement Extraction

Jacobson (1992) offers one more property that can further distinguish raising and control, namely
extraction. She claims that it is possible to extract the complement of control verbs, but
impossible to extract the complement of raising verbs.
In fact, this is what we see in Sundanese. Control verbs countenance extraction or
fronting of their complements, but raising verbs do not. This is exemplified in (15-16).

(15) Sina nga-latih Persib, Ujang di-titah (ku) abah téh.


So that AV-train P U PV-order by grandfather PAR
‗Ujang was ordered by grandfather to coach Persib.‘

(16) ?*Jang nga-latih Persib, Ujang di-harep-keun (ku) abah téh.


to AV-train P U PV-expect-APPL by grandfather. PAR
‗Ujang was expected to coach Persib.‘

In (15) sina ngalatih Persib ‗to coach Persib‘ is the complement clause of the control verb dititah
‗be ordered‘, which gets fronted. And the sentence is acceptable. However, fronting of the
raising complement is marginally acceptable (16).

2.2 Other Distinguishing Properties of Raising and Control

There are three additional properties peculiar to Sundanese that give rise to structural differences
between raising and control. All these properties have to do what kind of element can readily
appear in the complement clauses.
2.2.1 ‗Tensed‘ Auxiliaries7

Control verbs do not allow any ―tensed‖ auxiliaries in the complement clauses (17), but raising
verbs do. Thus, the control structure in (18) is ungrammatical with the tensed auxiliary.

(17) Ujang di-titah (ku) bapa-na (*rék) indit ka kota.


U. PV-order by father-DEF will go to town
‗Ujang was ordered by his father to go to town.‘

(18) Ujang di-anggap (ku) bapa-na rék indit ka kota.


U. PV-assume by father-DEF will go to town
‗Ujang was assumed by his father to go to town.‘

2.2.2 ‗Complementizers‘

Control complements can be prefaced by numerous yet synonymous elements, namely


sina/sangkan/ngarah. They are generally interchangeable, although there is slight difference.
Sangkan and ngarah allow resumptive pronouns to appear in the complement clause (19), but
sina does not (20).

(19) Ujang di-titah (ku) Ema sangkan/ngarah manéhna indit ka kota.


U. PV-order by mother so that he go to town
‗Ujang was ordered by mother to go to town.‘

(20) *Ujang di-titah (ku) Ema sina manéhna indit ka kota.


U. PV-order by mother so that he go to town
‗Ujang was ordered by mother to go to town.‘

Some of the raising predicates (believe and expect) optionally take jang-clauses, which are ill-
formed with the ‗tensed‘ auxiliaries (21), indicating a competition between jang and the tensed
auxiliaries for one single landing site.

(21) *Ujang di-percaya (ku) bapa jang rék ng-anteur-keun duit ka kota.
U. PV-believe by father to will AV-send-APPL money to town
‗Ujang was believed by the father that he would send the money to town.‘

7
The term ‗tensed‘ is put in quotation marks in conjunction with auxiliaries in light of the fact that it is not clear yet
how tense or finiteness operate in Sundanese. Although Arka (1998) analyzed equivalent morphemes in Indonesian
as finite auxiliaries, such an analysis needs to be warranted since just like Sundanese, Indonesian does not exhibit
outward signs of finiteness.
2.2.3 Adverbs

Epistemic adverbs such as sigana ‘possibly‘ can distinguish embedded CP clauses from raising
and control ones. Unsurprisingly, embedded CP clauses handily admit epistemic adverbs (22),
while raising (23) and control complements (24) do not.

(22) Acéng di-omong-keun (ku) Siti yén manéhna sigana rék munggah haji.
A. PV-talk-APPL by S. COMP he possibly will go pilgrimage
‗Siti said about Aceng that he would possibly go on pilgrimage.‘

(23) Acéng di-sangka (ku) Siti (*sigana) rék munggah haji.


A. PV-suspect by S. possibly will go pilgrimage
‗Aceng was suspected by Siti to have possibly gone on pilgrimage.‘

(24) Acéng di-pénta (ku) Siti sangkan (*sigana) munggah haji.


A. PV-ask by S. so that possibly go pilgrimage
‗Aceng was asked by Siti to possibly go on pilgrimage.‘

In summary, there are substantial syntactic and semantic differences between raising and
control in Sundanese.

3. Analysis

We have seen that control verbs do not admit the ‗tensed‘ auxiliaries in their complement
clauses, as in (17), repeated below.

(17) Ujang di-titah (ku) bapa-na (*rék) indit ka kota.


U. PV-order by father-DEF will go to town
‗Ujang was ordered by his father to go to town.‘

The ungrammaticality in (17) could be due to the fact that control complements optionally
introduced by sangkan-type morphemes cannot co-exist with the ‗tensed‘ auxiliary rék ‗will‘.8
One plausible explanation to capture this fact is that control complements simply do not project
tense, which is a potential landing site for the ‗tensed‘ auxiliaries.
This proposal is further backed up by the fact that the complements of control verbs can
be extracted or fronted, but those of raising ones cannot. Recall (15-16), repeated in the
following.

8
One might argue that the incompatibility of the ‗tensed‘ auxiliaries such as rék ‗will‘ could be due to some
semantic clash. This, however, is untenable since the semantics of control complements, which according to Bresnan
(1972) requires an ―unrealized future tense‖, should be well suited with the semantics of the future-denoting
auxiliaries such as rék .
(25) Sina nga-latih Persib, Ujang di-titah (ku) abah téh.
So that AV-train P U PV-order by grandfather PAR
‗Ujang was ordered by grandfather to coach Persib.‘

(26) ?*Jang nga-latih Persib, Ujang di-harep-keun (ku) abah téh.


to AV-train P U PV-expect-APPL by grandfather. PAR
‗Ujang was expected to coach Persib.‘

As illustrated above, it is licit to extract the complement of a control verb (15), but illicit to
extract the complement of a raising verb (16). This extractability of control complements points
to the relative smallness of their embedded structure. If the complement is no larger than a
VoiceP9, such fronting follows naturally. And, I propose that the sangkan-type morphemes will
be adjoined to this projection. The proposed structure for control complements is schematized as
follows.

(25) a. [FP sangkan/ngarah [VoiceP pro10 [vP <pro> embedded verb]]]


b.
VoiceP

sangkan/ngarah VoiceP

pro Voice‘

Voice vP
me-/di-
<pro> v‘

v VP
I propose pro merges in the specifier of vP in the lower clause due to the possibility of having a
resumptive pronoun obligatorily coreferential with the object of the higher clause. It originates in
Spec, vP because it is a position in which pro gets assigned an agent thematic role. It will then
move to the specifier of VoiceP to satisfy some requirement imposed by the voice marking on
the verb.11 And as noted earlier, sangkan-type morphemes are then adjoined to VoiceP. As such,
the morphemes will be able to license the occurrence of pro.

9
In spirit, I follow Son (2006) and Son & Cole (2008)‘s proposal of VoiceP, which is apparently an extended
proposal of Kratzer‘s (1996) and Pylkkänen‘s (2002). This voice projection is assumed to be responsible for
introducing an external argument. In my proposal, though, VoiceP will be responsible solely for licensing a
grammatical subject in active sentences, since voice markers in Sundanese can occur in various types of verbs
including unaccusatives and inchoatives. I will not delve into this proposal further here.
10
Recall that this possibility only goes with sangkan and ngarah based on the data in (19-20). When the control
complement headed by sina, however, PRO will be merged in place of pro.
11
Voice marking is observed in most of the verbs regardless the types, although in most cases, it goes together with
transitive verbs. It can be found pervasively in inchoative intransitive verbs such as nga-gedé-an ‗become bigger‘,
In regards to raising, as its complements allow the ‗tensed‘ auxiliaries, its embedded
structure includes a tense projection, making it a larger structure than that of control
complements. This proposal gains some support from the fact that jang cannot co-occur with the
‗tensed‘ auxiliaries, as illustrated in (21), repeated below.

(21) *Ujang di-percaya (ku) bapa jang rék ng-anteur-keun duit ka kota.
U. PV-believe by father to will AV-send-APPL money to town
‗Ujang was believed by the father that he would send the money to town.‘

Given its incompatibility with the ‗tensed‘ auxiliaries, jang can be analyzed as an ‗untensed‘
auxiliary that competes for the same slot with the ‗tensed‘ counterparts.
The impossibility of admitting epistemic adverbs into raising complements provides
further empirical support for this proposal. This is given in (23), repeated below.

(23) Acéng di-sangka (ku) Siti (*sigana) rék munggah haji.


A. PV-suspect by S. possibly will go pilgrimage
‗Aceng was suspected by Siti to have possibly gone on pilgrimage.‘

The ill-formedness of (23) is attributed to the fact that an epistemic adverb appears in the raising
complement. This leads me to conjecture that raising complements are not as structurally large as
regular CP complements in that they lack any node above tense projection. The proposed
structure for raising complements is given in (26).

(26) a. [TP jang/rék [VoiceP embedded subject [vP <embedded subject> embedded verb]]]
b.
TP

DP T‘

jang/rék VoiceP

<DP> Voice‘

Voice vP
me-/di-
<DP> v‘

v VP

nga-leutik-an ‗become smaller‘, manjang-an ‗become longer‘, mondok-an ‗become shorter‘, and etc. It can also
occur in some stative verbs such as ngalir ‗flow‘, ngeclak ‗drip‘, nga-golontor ‗flow rapidly‘, and etc. Hence,
VoiceP is postulated to harbor the voice marker on the verbs. I leave open what features trigger the movement of
subject to the specifier of VoiceP, since it goes beyond the issue under investigation.
The embedded subject DP originates in the [Spec, VoiceP] to be theta marked, after which it
raises to the [Spec, TP]. jang or rék can then head the tense projection.
To recap, raising and control differ structurally. Raising complements include a tense
projection, whereas control complements do not.

5. Implications and Remaining Issues

The Sundanese structures have implications for perhaps the most central debate in current
analyses of raising and control. In traditional approaches to control within the Principles and
Parameters framework, raising and control have been analyzed as two substantially different
structures. While raising is the by-product of movement, control is produced by having a base-
generated PRO as the subject of the lower clause coindexed with an argument of the higher
clause. This distinction is crucially based on the following major assumptions: (i) an argument
chain bears only one theta role; (ii) movement to a theta position is barred; (iii) theta roles are not
checkable features; (iv) PRO and DP-trace are distinct; and (v) the control module that governs
the controller of PRO and its interpretation exists. Consider (29-30).

(27) John tended to embarrass Kim.

(28) John tried to embarrass Kim.

How is a control construction like (28) derived under this traditional view? The
derivation for (28) is given in (30). PRO is merged in [Spec, vP], where it is assigned a thematic
role by the embedded verb. It raises to the embedded [Spec, TP] to satisfy the EPP feature on
embedded T. Null C is then merged, bearing the case feature [null], which values the case feature
on PRO. John is merged in the matrix [Spec, vP], where it receives a θ-role; and finally it ends
up in the matrix [Spec, TP]. The derivation for the raising sentence (27) significantly differs
from that of control, as shown in (29). John is base-generated in the embedded [Spec, vP] where
it gets a theta role. It moves first to the embedded [Spec, TP] for EPP reasons, then moves on to
the matrix [Spec, vP] to have its case feature valued by the case feature on the matrix T. John
finally sits in the matrix [Spec, TP] to satisfy EPP.

(29) [TP Johni [vP Johni tended [TP Johni to [vP Johni to embarrass Kim]]]]

(30) [TP Johni [vP Johni tried [CP [TP PROi to [vP PROi to embarrass Kim]]]]]

Hornstein (1999), in a more recent approach to control, has departed from this traditional
view of control and instead proposes what Landau (2003) referred to as the reductionist view of
control. Hornstein analyzes control as movement from one argument position to another. On this
‗radical‘ view, both raising and control are now derived via movement, and the difference being
merely that the movement under control is linked to a thematic position, whereas that of raising
is non-thematic. So, the sentence in (28) is derived as in (32). John merges in the lower vP,
where it receives a θ-role; it first moves to the embedded [Spec, TP]; it then moves further to the
matrix [Spec, vP], where it receives another θ-role; and finally it ends up in the matrix [Spec,
TP]. Meanwhile, the derivation for the raising sentence (27) is essentially identical to the one for
the control sentence, except that John is not theta marked when it moves to the matrix [Spec, vP]
on its way to the matrix [Spec, TP]. This is shown in (31).

(31) [TP John [vP John tended [TP John to [vP John to embarrass Kim]]]]

(32) [TP John [vP John tried [CP [TP John to [vP John to embarrass Kim]]]]]

One of the most far-reaching consequences of this innovation, despite its apparent
elegance, is that the contrast between control and raising is no longer substantial. In this view,
both constructions are now derived via movement, the difference being merely that the
movement under control is linked to a thematic position, whereas that of raising is non-thematic.
This reductionist view of control faced many challenges. Culicover and Jackendoff
(2001), for instance, argue that Hornstein‘s movement theory of control is untenable, by
presenting some data that show that ―the position of the controller is determined in part by
semantic constraints‖. A semantic account can naturally capture the generalizations in a manner
impossible for the purely syntactic account. Along the same lines, Landau (2003) argues
―Hornstein‘s system overgenerates nonexisting structures and interpretations … and a wide
empirical range of raising/control contrasts makes an overwhelming case against the reductionist
view‖.
One empirical question arises, i.e. is this movement theory of control consistent with the
facts of raising and control constructions in Sundanese? Recall that this movement view of
control essentially reduces certain cases of control to movement. Laundau (2003) noted that this
view is not an attempt to obliterate the distinction between raising and control, since it still
recognizes the single difference between the two, namely the fact that the raised DP is ascribed
two thematic roles in control but only one role in raising. This entails that raising and control
have the same syntactic structure, which appears to be a prerequisite for the movement theory of
control to stand. If one can reveal systematic differences between the two constructions, then the
theory will not be supported. Put differently, if it can be shown that there is an incontrovertible
structural contrast between raising and control in Sundanese, then it seems reasonable to argue
that the movement theory of control is untenable in Sundanese.
To that end, I have shown in Section Two that raising and control are profoundly
dissimilar in a number of peculiar properties summarized in Table 1.

Raising Control
complement extraction out of complement Very marginal Acceptable
clause
‗tensed‘ auxiliaries in complement clause Yes No
‗complementizers‘ jang sina/sangkan/ngarah
overt pronoun in complement clause No yes with sangkan/ngarah
Table 1. Some peculiar properties of raising and control in Sundanese
It is immediately clear from Table 1 that the Sundanese facts furnish such a compelling argument
against Hornstein‘s movement theory because raising and control exhibit salient structural
distinctions that are impossible to reduce solely to a semantic difference. The theory would be
forced to invoke a set of stipulations to handle each of those properties, which eventually
introduce more complexities than simplification. If this is what happens, then the theory has to
abandon its major promise, i.e. grammatical downsizing.
This means that until a new control theory is conceived, Sundanese raising and control
facts can be best accounted for by the classic, well-established control theory under the
Principles and Parameters framework.
Another important thing to note is that my proposed analysis for Sundanese runs counter
to the traditional assumption about the embedded structure of raising and control. In English and
most other languages for which raising and control constructions have been analyzed, the control
structure is 'larger' than the raising structure, but in Sundanese it is the opposite.
What remain to be done for future research are (i) determining the exact syntactic and
semantic contribution of the various morphemes—namely sina, sangkan, ngarah—that give rise
to different kinds of subjects in control complements; and (ii) examining the intimate relation
between sina and PRO on the one hand, and between sangkan/ngarah and little pro or
resumptive pronouns on the other.
Summarizing, I have argued, on the basis of syntactic and semantic facts, that the
traditional distinction between raising and control must be maintained, contra Hornstein‘s
movement theory of control. I have also shown that raising complements comprise ‗larger‘
structure than control counterparts, contra the traditional assumption that control predicates have
embedded CP structure.

References

Arka, I Wayan. 2000. Control theory and argument structure: Explaining control into subject in
Indonesian. Paper presented at the 4th International Malay and Indonesian Symposium,
Jakarta.

Carnie, Andrew. 2002. Syntax: a generative introduction. Oxford: Blackwell.

Chomsky, Noam. 1981. Lectures on government and binding. Dordrecht: Foris.

Chomsky, Noam. 1995. The minimalist program. Cambridge, Mas.: MIT Press.
Cinque, Guglielmo. 1999. Adverbs and functional heads: A cross-linguistic perspective. New
York: Oxford University Press.

Cullicover Peter W. and Ray Jackendoff. 2001. Control is not movement. Linguistic Inquiry,
Vol. 32, No. 3, pp. 493-512.
Davies, William D. 2005. Madurese prolepsis and its implications for a typology of raising.
Language 81:645–665.
Davies, William D., and Stanley Dubinsky. 2004. The grammar of raising: a course in syntactic
argumentation. Oxford: Blackwell.
Flegg, Jill and Ileana Paul. 2002. On the difference between Control and Raising. Paper
presented at the ninth annual meeting of the Austronesian Formal Linguistics
Association, Ithaca, NY, Cornell University, April 26-28.
Hornstein, Norbert. 1999. Movement and control. Linguistic Inquiry, Vol. 30, No. 1, pp. 69-96.
Jacobson, Pauline.1992. Raising without movement. In Control and Grammar, eds. by Richard
K. Larson, Sabine Iatridou, Utpal Lahiri, and James Higginbotham, 149–94. Dordrecht:
Kluwer.
Koizumi, Masatoshi. 1995. Phrase structure in minimalist syntax. Doctoral dissertation, MIT,
Cambridge, Mass.
Kratzer, A. 1996. Severing the external argument from its verb. In Phrase structure and lexicon,
ed. by Johan Rooryck and Laurie Zaring, 109-137. Dordrecht: Kluwer.

Laundau, Idan. 2003. Movement out of control. Linguistic Inquiry, Vol. 34, No. 3, pp. 471-498.
Pylkkänen, L. 2002. Introducing arguments. Doctoral dissertation, MIT, Cambridge, Mass.

Rosenbaum, Peter S. 1967. The grammar of English predicate complement constructions.


Cambridge, Mas.: MIT Press.
Son, M. 2006. Causation and syntactic decomposition of events. Doctoral dissertation,
University of Delaware.

Son, M. J. and Peter Cole. 2008. Syntactic Decomposition of Events in Korean and Standard
Indonesian. In Event Structures in Linguistic Form and Interpretation, eds. by Johannes
Dölling, Tatjana Heyde-Zybatow and Martin Schäfer, Language, Context and Cognition
5: 55-80, Mouton-de Gruyter.

Wechsler, S. and I. Wayan Arka. 1998. Syntactic ergativity in Balinese: an argument structure
based theory," Natural Language and Linguistic Theory 16:387-441.
Wouk, Fay and Malcolm Ross (eds.). 2002. The history and typology of western Austronesian
voice systems. Canberra: Pacific Linguistics.

Você também pode gostar