antipassive in austronesian alignment change edith aldridge
TRANSCRIPT
Antipassive in Austronesian Alignment Change
Edith Aldridge, University of Washington
1 Introduction
Linguists continue to debate the question of ergativity in Austronesian languages. Many have
made a case for an ergative analysis (Payne 1982; Cooreman 1982; Hopper 1983; Gerdts 1988;
De Guzman 1988; and Verhaar 1988; Gibson and Starosta 1990; Huang 1994; Liao 2002, 2004;
Aldridge 2004), while others maintain an accusative perspective (Keenan 1976, Bell 1983,
Davies 1991, Kroeger 1993, Rackowski 2002, Rackowski and Richards 2005). Yet others have
put forth split-ergative proposals (Maclachlan 1996, Maclachlan and Nakamura 1997, Chang
1997, Wechsler and Arka 1998, Arka 1998, van de Visser 2003, Paul and Travis 2006) or
relegate these languages to their own typological class (Schachter 1976, 1984, 1994). What I
argue in this paper is that these languages do not all belong to a single typological class. Rather,
some are ergative, many others exhibit a split-ergative system, and yet others are predominantly
accusative. But what makes this variation interesting is that it can be accounted for in terms of a
historical continuum: an ergative language evolves into a split-ergative language, which in turn
eventually becomes an accusative language.
It is well-known that some accusative languages have developed an ergative case-marking
pattern by reanalyzing passive clauses as active and transitive (Anderson 1977, Estival & Myhill
1988, and others). In a passive clause, an internal argument is promoted to subject status, while
the underlying subject is treated as an oblique. If this clause type is reanalyzed as transitive, it
takes on the appearance of a transitive clause in an ergative language. This progression can be
1
schematized as follows. Case-marker A is nominative or absolutive. Marker B represents non-
nominative/absolutive, including oblique, ergative, and accusative.
(1) Accusative Transitive => Passive => Ergative
V NPA NPB V (NPB) NPA V NPB NPA
As for the transition from ergative to accusative, I propose that this process also begins in an
intransitive construction, specifically an antipassive. An antipassive is semantically transitive, in
that it contains two DP arguments. However, case-marking follows an intransitive pattern: the
external argument has absolutive case, while the internal argument is marked as an oblique. This
yields a mapping from semantic to grammatical relations which is parallel to transitive clauses in
accusative languages.
(2) Intransitive Clauses in Ergative Languages
VI NPA
VAP NPA NPB
I propose that an ergative language becomes split-ergative by reanalysis of its antipassive
construction as syntactically transitive. A split-ergative language then can evolve into an
accusative language through the further reanalysis of transitive ergative clauses as passive. I
illustrate this continuum below with the ergative language Tagalog, the split-ergative languages
Malagasy and Seediq, and the predominantly accusative standard Indonesian.
2
2 Analysis of ergativity
The analysis of case in an ergative language like Tagalog makes it particularly clear how the
change from ergative to accusative syntax can begin in intransitive, especially antipassive
clauses. Contra many well-known approaches to case in ergative languages, in which absolutive
case is associated across the board with subject position (Murasugi 1992, Bittner and Hale 1996,
Ura 2000, and others), I follow Aldridge (2004, 2008) in proposing that absolutive case-valuing
is shared by T and v1. Absolutive case is valued by T only in intransitive clauses. In a transitive
clause, v values absolutive case on the first DP in its c-command domain. Following Mahajan
(1989), Woolford (1997, 2006), and Legate (2002, 2008), I assume that ergative is inherent case
assigned by v to the external argument.
(3) a B<in>ili ng babae ang isda.
<Tr.Perf>buy Erg woman Abs fish
‘The woman bought the fish.’
b TP
V+v+T vP DP[Erg] v’ tV+v[uCase:Abs] VP tV DP[Case: ]
In intransitive clauses, v does not have a case feature, so case must be valued on the subject by T.
The object receives inherent oblique case from the verb. Intuitively, this suggests that
absolutives are expected to have the characteristics of direct objects in transitive clauses and
3
behave as subjects only in intransitive clauses. In terms of the historical analysis being
developed here, the prediction is that reanalysis of absolutives as subjects should take place in
intransitive, e.g. antipassive, constructions.
(4) a B<um>ili ang babae ng isda.
<Intr.Perf>buy Abs woman Obl fish
‘The woman bought a fish.
b TP
V+v+T[uCase :Abs] vP DP[Case: ] v’ tV+v VP tV DP[Obl]
The case features valued by T and transitive v are summarized below. Note that the absolutive
case feature on T is optional. Since this feature is uninterpretable, the derivation will converge
only if T has an absolutive case feature in intransitive clauses and does not have a case feature in
transitive clauses. This is because it is only in intransitive clauses that there will be a DP with an
unvalued case-feature to check the uninterpretable case feature on T.
4
(5) Ergative language (Tagalog)
vTr: Inherent ergative case
[uCase:Abs]
vIntr: No case feature
TFin: Optional [uCase:Abs]
2.1 Subject properties of the external argument
The analysis sketched above accords well with the well-documented split in subject properties
found in ergative languages. As shown by Anderson (1976), Larsen and Norman (1979), Payne
(1982), Dixon (1994), Manning (1996), among many others, subject properties relating to
binding and control reside with the ergative DP in a transitive clause and the absolutive in an
intransitive clause. For example, (6a) shows that an ergative antecedent binds an absolutive
reflexive. In the antipassive in (6b), the absolutive external argument is the binder.
(6) a P<in>igil ng lalaki ang sarili=niya.
<Tr.Perf>control Erg man Abs self=3sg.Gen
‘The man controlled himself.’
b Nag-pigil=siya sa sarili=niya.
Intr.Perf-control=3sg.Abs Dat self=3sg.Gen
‘He controlled himself.’
In (7), the ergative and absolutive external arguments serve as imperative and hortative
addressees.
5
(7) a Bigy-an=mo=siya ng kape.
give-App=2sg.Erg=3sg.Abs Obl coffee
‘Give him the coffee.’
b K<um>ain=na=tayo.
<Intr.Perf>eat=now=1p.Abs
‘Let’s eat now!’
Controlled PRO also appears in external argument position.
(8) a Nagba-balak si Maria-ng [PRO tulung-an si Pedro]
Intr.Prog-plan Abs Maria-Lk (Erg) help-App Abs Pedro
‘Maria is planning to help Pedro.’
b Gusto ni Maria-ng [PRO b<um>ili ng libro]
want Erg Maria-Lk (Abs) <Intr.Perf>buy Obl book
‘Maria wants to buy a book.’
6
Note that both the matrix and embedded clauses in (8a) contain an overt absolutive DP,
indicating that PRO need not appear in the slot reserved for absolutive DPs. This also provides
further support for the analysis of case sketched above, i.e. that absolutive case in transitive
clauses is valued by v and not T. If the source of absolutive case were uniformly T (or whatever
vP-external functional head responsible for case-licensing subjects), as proposed by Murasugi
(1992), Bittner and Hale (1996), Ura (2000), and others, then we would not expect it to be
available in nonfinite contexts, where T is defective and unable to value case.
2.2 Division of labor in case valuing
The preceding discussion has shown that ergative DPs in transitive clauses and absolutives in
intransitive clauses exhibit syntactic behavior expected of subjects, indirectly supporting the
case-valuing analysis proposed at the beginning of this section. This subsection provides direct
evidence that absolutive case-valuing is shared by T and v.
Recall first from (8a) above that absolutive case is available for an internal argument in a
transitive nonfinite clause, indicating that T cannot be the source of this case. In contrast, (8b)
shows that PRO appears in absolutive position in an intransitive nonfinite clause. This supports
the current proposal that T values absolutive case in intransitive clauses, while v is the source of
this case in transitive contexts.
It is possible, however, for an overt absolutive to appear in subject position in a nonfinite
clause, provided that case is available exceptionally from matrix v, i.e. when matrix v is
transitive, as in (9a). Interestingly, antipassive, i.e. intransitive, v in (9b) is not able to case
license the embedded subject.
7
(9) a Bina-balak ni Maria-ng
Tr.Prog-plan Erg Maria-Lk
[makapagaral ang anak=niya sa UP]
Intr.study Abs child=3sg.Gen at UP
‘Maria is planning for her child to study at the University of the Philippines.’
b *Nagba-balak si Maria-ng
Intr.Prog-plan Abs Maria-Lk
[makapag-aral ang anak=niya sa UP]
Intr-study Abs child=3sg.Gen at UP
‘Maria is planning for her child to study at the University of the Philippines.’
This again demonstrates that the source of absolutive case is different in transitive and
intransitive clauses. Specifically, absolutive case is not available from intransitive – even
antipassive – v. A similar asymmetry can be observed in Tagalog applicative constructions. Only
transitive verbs can host applicative morphology, and the applied arguments always have
absolutive status. This is expected, since transitive v has an absolutive case feature which can
value case on the applied object. (10a) shows a transitive clause, with a benefactive PP
argument. In (10b), with the benefactive applicative i- prefixed to the verb, the benefactive
argument appears as the absolutive DP.
(10) a B<in>ili=ko ang libro para sa babae.
<Tr.Perf>buy=1sg.Erg Abs book for Dat woman
‘I bought the book for the woman.’
8
b I-b<in>ili=ko ng libro ang babae.
App-<Tr.Perf>buy=1sg.Erg Obl book Abs woman
‘I bought the woman a book.’
Tagalog antipassive verbs, on the other hand, cannot take applicative affixes. In the antipassive
in (11a), the benefactive argument can appear as a PP, but it cannot appear as a DP, with the
applicative prefix on the verb, as shown in (11b). This is unsurprising, since antipassive v is not
capable of case-licensing the applied DP. Note further that inherent oblique case should not be
available either, on the assumption that the applied argument is base merged in the specifier of
ApplP (in the sense of Pylkkanen 2002), rather than being selected by the lexical verb.
(11) a B<um>ili=ako ng libro para sa babae.
<Intr.Perf>buy=1sg.Abs Obl book for Dat woman
‘I bought a book for the woman.’
b *I-b<um>ili=ako ng libro ang babae.
App-<Tr.Perf>buy=1sg.Erg Obl book Abs woman
‘I bought the woman a book.’
The same point can be made with small clause subjects. Small clause subjects in Tagalog require
transitive morphology on v in order to be case-licensed. Under Hoekstra’s (1988, 1992) analysis
of small clauses, in which the embedded subject is a constituent of the small clause and not
selected by the lexical verb, inherent case is not available. Therefore, the embedded subject is
dependent on v for case-licensing.
9
Tagalog
(12) a Gina-gamit=niya [SC ang lalaki-ng alipin].
Tr.Prog-use=3sg.Erg Abs man-Lk slave
‘He/she uses the man as a slave.’
b *Guma-gamit=siya [SC ng lalaki-ng alipin].
Intr.Prog-use=3sg.Abs Obl man-Lk slave
‘He/she uses the man as a slave.’
This section has argued that absolutives behave as subjects only in intransitive clauses,
suggesting that the reanalysis of the absolutive as a subject should begin in intransitive contexts.
I have further shown that only transitive, but not intransitive or antipassive, v in Tagalog is
capable of valuing absolutive case on an internal argument. The source of absolutive case in
intransitive clauses must therefore be T. In the next section, I show that the first step in the
change from ergative to accusative syntax is the reanalysis of the antipassive construction as
transitive. Specifically, I show in the next section that Malagasy is distinguished from Tagalog in
that antipassive v is able to value structural accusative case.
3 First stage in the historical change - Malagasy
Malagasy is still predominantly ergative. (13) shows an ergative case-marking pattern. Malagasy
is a VOS language. As (13) shows, absolutives appear in clause-final position.
10
(13) a Nohanin’ny gidro ilay voankazo. (Pearson 2001: 88)
Past.AccP.eat.Det lemur that fruit
‘Rajaona bought the book.’
b Mandihy Rabe. (Paul & Travis 2006: 321)
AT.dance Rabe
‘Rabe is dancing.’
Ergative clauses like (13a) are clearly transitive and have not been reanalyzed as passive.
Evidence for this is that the ergative DP has the same subject properties as ergative nominals in
Tagalog. For example, the ergative DP can antecede reflexives and serve as the addressee in an
imperative construction, indicating that it has not been demoted to oblique status.
(14) a Hajain’ny vehivavyi ny tenanyi.
Respect.Gen.Det woman Det self
‘The woman respects herself.’
b Sasao ny lamba!
TT.wash.Imp Det cloth
‘Wash the clothes!’ (Paul & Travis 2006: 319)
Controlled PRO also appears in the ergative slot in a transitive nonfinite clause.
11
(15) Kasain-dRasoa [PRO hosasana ny zaza]
intend.TT.Gen.Rasoa Fut.TT.wash Det child
‘Rasoa intends to wash the child.’ (Paul and Travis 2006: 318)
However, Malagasy does not have an antipassive construction. The construction historically
descended from the antipassive2 appears to have the characteristics of an active transitive clause.
It is well known that objects in antipassive constructions cross-linguistically generally receive an
indefinite, narrow scope interpretation (Bittner 1987, 1994, 1995; Bittner and Hale 1996; Kalmar
1979; Cooreman 1994; Campbell 2000). However, the direct object, in a Malagasy antipassive
can be definite or can take wide scope over the external argument in Malagasy. Actor topic verbs
are prefixed with maN- (naN- in the past tense), which is cognate with the Tagalog intransitive
prefix maN-.
(16) a Nanapaka ity hazo ity tamin’ny antsy i Sahondra.
Past.AT.cut this tree this Past.P.Gen.Det knife Sahondra
‘Sahondra cut this tree with the knife.’
b Namaky ny boky roa ny mpianatra tsirairay.
Past.AT.read Det book two Det student each
‘Each student read two books.’ (2>ALL) (Paul & Travis 2006: 323)
This contrasts with Tagalog, which as an ergative language retains its antipassive construction.
Antipassive obliques in Tagalog are typically indefinite and nonspecific.
12
(17) K<um>ain=ako ng isda.
<Intr.Perf>eat=1sg.Abs Obl fish
‘I ate (a)/*the fish.’
Antipassive obliques in Tagalog also must take narrow scope with respect to the external
argument, as shown in (18a). In transitive clauses, in contrast, an absolutive object will take wide
scope over the ergative DP, as shown in (18b).
(18) a Nag-basa ang [lahat ng bata] ng [marami-ng libro].
-Intr.Perf-read Abs all Gen child Obl many-Lk book
‘All the children read many books.’
ALL > MANY
b B<in>asa ng [lahat ng bata] ang [marami-ng libro].
<Tr.Perf>read Erg all Gen child Abs many-Lk book
‘All the children read many books.’
MANY > ALL
Direct evidence that Malagasy active clauses are transitive and no longer antipassive comes from
the fact that structural case is available for the direct object. Malagasy has a type of object
promotion in active clauses whereby a PP can become a direct object. The PP in (19a) can be
repackaged as a DP direct object in (19b). As a DP, it requires case, indicating that the active
verb must have a case feature to value with it.
13
(19) a Nikapa ny hazo [PP tamin’ny famaky] i Soa.
Past-AT.cut Det tree Past.with.Gen.Det axe Det Soa
‘Soa cut the tree with the axe.’
b Nikapa famaky ny hazo i Soa.
Past.AT.cut axe Det tree Det Soa
‘Soa cut the tree with the axe.’ (Ileana Paul, personal communication)
Further evidence comes from the fact that antipassive v in Malagasy can exceptionally case-mark
an embedded subject. Note from section 2.2 that this was impossible in Tagalog.
(20) a Hitan’ ny mpampianatra [ianao namaky boky].
TT(Tr).see Det teacher 2sg.Nom Past.AT.read book
‘The teacher saw you reading a book.’
b Nahita [anao namaky boky] ny mpampianatra.
Past.AT(AP).see 2sg.Acc Past.AT.read book Det teacher
‘The teacher saw you reading a book.’ (Pearson 2008)
We have seen that ergative clauses in Malagasy continue to function as active and transitive. I
analyze Malagasy ergative clauses as parallel to their Tagalog counterparts. v values absolutive
case with an internal argument, while ergative case is inherent. However, the Malagasy
antipassive construction has been reanalyzed as transitive, i.e. antipassive v has acquired an
accusative case feature. This is the distinguishing feature differentiating Malagasy from Tagalog
in case-licensing.
14
(21) Malagasy: v-type split-ergative language
vTr: Inherent ergative case
[uCase:Abs]
vAP: [uCase:Acc]
vIntr: No case feature
TFin: Optional [uCase:Abs]
Malagasy can therefore be characterized as a split-ergative language, with accusative syntax
emerging in the former antipassive construction. The language otherwise remains ergative.
4. Second stage - Seediq
Further support for the proposal that change from ergative to accusative alignment begins in the
antipassive construction comes from Seediq, an Ataylic language spoken in Taiwan. Not only
has the antipassive construction been reanalyzed as transitive but the ergative construction has
taken one step toward being reanalyzed as intransitive. (22) shows that Seediq retains an ergative
case-marking pattern. Like Malagasy, Seediq is a VOS language and absolutives appear clause-
finally.
(22) a Wada kudurjak ka qedin=na.
Past flee Abs wife=3sg.Gen
‘His wife ran away.’
15
b Wada bube-un na Pihu ka dangi=na.
Past hit-Tr Erg Pihu Abs friend=3sg.Gen
‘Pihu hit his friend.’
Also like Malagasy, Seediq has lost its antipassive construction. The direct object can be
definite, as in (23).
(23) Wada beebu Pawan ka Awi-ni.
Past hit Pawan Abs Awi-Def
‘Awi hit Pawan.’
The object in an antipassive can also take wide scope with respect to the external argument, as in
(24b). (24a) provides the preceding context for (24b).
(24) a Hatang=ku m-bari teru bale, bulequn=ku suburo.
plan=1sg.Abs Intr.Fut-buy three only well=1sg.Abs rotten
‘I planned to buy only three, but when I took a good look, they were rotten.’
b Kiyaka ini=ku bari kanna.
so Neg=1sg.Abs buy.Intr.Irr all
‘So I didn’t buy all of them.” (ALL > Neg’
Crucially, structural case is available for the object. (25) shows exceptional case-marking. Recall
from section 2.2 that exceptional case-licensing is not available from antipassive v in Tagalog.
16
(25) Wada=ku m-ita [TP [vP m-imah sino] tama].
Past=1sg.Abs Intr.see Intr.drink wine father
‘I saw father drinking wine.’
Likewise, structural case is available in an antipassive for a small clause subject. Again, this was
shown to be impossible in Tagalog. Note that the bracketed constituent is clausal and not
nominal. In a nominal phrase with an adjectival modifier, the adjective would follow the noun,
e.g. qutsurh meluk ‘raw fish’.
(26) a Tanah tunuh m-ekan [meluk qutsuruh].
red head Intr-eat raw fish
‘Japanese (lit. red heads) eat [fish raw].’
b Wada bulebin [knedis sunuyuk] ka laqi.
Past pull.Intr long string Abs child
‘The child pulled [the string long].’
What differentiates Seediq from Malagasy is the lack of a structural case feature on transitive v.
Evidence for this comes from the fact that absolutive case never appears in a nonfinite clause. If
there were an absolutive case feature on transitive v, then an absolutive DP should be licensed in
a nonfinite clause and (27b) should be grammatical, contrary to fact. In other words, nonfinite
clauses must all be intransitive or antipassive, with PRO occupying subject position. The fact
that absolutive case does not surface in nonfinite clauses suggests that T is the sole source of this
17
case.
(27) a M-n-osa [PRO m-ari patis taihoku] ka Ape.
Intr-Perf-go Intr-buy book Taipei Abs Ape
‘Ape went to buy books in Taipei.’
b *M-n-osa [PRO burig-un taihoku (ka) patis] ka Ape.
Intr-Perf-go buy-Tr Taipei Abs book Abs Ape
‘Ape went to buy books in Taipei.’
The change from Malagasy to Seediq, then, is the loss of the case feature on transitive v.
Consequently, the absolutive case feature has also become obligatory on finite T.
(28) Seediq: T-type split-ergative language
vTr: Inherent ergative case
No case feature
vAP: [uCase:Acc]
vIntr: No case feature
TFin: [uCase:Abs]
The fact that two changes have taken place in Seediq indicates that Seediq has developed further
in the direction of accusative syntax. The fact that the transitivization of antipassive v in
Malagasy is the sole parameter distinguishing the case system of Malagasy from Tagalog argues
that reanalysis of the antipassive is the first stage in the historical continuum. This step can then
18
be followed by detransitivization of ergative v, as observed in Seediq.
Before turning to the next section, I point out that a word order change can be observed to
have taken place in these languages concomitant with alignment change. Tagalog is a VSO
language with relatively free word order. Malagasy and Seediq are both VOS. Particularly
relevant is the fact that the absolutive DP must appear in clause-final position. Aldridge (2010)
proposes that VOS order correlates high topicality of the absolutive DP, requiring this argument
to move to a clause-peripheral topic position. Standard Indonesian, which I discuss in the next
section, is an SVO language. The clause-initial DP has characteristics of both A’-position topics
and A-position subjects, suggesting that it is presently in the process of being reanalyzed from
topic to subject. I suggest here that this change is at least indirectly related to the alignment
change. Since the change from ergative to accusative alignment begins in the antipassive
construction, in which the absolutive argument is also a semantic subject, then one characteristic
of alignment change is the emergence of a subject grammatical function. The reanalysis of the
absolutive as subject is mediated by topicalization, i.e. by giving prominence to this grammatical
function by dislocating it to a clause-peripheral topic position.
I additionally point out that Malagasy and Seediq also have relatively impoverished
morphological case. Absolutives appear in clause-final topic position, but no morphological
case-marking distinguishes the clause-final argument from a clause-internal direct object in
Malagasy. This can be seen in (20) above. Ny is a determiner and does not mark case. Garrett
(1990) has proposed that a change in a language’s case-marking pattern can play a role in
accusative and ergative alignment change. It could be suggested for Malagasy that the structural
reanalysis in the former antipassive construction is the result of the loss of oblique case-marking
on the direct object. In theoretical terms, the disappearance of the case-marker might signal the
19
loss of inherent case and subsequent emergence of a structural case feature on v for the direct
object.
However, if we look at the Seediq examples, the absolutive case-marker is retained. In this
language, there is clearly a three-way case distinction: ergative (na), absolutive (ka), and
accusative (null). Therefore, there is no evidence to suggest that the presence or absence of
morphological case-marking is the trigger for either alignment or word order change. Seediq
makes it clear that changes in structural case features on T and v, as well as the change to fixed
VOS word order, took place before the loss of morphological case.
5 Third stage exemplified - Indonesian
The transition which reanalyzes transitive ergative clauses as passive is exemplified by standard
Indonesian. Indonesian is nearly a fully accusative language and has no antipassive construction.
However, it retains some remnant aspects of ergative syntax in the passive construction.
Standard Indonesian exhibits a nominative/accusative case-marking pattern. Nominative
subjects appear in clause-initial position. Note the meN- prefix on the active transitive verb in
(29a). This meN- is again cognate with Malagasy maN- and intransitive maN- in Tagalog.
(29) a Ali mem-beli buku.
Ali Act-buy buku
‘Ali bought a book.’
b Ali bekerja.
Ali work
‘Ali works.’
20
I showed in section 3 that maN- verbs in Malagasy are transitive; accusative case is available to
value with an internal argument. This is also true in Indonesian. Applicatives can occur on
active verbs. (30a) shows a mono-transitive clause with a DP direct object followed by a
benefactive PP. In (30b), the benefactive has been promoted to direct object status and appears
as a DP. Promotion of the object indicates that structural case must be available to license it.
(30) a Ali mem-beli buku pada Nuri.
Ali Act-buy book for Nuri
‘Ali bought a book for Nuri.’
b Ali mem-beli-kan Nuri buku.
Ali Act-buy-App Nuri book
‘Ali bought Nuri a book.’
Note additionally that the applied object will become the subject if the clause is passivized. This
provides additional evidence that this DP is the accusative case-marked direct object in (30b)
above.
(31) Nuri di-beli-kan oleh Ali.
Nuri Pass-buy-App buy Ali
‘Nuri was bought a book by Ali.’
21
Passives are clearly intransitive in some instances: they occur freely without an agent, indicating
that this argument has been demoted to oblique status.
(32) Banyak karya seni dapat di-beli di Indonesia.
many work art can Pass-buy in Indonesia
‘Many works of art can be bought in Indonesia.’ (Verhaar 1988: 350)
This indicates that active clauses are fully transitive in having structural case available for the
direct object, while passives are clearly intransitive. There is also evidence that, as in Seediq, T
is the sole source of nominative case. In nonfinite clauses, PRO will always occur in subject
position. Overt nominatives will not appear in nonfinite embedded clauses.
(33) a Saya ingin [PRO mem-ilih dia sebagai presiden.
1s want Act-choose 3s as president
‘I want to choose him as president.’
b Saya ingin [PRO di-pilih sebagai presiden.
1s want Pass-choose as president
‘I want to be chosen as president.’
Clearly, then, Indonesian passive v lacks both a case feature and the ability to select an external
argument.
The preceding discussion suggests that Indonesian is a fully accusative language, with T
valuing nominative case and v valuing accusative case. However, there are still remnants of
22
ergative syntax in Indonesian. It is well-known that not all passives in Indonesian are of the
canonical type seen in (32) (Arka and Manning, 1998; Chung, 1976; Cole and Hermon, 2005;
Guilfyole, Hung, and Travis, 1992; Musgrave, 2001a, 2001b; Sneddon, 1996; among others).
There is another type of passive3, in which the agent is expressed as a pronoun. The pronoun can
be a free or bound form. The free form pronominals can express any person, as shown in (34a).
1st and 2nd person have alternate proclitic forms, as shown in (34b).
Indonesian (Arka & Manning 1998:3)
(34) a. Buku itu saya/kamu/dia baca.
book that 1sg/2sg/3sg read
‘The book, I/you/(s)he read.’
b. Buku itu ku-/kau-baca.
book that 1sg/2sg-read
‘The book, I/you read.’
Crucially, there is evidence that pronominal passives are transitive, i.e. can select an external
argument. This is shown by the fact that these agents exhibit the behavior of subjects in that they
can antecede reflexives. (35a) shows examples of all types of pronominal passive: preverbal free
form, proclitic, and enclitic. These all pattern with subjects in active clauses with meN-, which
also have the ability to bind reflexives, as shown in (35b). In contrast, the agent in the cannonical
passives in (35c) is unable to bind a reflexive.
23
(35) a. Diri-saya saya serah-kan ke polisi. (Arka & Manning 1998:8)
self-1sg 1sg surrender-App to police
‘I surrendered myself to the police.’
b Saya men-yerah-kan diri saya ke polisi. (Arka & Manning 1998:4)
1sg Act-surrender-App self 1sg to police
‘I surrendered myself to the police.’
c ?*Diri-nya di-serah-kan ke polisi oleh Amir.
self-3sg.Gen Pass-surrender-App to police by Amir
“Himself was surrendered to the police by Amir.” (Arka & Manning 1998:5)
Further evidence of the transitivity of pronominal passives comes from the fact that agents can
serve as imperative addressees,1 a fact pointed out by Verhaar (1988).
(36) Kerja-kan hitungan itu!
solve.Pass-App sum that
‘Solve those sums!’
The evidence presented above indicates that the Indonesian clause type which descends from
antipassive is fully transitive. However, as shown above, remnant ergative syntax can still be
observed in pronominal passive clauses. In terms of the historical continuum I have proposed in
this paper, reanalysis of antipassive to transitive is complete. But reanalysis of transitive to
1 When the agent is first or second person, a passive verb does not take the di- prefix but rather appears as the bare
stem.
24
passive is still in progress. Again, this fact supports my position that the change from ergative to
accusative syntax in Austronesian languages is initiated in the antipassive construction and only
later extends to ergative clauses.
I will add before concluding this paper that historical change from ergative to accusative
syntax originating in an antipassive construction has been independently demonstrated for
languages outside of the Austronesian family. Harris (1985) shows that the direction of change
in Kartvelian languages is from ergative to accusative or active and that the series I accusative
case-marking pattern in modern Georgian has its historical origin in a derived intransitive
construction which had the case-marking and aspectual properties generally associated with
antipassives.
6. Conclusion
In this paper, I have proposed an account of varying degrees of ergative syntax among Western
Austronesian languages in terms of a historical continuum. I have shown how an ergative
language can evolve into a split-ergative language through the reanalysis of its antipassive
construction as active and transitive. Subsequently, a split-ergative language can become
accusative after the reanalysis of ergative clauses as passive.
I have suggested that the trigger for this change is the reinterpretation of the absolutive
nominal as a subject, which is most directly accomplished in intransitive clauses, including
antipassives, since intransitive absolutives have the syntactic properties generally attributed to
subjects in accusative languages.
25
References
Aldridge, Edith (2004). Ergativity and Word Order in Austronesian Languages. Ph.D.
dissertation, Cornell University.
Aldridge, Edith (2008). ‘Minimalist Analysis of Ergativity, Sophia Linguistica 55: 123-142.
Aldridge, Edith (2010). ‘Directionality in Word Order Change in Austronesian Languages’, in A.
Breitbarth, C. Lucas, S. Watts, D. Willis (eds.), Continuity and Change in Grammar. John
Benjamins, 169-180.
Anderson, Stephen (1976). ‘On the Notion of Subject in Ergative Languages’, in C. Li (ed.),
Subject and Topic. New York: Academic Press, 3-23.
Anderson, Stephen (1977). ‘On Mechanisms by Which Languages Become Ergative’, in C. Li
(ed.), Mechanisms of Syntactic Change. Austin: University of Texas Press, 317-363.
Arka, Wayan (1998). From Morphosyntax to Pragmatics in Balinese. Ph.D. dissertation,
University of Sydney.
Arka, Wayan and Christopher Manning (1998). ‘Voice and Grammatical Relations in
Indonesian: A New Perspective’, in M. Butt and T. H. King (eds.) Proceedings of the
LFG98 Conference. CSLI Online Publications.
Bell, Sarah (1983). ‘Advancements and Ascensions in Cebuano’, in D. Perlmutter (ed.), Studies
in Relational Grammar. Chicago: University of Chicago Press, 143-218.
Bittner, Maria (1987). ‘On the Semantics of the Greenlandic Antipassive and Related
Constructions’. International Journal of American Linguistics 53: 194-231.
Bittner, Maria (1994). ‘Case, Scope, and Binding’. Dordrecht: Kluwer Academic Publishers.
Bittner, Maria (1995). ‘Quantification in Eskimo: A Challenge for Compositional Semantics’, in
E. Bach, E. Jelinek, A. Kratzer, and B. Partee (eds.), Quantification in Natural
26
Languages. Dordrecht: Kluwer Academic Publishers, 59-80.
Bittner, Maria and Ken Hale (1996). ‘Ergativity: Toward a Theory of a Heterogeneous Class’.
Linguistic Inquiry 27: 531-604.
Campbell, Lyle (2000). ‘Valency-changing Derivations in K’iche’. in R. M. W. Dixon and A. Y.
Aikhenvald (eds.), Changing Valency: Case Studies in Transitivity. Cambridge:
Cambridge University Press, 236-281.
Chang, Yung-li (1997). Voice, Case and Agreement in Seediq and Kavalan. Ph.D. dissertation,
Hsinchu, Taiwan: National Tsing Hua University.
Chung, Sandra (1976). ‘On the Subject of Two Passives in Indonesian’, in C. Li (ed.), Subject
and Topic. Academic Press, New York, 59-98.
Cole, Peter and Gabriella Hermon (2005). ‘Subject and Non-subject Relativization in
Indonesian’. Journal of East Asian Linguistics 14: 59-88.
Cooreman, Ann (1982). ‘Topicality, Ergativity, and Transitivity in Narrative Discourse:
Evidence from Chamorro’. Studies in Language 3: 343-374.
Cooreman, Ann (1994). ‘A Functional Typology of Antipassive’ in B. Fox and P. Hopper (eds.),
Voice: Form and Function. Amsterdam: John Benjamins, 49-86.
Davies, William (1991). ‘Against an Ergative Analysis of Eastern Javanese’. Proceedings of
ESCOL, 77-89.
De Guzman, Videa P. (1988). ‘Ergative Analysis for Philippine Languages: An Analysis’, in R.
McGinn (ed.), Studies in Austronesian Linguistics. Athens, Ohio: Ohio University Center
for International Studies, 323-345.
Dixon, R.M.W (1994). Ergativity. Cambridge University Press.
Estival, Dominique and John Myhill (1988). ‘Formal and Functional Aspects of the
27
Development from Passive to Ergative Systems’ in M. Shibatani (ed.), Passive and
Voice. Amsterdam: John Benjamins, 441-491.
Garrett, Andrew (1990). ‘The Origin of NP Split Ergativity’. Language 66. 2: 261-296.
Gerdts, Donna B. (1988). ‘Antipassives and Causatives in Ilokana: Evidence for an Ergative
Analysis’, in R. McGinn (ed.), Studies in Austronesian Linguistics. Athens, Ohio: Ohio
University Center for International Studies, 295-321.
Gibson, Jeanne and Stanley Starosta (1990). ‘Ergativity East and West’, in P. Baldi (ed.),
Linguistic Change and Reconstruction Methodology. Berlin: Mouton de Gruyter, 195-
210.
Guilfoyle, Eithne, Henrietta Hung, and Lisa Travis (1992). ‘Spec of IP and Spec of VP: Two
Subjects in Austronesian Languages’, Natural Language and Linguistic Theory 10: 375-414.
Harris, Alice (1985). Diachronic Syntax: The Kartvelian Case. (Syntax and Semantics 18) New
York: Academic Press.
Hoekstra, Teun (1988). ‘Small Clause Results’. Lingua 74: 101-139.
Hoekstra, Teun (1992). ‘Aspect and Theta Theory’, in I.M. Roca (ed.), Thematic Structure: Its
Role in Grammar. Berlin: Foris Publications, 145-174.
Hopper, Paul (1983). ‘Ergative, Passive, and Active in Malay Narrative’ in F. Klein-Andreu
(ed.), Discourse Perspectives on Syntax. New York: Academic Press, 67-88.
Huang, Lillian (1994). ‘Ergativity in Atayal’. Oceanic Linguistics 33.1: 129-143.
Kalmar, I. (1979). ‘The Antipassive and Grammatical Relations in Eskimo’, in F. Plank (ed.),
Ergativity. New York: Academic Press, 117-143.
Keenan, Edward (1976). ‘Remarkable Subjects in Malagasy’, in C. Li (ed.), Subject and Topic.
New York: Academic Press, 249-299.
28
Kikusawa. Ritsuko (2002). Proto Central Pacific Ergativity: Its reconstruction and development
in the Fijian, Rotuman and Polynesian Languages. Canberra: Pacific Linguistics.
Kroeger, Paul (1993). Phrase Structure and Grammatical Relations in Tagalog. Stanford
University: Center for the Study of Language and Information.
Larsen, T. W. and W. M. Norman (1979). ‘Correlates of Ergativity in Mayan Grammar’, in F.
Plank (ed.), Ergativity: Towards a Theory of Grammatical Relations. New York:
Academic Press, 347-370.
Legate, Julie (2002). Warlpiri: Theoretical Implications. Ph.D. Dissertation, MIT.
Legate, Julie (2008). ‘Morphological and Abstract Case’. Linguistic Inquiry 39: 55-101.
Liao, Hsiu-chuan (2002). ‘The Interpretation of tu and Kavalan Ergativity’. Oceanic Linguistics
41.1: 140-158.
Liao, Hsiu-chuan (2004). Transitivity and Ergativity in Formosan and Philippine Languages.
Ph.D. dissertation, University of Hawaii.
Maclachlan, Anna. (1996). Aspects of Ergativity in Tagalog. Ph.D. dissertation, McGill
University.
Maclachlan, Anna and Masanori Nakamura (1997). ‘Case-checking and Specificity in Tagalog’.
The Linguistic Review 14: 307-333.
Mahajan, Anoop (1989). ‘Agreement and Agreement Projections’, in I. Laka & A. Mahajan
(eds.), Functional Heads and Clause Structure. MIT Working papers in Linguistics 10.
Cambridge, MA: MIT Working Papers in Linguistics, 217-252.
Manning, Christopher (1996). Ergativity: Argument Structure and Grammatical Relations.
Stanford: Center for the Study of Language and Information.
Murasugi, Kumiko G. (1992). Crossing and Nested Paths: NP Movement in Accusative and
29
Ergative Languages. Ph.D. dissertation, MIT.
Musgrave, Simon (2001a). Non-Subject Arguments in Indonesian. Ph.D. dissertation, University
of Melbourne.
Musgrave, Simon (2001b). ‘Pronouns and Morphology: Undergoer Subject Clauses in
Indonesian’, in: Booij, G. and van Marle, J. (eds.), Yearbook of Morphology 2000. Kluwer
Academic Publishers, Dordrecht, 155-186.
Palmer, F. R. (1994). Grammatical Roles and Relations. Cambridge: Cambridge University
Press.
Paul, Ileana and Lisa Travis (2006). ‘Ergativity in Austronesian Languages’, in A. Johns, D.
Massam, J. Ndayiragije (eds.), Ergativity: Emerging Issues. Dordrecht: Springer, 315-335.
Payne, Thomas E. (1982). ‘Role and Reference Related Subject Properties and Ergativity in
Yup’ip Eskimo and Tagalog’. Studies in Language 6.1: 75-106.
Pearson, Matthew (2001). The Clause Structure of Malagasy: A Minimalist Approach. Ph.D.
dissertation, UCLA.
Pearson, Matthew (2008). ‘Predicate-fronting and Constituent Order in Malagasy’. ms, Reed
College.
Pylkkanen, Liina (2002). Introducing Arguments. Ph.D. dissertation, MIT.
Rackowski, Andrea (2002). The Structure of Tagalog: Specificity, Voice, and the Distribution of
Arguments. Ph.D. dissertation, MIT.
Rackowski, Andrea and Norvin Richards (2005). ‘Phase Edge and Extraction: A Tagalog Case
Study’. Linguistic Inquiry 36.4: 565-599.
Schachter, Paul (1976). ‘The Subject in Philippine Languages: Topic, actor, actor-topic, or none
of the above’, in C. Li (ed), Subject and Topic. New York: Academic Press, 491-518.
30
31
Schachter, Paul (1984). ‘Semantic-Role-Based Syntax in Toba Batak’. UCLA Occasional Papers
in Linguistics 5: 122-149.
Schachter, Paul. 1994. The Subject in Tagalog: Still None of the Above. UCLA Occasional
Papers in Linguistics 15.
Sneddon, James Neil (1996). Indonesian: A Comprehensive Grammar. New York: Routledge.
Ura, Hiroyuki (2000). Checking Theory and Grammatical Functions in Universal Grammar.
Oxford University Press.
Verharr, John (1988). ‘Syntactic Ergativity in Contemporary Indonesian’, in R. McGinn (ed.),
Studies in Austronesian Linguistics. Athens, Ohio: Ohio University Center for
International Studies, 347-384.
van de Visser, Mario (2003). ‘Syntactic Ergativity in Dyirbal and Balinese’. Linguistics in the
Netherlands 2003: 177-188.
Wechsler, Stephen and Wayan Arka (1998). ‘Syntactic Ergativity in Balinese: An Argument
Structure Based Theory’. Natural Language and Linguistic Theory 16: 387-441.
Woolford, Ellen (1997). ‘Four-way Case Systems: Ergative, Nominative, Objective, and
Accusative’. Natural Language and Linguistic Theory 11: 679-728.
Woolford, Ellen (2006). ‘Lexical Case, Inherent Case, and Argument Structure’. Linguistic
Inquiry 37.1: 111-130.
1 Legate (2002, 2008) also proposes that T and v share the function of case licensing of
absolutive arguments in Warlpiri. A crucial difference between Legate’s and the current
approach, however, is that Legate assumes that the case features valued by T and v are different
– nominative and accusative, respectively, as in an accusative language – but are spelled out
morphologically as a single default form post-syntactically. However, this approach is not fully
32
satisfactory in accounting for the variation in case systems among the Austronesian languages,
since the split-ergative languages make use of both absolutive and accusative case for objects.
2 This construction is generally referred to in the literature as the “active” or “actor topic”
construction.
3 This construction has been referred to in various ways in the literature: ‘object preposing’
(Chung, 1976), ‘objective voice’ (Arka and Manning, 1998), ‘passive type two’ (Cole and
Hermon, 2005; Sneddon, 1996), etc. I employ the descriptive term ‘pronominal passive’ for the
expository part of the present section but adopt the formal designation ‘ergative’ later.